Chapter 15

MulteFire Alliance IoT technologies

Abstract

In this chapter, we describe how the MulteFire Alliance have adapted LTE-M and NB-IoT to support operation in unlicensed frequency bands according to FCC and ETSI regulations. The physical layer design as well as the procedures performed in idle and connected mode are presented for both technologies. The MulteFire design closely follows the 3GPP design, and the description is focused on the adaptations introduced to facilitate operation in unlicensed frequency bands. The chapter also presents the technical potential of the technologies in terms of the performance that can be expected when deploying these technologies.

Keywords

LTE-M; LTE-M-U; eMTC-U; NB-IoT; NB-IoT-U; LPWAN; ETSI; FCC; Unlicensed frequency band

1. 15.1 Background 
2. 15.2 LTE-M-U 
   1. 15.2.1 Radio access design principles 
      1. 15.2.1.1 FCC regulations 
      2. 15.2.1.2 ETSI regulations 
   2. 15.2.2 Physical layer 
      1. 15.2.2.1 Physical resources 
      2. 15.2.2.2 Transmission schemes 
      3. 15.2.2.3 Downlink physical channels and signals 
      4. 15.2.2.4 Uplink physical channels and signals 
   3. 15.2.3 Idle and connected mode procedures 
      1. 15.2.3.1 Cell selection and system information acquisition 
      2. 15.2.3.2 Paging 
      3. 15.2.3.3 Power control 
      4. 15.2.3.4 Medium utilization 
3. 15.3 NB-IoT-U 
   1. 15.3.1 Radio access design principles 
      1. 15.3.1.1 FCC regulations 
      2. 15.3.1.2 ETSI regulations 
   2. 15.3.2 Physical layer 
      1. 15.3.2.1 Physical resources 
      2. 15.3.2.2 Transmission schemes 
      3. 15.3.2.3 Downlink physical channels and signals 
      4. 15.3.2.4 Uplink physical channels and signals 
   3. 15.3.3 Idle and connected mode procedures 
      1. 15.3.3.1 Cell selection and system information acquisition 
      2. 15.3.3.2 Power control 
4. 15.4 Performance 
   1. 15.4.1 Performance objectives 
   2. 15.4.2 Coverage and data rates 
   3. 15.4.3 Latency 
   4. 15.4.4 Battery life 
5. References 

15.1. Background

The MulteFire Alliance (MFA) is a standardization organization that develops wireless technologies for operation in unlicensed and shared spectrum. The MFA specifications are using the 3GPP technologies as baseline and add adaptations needed for operation in unlicensed spectrum. In its first Release 1.0, the MFA developed an unlicensed-spectrum version of 3GPP LTE using the LTE Licensed-Assisted Access feature as baseline. The specification is also aligned with the 3GPP Release 14 enhanced Licensed-Assisted Access work, which was developed in parallel by 3GPP. In Release 1.1, the MFA turned its attention toward the 3GPP IoT technologies LTE-M and NB-IoT, which are described in Chapters 5 to 8. These two systems were modified for operation in unlicensed spectrum governed by ETSI and FCC.

Section 15.2 describes how LTE-M was redesigned to enable operation in the global unlicensed 2.4   GHz frequency band. The MFA named this technology eMTC-U, after the 3GPP Release 13 work item Further LTE Physical Layer Enhancements for MTC [3], which from time to time is referred to as the eMTC work item. In this book we use the term LTE-M when describing the set of LTE MTC enhancements specified in 3GPP (see Chapter 5), and LTE-M-U when describing the work done in MFA.

Section 15.3 focuses on the NB-IoT design for operation in the Short Range Device (SRD) and Industrial, Scientific and Medical (ISM) frequency bands defined in the range 863–870   MHz in Europe and 902–928   MHz in the US, respectively. The MFA has named this technology NB-IoT-U and in this book we use the same naming convention.

The descriptions of NB-IoT-U and LTE-M-U provided in the next 3 sections are based on the MulteFire Release 1.1 physical layer specifications [3–6] and higher layer specifications [7–9].

15.2. LTE-M-U

The interest among industries and enterprises to use wireless technologies for operating a private network to serve their needs has grown rapidly during the past few years. To use the 3GPP technologies typically requires a partnership with a spectrum license holder, often a mobile network operator, who grants a permit for sub-leasing the needed spectrum in a limited geographical area. An alternative available in certain regions is to use the 3GPP technologies in shared spectrum such as the Citizens Broadband Radio Service bands currently made available in the United States. The concept of sub-leasing is still in 2019 rarely used, while the availability of shared bands that allow operation of 3GPP technologies can at best be said to be emerging.

The MFA developed LTE-M-U to lower the threshold for industries and enterprises to make use of 3GPP technologies. The system is designed to operate in the 2400–2483.5   MHz license exempt frequency band. This is an attractive band due to its significant bandwidth and global availability, which makes the technology deployable in most countries in the world.

The next three sections introduce the design principles followed by MFA when developing LTE-M-U, the LTE-M-U physical layer functionality and the procedures used in RRC idle and connected mode.

15.2.1. Radio access design principles

The LTE-M-U design principles are based on three components: The 3GPP LTE-M design described in Chapter 5, the ETSI regulations for the 2.4   GHz band [1] and the FCC regulations for the same band [2]. In Release 1.1 the MFA fully focused on the ETSI and FCC regulations to prioritize an introduction of LTE-M-U in the European and US markets.

The physical layer of LTE-M-U closely follows the LTE-M design. The downlink is OFDM modulated while the uplink used DFT-precoded OFDM. LTE-M-U is designed as a time-division duplex (TDD) frequency hopping system with a system bandwidth of 6 physical resource blocks (PRB). Each PRB is of 180   kHz bandwidth which results in a useful system bandwidth of 1.08   MHz, usually referred to as a narrowband (NB) in the context of LTE-M and LTE-M-U. LTE-M-U supports CE mode A but not CE mode B (cf. Section 5.1.2.2). The rationale for this is that LTE-M-U is designed for small-cell private network deployments. For this use case, the CE mode A coverage range was deemed fully adequate. This design baseline has been carefully chosen to provide a single solution that complies with both FCC and ETSI regulations.

The next two sections introduce the FCC and ETSI regulations. Detailed descriptions of parts of the regulations are also provided e.g. in Sections 15.2.2.2.3, 15.2.2.2.4, 15.2.3.3 and 15.2.3.4.

15.2.1.1. FCC regulations

In the US, the FCC regulations specify support for three alternative system designs in the 2.4   GHz band: digitally modulated systems, frequency hopping spread spectrum (FHSS) systems or a hybrid system design which is a combination of the first two options.

The LTE-M-U downlink is operating according to the FCC rules for a digitally modulated system while the uplink is designed according to the regulations specified for a hybrid system. A digitally modulated system is required to have a transmit bandwidth of at least 500   kHz and meet a conducted power spectral density (PSD) limitation of 8   dBm/3   kHz, which corresponds to 25.8 dBm/180   kHz. As LTE-M-U base stations are mandated to transmit Cell-Specific Reference Signals (CRS) over the full downlink narrowband bandwidth of 1.08   MHz, and the base station conducted output power is limited to at most 30 dBm, which corresponds to a PSD of 4.4   dBm/3   kHz, both requirements are fulfilled. The uplink transmission bandwidth occupies between 1 and 6 physical resource blocks (PRBs), which implies that the digital modulation bandwidth requirement is not always met. Instead the uplink design complies with the FCC regulations for a hybrid system which is required to frequency hop and at the same time comply with the digital modulation PSD limitation of 8 dBm/3   kHz. As the LTE-M-U narrowband is frequency hopping over 16 or 32 carrier frequencies according to the FHSS regulations and the device PSD never exceeds 20   dBm/180   kHz, both the hybrid system frequency hopping and PSD criterions are met for the uplink design.

The uplink could in principle also be qualified as a frequency hopping system instead of a hybrid system. For a frequency hopping system, the device power class is defined by the maximum peak power, and not the maximum average power which is the case for digitally modulated and hybrid system designs. For LTE-M-U, being an OFDM system with a significant peak to average power ratio (PAPR), it is a benefit to declare the output power according to average, and not peak, power measurement procedures. This explains why the LTE-M-U uplink is declared as a hybrid system, and not a FHSS system when operating under FCC regulations.

As the downlink frequency hops with the uplink one may be led to believe that also the downlink can be declared as a FHSS. The FCC FHSS regulation however mandates an equal visitation of each frequency and a pseudo-random frequency hopping pattern. This condition is not fulfilled for the downlink as the anchor channel, carrying synchronization and broadcast signaling, is mapped on a fixed carrier frequency and is visited more frequently than the data channels. Note that visitation of a channel is here considered to be equivalent to transmission on the channel.

15.2.1.2. ETSI regulations

In Europe, ETSI specifies two categories of wideband data transmission systems in the 2.4   GHz band: FHSS systems and wideband modulation systems which include techniques such as OFDM and direct sequence spread spectrum modulation. The equipment can be adaptive or non-adaptive. Adaptive equipment is not permitted to transmit on an occupied radio channel and is required to sense the channel occupancy before accessing a channel. Non-adaptive equipment is not required to sense the channel, but needs to follow a specified medium utilization (MU) cycle which limits the channel occupancy time.

The ETSI wideband modulation regulation defines a PSD limitation of 10 dBm per MHz. Due to this restriction, LTE-M-U is declared as an FHSS system. The uplink is designed according to the non-adaptive FHSS regulations which implies that the LTE-M-U devices are limited to use a 10% MU when operating at 20   dBm Equivalent Isotropically Radiated Power (EIRP). While 3GPP specifications commonly refer to conducted power, ETSI and FCC regulations often define limits in terms of radiated power. To avoid a limitation of the downlink resource utilization, the LTE-M-U base stations are designed to follow the adaptive FHSS regulation. This requires the base station to perform channel sensing before accessing the medium.

Table 15.1 summarizes the radio access design principles followed by LTE-M-U. It is important to understand that LTE-M-U has a single design baseline that is defined by the most stringent requirements in the ETSI and FCC regulations. This means for example that the base station is expected to perform channel sensing not only when operating under ETSI regulations, but also when operating in the US under FCC regulations.

Table 15.1

LTE-M-U radio access design principles.

Regulation	Link direction	Design basis
FCC	Downlink	Digitally modulated
	Uplink	Hybrid
ETSI	Downlink	Adaptive FHSS
	Uplink	Non-adaptive FHSS

15.2.2. Physical layer

The description of the physical layer is distributed in four parts. The Physical resources section 15.2.2.1 presents the basic time and frequency resources used for transmitting the LTE-M-U signal in the 2.4   GHz band. The Transmission schemes section 15.2.2.2 presents generic aspects and functionality that applies to the physical layer design. The last two sections 15.2.2.3 and 15.2.2.4 present the details of the downlink and uplink physical channels and signals.

15.2.2.1. Physical resources

15.2.2.1.1. Channel raster

In 3GPP the channel bandwidth is defined as the occupied bandwidth of a modulated waveform which corresponds to the frequency range containing 99% of the total power of the modulated signal. An LTE-M-U narrowband of 6 PRBs have a 1.4   MHz channel bandwidth according to this definition which is reused by ETSI. The FCC regulations defines the channel bandwidth by its 20-dB bandwidth requirement. This corresponds to the bandwidth where the signal emissions are attenuated by 20   dB relative the power measured in the center of the carrier. The 20-dB bandwidth definition is more stringent than the 3GPP and ETSI occupied bandwidth requirement. MFA consequently concluded that the 20-dB bandwidth of an LTE-M-U narrowband should be within 1.8   MHz.

LTE-M-U supports the 2.4   GHz license exempt frequency band, defined by the frequency range 2400–2483.5   MHz. Across this band 43 radio frequency (RF) channels are distributed using a channel raster with a granularity of 1.8   MHz, which is aligned with the assumed 20-dB bandwidth. Three RF channels are reserved for the so-called anchor channel, which carries the synchronization signals and the most important broadcast and system information messages.

Each LTE-M-U cell is associated with one out of the three anchor channels. When an LTE-M-U device makes its initial cell selection it needs to scan only the three anchor RF channels in its search for a cell to camp on. The rationale for limiting the number of anchor frequencies to three is to limit the time it takes for a device to perform the initial system acquisition, and to maximize the number of channels available for data transmission.

The remaining 40 channels are dedicated for data transmission. They are grouped into 10 sets, each containing 4 channels. When frequency hopping is configured over 16 data channels, 4 of the 10 sets are configured for carrying physical uplink and downlink control channels, shared channels and reference signals. With frequency hopping configured over 32 channels, 8 of the sets are activated. The frequency hopping is further described in Section 15.2.3.6. Fig. 15.1 provides an illustration of a potential arrangement of the 43 LTE-M-U channels distributed over in total 77.4   MHz of the totally available 83.5   MHz in the 2.4   GHz band.

Fig. 15.1 Potential LTE-M-U channel arrangement in the 2.4   GHz band.

The upper edge of 3GPP band 40 is directly adjacent to the lower edge of the 2.4   GHz band. 3GPP Release 16 is studying a new band located immediately above the 2.4   GHz band. To secure coexistence with 3GPP systems operating in the adjacent bands, the outer LTE-M-U channel locations are expected to be shifted from the 2.4   GHz band edges. This will provide guard-band between the 3GPP bands and LTE-M-U. Coexistence between MFA and 3GPP technologies are for natural reasons important.

15.2.2.1.2. Frame structure

LTE-M-U introduces a new frame structure, called Type 3M. It is a TDD type of frame structure which is a natural choice for the 2.4   GHz unlicensed band since it is defined as a single unpaired band, without any reserved frequency ranges for uplink and downlink transmissions.

The frame structure is illustrated in Fig. 15.2, including the new mframe concept. One mframe contains 8 radio frames and spans 80   ms. The mframe numbering run across 8 hyperframes before it wraps around and starts over. Besides the mframe, the LTE-M-U frame numbering follows the same conventions as LTE.

The mframe is shown in Fig. 15.3. The first 5   ms is known as the anchor segment. It is dedicated to downlink transmissions of the LTE-M-U PSS, SSS, PBCH and the SIB1-A message, which are described in detail in Section 15.2.2.3. It is transmitted on an anchor frequency channel associated with the camped-on cell, as discussed in Section 15.2.2.1.1. Grouping these signals and channels together in the beginning of the mframe facilitates a frequency hopping design of the second part of the mframe known as the data segment.

The data segment spans 75   ms and is divided into uplink and downlink portions according to a set of specified uplink-downlink configurations introduced in Section 15.2.2.2. The uplink and downlink parts of the data segment support both control and data channel transmission that are mapped on a set of frequencies according to the configured frequency hopping pattern.

The subframes in an mframe are numbered according to a relative subframe number defined in the range 0–79.

Fig. 15.2 LTE-M-U Frame structure 3M including the mframe concept.

Fig. 15.3 LTE-M-U mframe including anchor and data segments.

15.2.2.1.3. Resource grid

The resource grid is just as for LTE-M defined by a PRB pair transmitted during a subframe of length 1   ms. In the frequency domain the PRB spans 180   kHz. LTE-M-U transmissions are mapped over 1 to 6   PRBs, with a granularity of 1 PRB.

Each PRB pair is mapped onto 2 consecutive time slots. Each slot is defined by a resource grid specified by 7 OFDM symbols (OS) in time and 12 subcarriers in frequency domain. Each subcarrier is 15   kHz wide.

The smallest building block in the resource grid is the resource element, defined by one OFDM symbol transmitted on one subcarrier. Section 5.2.1.3 provides an illustration of a PRB pair.

15.2.2.2. Transmission schemes

15.2.2.2.1. Anchor and data segment transmissions

The operation of LTE-M-U is based on the mframe concept introduced in Section 15.2.2.1.2. The first 5   ms in the mframe are used by the anchor channel while the remaining 75   ms are reserved for data transmissions. The anchor channel segment is dedicated to downlink transmissions. The exclusive use of downlink transmissions avoids near-far uplink-to-downlink interference situations that can otherwise be difficult to avoid in shared spectrum with uncoordinated deployments. The anchor frequencies are hence protected from uplink interference from LTE-M-U devices, but not from interference caused by other systems operating in the shared band.

The data segment is divided into uplink and downlink portions according to a set of specified uplink-downlink configurations. These are presented in Table 15.2. In LTE-M-U, the uplink-to-downlink switching occurs once or twice across 80 subframes, which is less frequent than in LTE-M. Remember that the LTE-M uplink-downlink configurations are defined across 10 subframes with 1 or 2 switching points. The long consecutive intervals of uplink or downlink subframes defined for LTE-M-U are intended to facilitate the concept of cross-subframe channel estimation (see e.g. Fig. 8.2) which improves operation under low SNR conditions. The taken approach of less frequent switching also optimizes the LTE-M-U channel capacity. This is an important aspect since LTE-M-U is a single-narrowband system where the configuration of additional narrowbands to increase system capacity is not supported. Uplink-downlink configuration 7 defines a somewhat more frequent switching to support a latency closer to what LTE-M TDD is capable of achieving.

The last two OS in every downlink subframe preceding an uplink subframe are left unused. This leaves room for RF switching and provides a guard-period that supports a cell radius of roughly 21   km. This explains why no special subframes are indicated in Table 15.2 (cf. Table 5.1 for LTE-M).

The system bandwidth is defined by a single narrowband containing 6 PRBs spanning a bandwidth of 1.08   MHz, or 1.4   MHz when including guard-bands. The anchor and data segments downlink transmissions are always spanning the full narrowband thanks to the CRS transmissions. The uplink transmission bandwidth ranges between 1 and 6 PRBs. Thanks to the frequency hopping the system may operate across the full 2.4   GHz band excluding the guard bands toward adjacent 3GPP bands.

Table 15.2

LTE-M-U data segment uplink-downlink configurations.

Configuration	Sequence of downlink (DL) and uplink (UL) subframes in the data segment
0		DL: 55				UL: 20
1		DL: 45				UL: 30
2		DL: 35				UL: 40
3		DL: 20				UL: 55
4	DL: 25		UL: 15		DL: 20		UL: 15
5	DL: 15		UL: 25		DL: 10		UL: 25
6	DL: 30		UL: 10		DL: 20		UL: 15
7	DL: 15	UL: 10		DL: 15	UL: 10	DL: 10		UL: 15

15.2.2.2.2. Transmission modes

The LTE-M downlink transmission modes TM1, TM2, TM6 and TM9 for up to 4 antenna ports are supported for Physical Downlink Shared Channel (PDSCH) transmissions on the data segment. The antenna port is a logical concept for defining the relation between the data and reference signal transmissions. It is also fundamental for supporting the precoding functionality categorized by the transmission modes. For all transmission modes LTE-M-U, just as LTE-M and in contrast to LTE, only supports single layer transmissions.

TM1 is defined for single antenna port transmission. TM1 can typically be associated with a single beamforming radiation pattern intended to provide coverage across an entire served cell. The applied precoder is transparent and common to all devices in a cell. TM2 is specified for 2 or 4 antenna ports and supports transmit diversity by means of Space Frequency Block Coding (SFBC). TM6 is for closed-loop precoding, where the base station applies a precoder, selected from a set of specified precoders, to create a beam optimizing the link quality. The device estimates the channel state based on the CRS transmissions and suggests an optimal precoder, from the available set, to the base station. TM9 is for open-loop precoding where the precoder applied by the network is transparent to the device. This requires the introduction of device-specific reference symbols that are precoded together with the data symbols. Section 5.2.4.7 in some further detail discusses precoding. Section 5.2.4.3 presents the LTE-M-U downlink reference signals and their use in the different transmission modes.

15.2.2.2.3. Listen-before-talk

Adaptive equipment operating according to ETSI regulations, i.e. LTE-M-U base stations, with an EIRP exceeding 10   dBm must perform a listen-before-talk based detect-and-avoid procedure before transmitting. This means that a base station must perform a clear channel assessment (CCA), and possibly an extended clear channel assessment (eCCA), in the beginning of each downlink dwell. A dwell is equivalent to a period of downlink or uplink transmission opportunities.

The CCA and eCCA are based on that no signal is detected at a power level exceeding a threshold TL defined as:

TL = − 70 dBm MHz + 10 log 10 ( P 100 mW )

(15.1)

P corresponds to the output EIRP of the LTE-M-U base station. A base station operating at the maximum output power of 20   dBm EIRP should hence refrain from transmitting on a downlink dwell in the beginning of which an energy exceeding -70   dBm per MHz is measured. At lower detected signal levels, the channel is available for transmissions.

In the context of low-power wide-area networks and compared to the sensitivity offered by a typical 3GPP system -70   dBm can be considered to correspond to a fairly high received signal level. For LTE-M-U being designed for industrial use cases with more moderate requirements on range and receiver sensitivity -70   dBm is expected to be a more representative receive signal level and therefore suitable as CCA threshold.

The CCA should be performed for 0.2% of the intended channel occupancy time, but for no less than 18   μs. In the context of CCA, occupancy is equivalent to transmission. If the CCA fails, the base station must perform an eCCA with a duration of up to 5% of the channel occupancy time. The application of the CCA and the eCCA procedures are in further detail described in Section 15.2.2.3.1 for the anchor segment and in Section 15.2.2.3.2.1 for the data segment.

15.2.2.2.4. Frequency hopping

Non-adaptive FHSS equipment operating according to ETSI regulations, i.e. LTE-M-U devices, shall meet the following conditions:

1. • Frequency hop over at least N frequency channels defined by 15   MHz divided by the channel separation. With a channel separate of 1.8   MHz, N equals 9.
2. • Each channel should be visited at least once every four frequency hopping cycles.

A hybrid system operating according to FCC regulations, i.e. LTE-M-U devices, shall in addition:

1. • Frequency hop over at least 15 channels.
2. • Follow a pseudo random frequency hopping pattern.
3. • Visit each channel at most 400   ms during N ⋅ 400 m s , with N defined as the number of channels in the hopping set.

Adaptive FHSS equipment operating according to ETSI regulations, i.e. LTE-M-U base stations, shall meet the following conditions:

1. • The base station should be able of frequency hopping across 70% of the 2.4   GHz frequency band.
2. • The accumulated transmit time should not exceed 400   ms during N ⋅ 400 m s , with N defined as the number of channels in the frequency hopping set.
3. • Each channel should be visited at least once every four frequency hopping cycles.

These ETSI and FCC requirements are met by LTE-M-U transmissions which hops over 16 or 32 frequencies, located on a 1.8   MHz frequency raster and spread across the band according to the selected frequency allocation. The hopping follows a pseudo random pattern that guarantees an equal visitation of each channel. It takes the physical-cell identity and mframe number as input and generates a cell-specific hopping sequence that repeats itself after every eighth hyperframe.

Fig. 15.4 illustrates the LTE-M-U frequency hopping including the anchor segment periodicity, which equals 80 or 160   ms when hopping over 16 or 32 data channels, respectively. The anchor segment is described in detail in Section 15.2.2.3.1.

15.2.2.3. Downlink physical channels and signals

LTE-M-U supports the basic set of logical and transport channels defined for LTE-M in 3GPP Release 13, which are depicted in Section 5.2.4. In the downlink the Primary and Secondary Synchronization Signals (PSS and SSS), the Cell-Specific and Demodulation Reference Signals (CRS and DMRS), the MTC Physical Downlink Control Channel (MPDCCH) and the PDSCH are all supported. In addition, LTE-M-U supports the Presence Detection Reference Signal (PDRS) for indicating that a basestation has successfully performed a CCA in the beginning of the mframe data segment.

Fig. 15.4 LTE-M-U frequency hopping.

The main change compared to LTE-M design is the introduction of the CCA procedure. It has an influence on all the downlink transmissions. Other noteworthy changes are the introduction of the PDRS and new mappings of the LTE-M-U PSS, SSS and PBCH on the resource grid. The LTE-M-U PDSCH and MPDCCH design is close to identical to the LTE-M design.

In LTE-M a bitmap is broadcasted for indicating which subframes are valid for LTE-M downlink transmissions (see Section 5.2.4.1). In LTE-M-U all downlink subframes being part of the uplink-downlink configuration are available for downlink operation.

To distinguish between the LTE-M physical channels and signals and the corresponding LTE-M-U counterparts, a prefix ‘u’ is from this point on added to the latter.

15.2.2.3.1. Synchronization signals

The LTE-M-U Primary and Secondary Synchronization Signal (uPSS, uSSS) transmissions start after an RF retuning from the frequency hopping data segment and a CCA have been performed. The CCA, introduced in Section 15.2.2.2.3, is performed to sense that the channel is free and available for transmission. As the anchor channel occupancy time is at most ∼5   ms, it is sufficient to perform the CCA over the minimum required time of 18   μs.

A total budget of two OS is reserved for the combination of RF retuning and CCA at the beginning of the anchor channel segment. In case the CCA is successfully performed, the synchronization signals are transmitted starting from OFDM symbol 2 in the first subframe n   =   0. This is depicted in Fig. 15.5 which shows that each of uPSS and uSSS is repeated 4 times during the first two subframes.

Fig. 15.5 uPSS and uSSS mapping on the center 62 subcarriers, excluding the DC subcarrier, of the anchor channel resource grid in subframe n   =   0 or 1 depending on the CCA procedure.

If the CCA fails, the base station must perform an eCCA. For a 5-ms channel occupancy the eCCA should be performed for a random time drawn between 18 and 250   μs. If the eNB is able to perform RF retuning, the CCA and the eCCA within the reserved 2 OS then it can start the synchronization signal transmission according to Fig. 15.5 with the subframe number n   =   0. If the time needed exceeds 2 OS the transmission is postponed 1   ms and may start first from OFDM symbol 2 in subframe n   =   1. In this case, the channel occupancy time is ∼4   ms which results in a required eCCA period of up to 200   μs.

In case neither the CCA or the eCCA is successful then the anchor channel transmission is canceled.

The uPSS and uSSS transmissions in the anchor segment are densified compared to LTE-M. This is to compensate for the infrequent anchor segment transmission. When the data segment is frequency hopping over 16 data channels the anchor segment is transmitted every 80   ms on the anchor frequency in the beginning of every mframe. With frequency hopping performed over 32 data channels only every other anchor segment is transmitted. This is intended to decrease the anchor channel occupancy and reduce anchor channel interference levels across cells. The periodicity then equals 160   ms as shown in Fig. 15.4. The densification of the uPSS and uNSS increases the likelihood for a device to detect and synchronize to the anchor channel segment. The increased detection probability will to some extent compensate for the long anchor channel periodicity.

15.2.2.3.2. uPSS

Relative subframe	0, 1 or 1, 2 depending on the (e)CCA
Periodicity	80   ms, 160   ms
Subcarrier spacing	15   kHz
Bandwidth	62 subcarriers (not including the DC subcarrier)
Frequency location	At the center of the LTE-M-U narrowband

The uPSS fills the same purpose as the LTE-M PSS, that is to provide time and frequency synchronization and part of the Physical Cell Identity (PCID). LTE-M-U supports, just as LTE-M, 504 PCIDs divided into 167 physical-layer cell-identity groups N ₁ , each containing 3 cell identities N ₂ . The uPSS waveform is identical to the PSS and is defined by a 63-element Zadoff-Chu sequence with root indices u equal to 25, 29 or 34.

p u ( n ) = e − j u π n ( n + 1 ) 63 , n = 0,1 , … , 62

(15.2)

The sequence is mapped to the 63 subcarriers located around the center in LTE-M-U narrowband. The center subcarrier itself, which is commonly associated with a DC component in the receiver, is punctured, and not transmitted. This means that the uPSS waveform in practice is defined by the Zadoff-Chu elements 0–30 and 32–62.

Each of the 3 roots is associated with one of the 3 cell identities N ₂ , defined within a physical-layer cell-identity group. When detecting the uPSS a device determines the root index and the corresponding cell identity.

The Zadoff-Zhu sequence provides both good auto-correlation properties and peak-to-average power ratio (PAPR) characteristics. In the time domain the uPSS sequence is first mapped on OFDM symbol (OS) 2 and repeated on OS 6 in the same subframe, and then on OS 2 and 5 in the next subframe. Fig. 15.5 shows this asymmetric mapping of the uPSS, and the uSSS, onto the resource grid across subframes 0 and 1. This allows a device synchronizing to the downlink frame structure to distinguish between the first and second subframes and obtain radio frame timing.

15.2.2.3.3. uSSS

Relative subframe	0, 1 or 1, 2 depending on (e)CCA
Periodicity	80   ms, 160   ms
Subcarrier spacing	15   kHz
Bandwidth	62 subcarriers (not including the DC subcarrier)
Frequency location	At the center of the LTE-M-U narrowband

The uSSS shares definition with the LTE SSS mapped to subframe 0. It is defined by the 62 symbol sequence p(n) which is determined by the interleaving of 2 length-31   m-sequences s 0 ( m 0 ) and s 1 ( m 1 ) , which in turn are scrambled by m-sequences z 1 ( m 0 ) and c 0 :

p ( 2 n ) = s 0 ( m 0 ) ( n ) c 0 ( n ) p ( 2 n + 1 ) = s 1 ( m 1 ) ( n ) c 0 ( n ) z 1 ( m 0 ) ( n ) n = 0,1 , … , 30

(15.3)

The sequence s 0 ( m 0 ) , s 1 ( m 1 ) and z 1 ( m 0 ) are all generated by cyclic shifts m ₀ and m ₁ of m-sequences s(n) and z(n), respectively. 167 unique pairs of shifts (m ₀ , m ₁ ) are defined, where each pair is associated to one of the 167 physical-layer cell-identity groups N ₁ .

c 0 ( n ) is generated by a cyclic shift of the m-sequence c(n), with the length of the cyclic shift dependent on the cell identity N ₂ . As the cell identity N ₂ is signaled by the uPSS, this creates a coupling between the uPSS and uSSS transmissions from the same cell.

The m-sequence is a suitable choice for providing physical-layer cell identity signaling by means of sequence detection thanks to its good auto-correlation properties. It is also straightforward to generate the m-sequences by means of shift registers. An interesting property of the m-sequence is that a cyclic shift of the sequence results in a new m-sequence.

The uSSS is just as the uPSS mapped in the frequency domain on the center 62 sub-carriers, excluding the DC subcarrier. In the time domain the signal is mapped on OS 3 adjacent to the first uPSS. It is repeated on OS 5 in the same subframe, and on OS 3 and 6 in the next subframe.

LTE-M concatenates the sequences s 0 ( m 0 ) and s 1 ( m 1 ) in two alternative fashions to create two versions of the SSS. The two versions are then mapped on different subframes, e.g. 0 and 5 for frequency-division duplex (FDD) LTE, to allow a device to obtain frame timing. LTE-M-U defines a single uSSS version that is repeated on all four OS. An LTE-M-U device therefore needs to use the uPSS and uSSS mapping to determine if it has detected the first or second subframe. It also needs to detect up to two uPSS and uSSS OS pairs to determine the OS timing within the subframe.

15.2.2.3.4. Downlink reference signals

In the downlink LTE-M-U supports the Cell-Specific Reference Signal (CRS), the Device-Specific Reference Signal and the PDRS. The CRS follows the LTE design and supports similar functions as LTE, i.e. channel estimation and frequency tracking for coherent uPDSCH demodulation and radio link monitoring in connected and idle mode. The CRS supports uPDSCH demodulation for transmission modes 1, 2 and 6. For the sake of radio link monitoring a device can assume that the CRS is present in case the channel sensing is successful which is indicated by the transmission of the uPSS, uSSS and uPBCH on the anchor channel and the PDRS on the data channel.

The Device-Specific Reference Signal, also known as the Demodulation Reference Signal (DMRS), supports demodulation of the uMPDCCH, and the uPDSCH in case of transmission mode 9. The DMRS is transmitted on the same antenna ports as the uMPDCCH and uPDSCH to support e.g. device-specific beamforming.

More information on the downlink reference signals can be found in Section 5.2.4.3. Next, we describe the PDRS which is unique to LTE-M-U.

15.2.2.3.4.1. PDRS

Relative subframe	5 for CCA or 6, 7 or 8 for eCCA
Basic TTI	1   ms
Repetitions	1
Subcarrier spacing	15   kHz
Bandwidth	72 subcarriers (not including the DC subcarrier)
Frequency location	Across the LTE-M-U narrowband

A base station is required to perform the CCA procedure (see Section 15.2.2.2.3) before initiating a downlink transmission on the mframe data segment. Use of the downlink parts in the applicable uplink-downlink configuration is only permitted if the channel is clear. The PDRS serves as an indication to the devices in a cell that the channel is clear and that CRS, uPDCCH and uPDSCH transmissions can be expected. Its functionality can be compared to that offered by the Wake-Up Signal specified for LTE-M in Release 15 (see Section 5.2.4.5).

The channel occupancy time for a downlink data transmission is defined from the first downlink subframe to the last downlink subframe in a data segment mapped on the mframe. This results in a channel occupancy time in the range of 20–60   ms depending on the uplink-downlink configuration. The corresponding CCA must be performed for a duration between 40 and 120   μs, i.e. less than 2 OS. The eCCA needs to be performed during in-between 1 and 3   ms.

The PDRS is a defined by a cell-specific pseudo-random bit sequence, generated just as the CRS, that is modulated into QPSK symbols. The QPSK symbols are mapped over a full narrowband on the last 12 OS of a subframe. Fig. 15.6 for simplicity illustrates the mapping on 1 of the 6 PRBs for the case of 2 base station antenna ports.

The first two OS on subframe n are left unused to support the CCA procedure. In case the channel is clear the PDRS transmission may start in subframe 5, which corresponds to the first subframe in the data segment. If the eCCA needs to be performed for k   =   1, 2 or 3 subframes, the PDRS transmission starts first in subframe n   =   5   +   k. This means that the PDRS may be transmitted in subframes n   =   5, 6, 7, or 8, and the number of valid subframes for the transmission of downlink physical channels will be reduced accordingly.

Fig. 15.6 PDRS resource element mapping illustrated for 2 antenna ports.

15.2.2.3.5. uPBCH

Relative subframe	0, 1 or 1, 2 depending on (e)CCA
Basic TTI	2   ms
Periodicity	80   ms, 160   ms
Repetitions	3
Subcarrier spacing	15   kHz
Bandwidth	62 subcarriers (not including the DC subcarrier)
Frequency location	At the center of the LTE-M-U narrowband

The uPBCH design is inspired by the LTE PBCH. It carries the Master Information Block (MIB) that together with the System Information Block A (see Section 15.2.3.1) contains the most vital parts of the system configuration information. The 24-bit MIB is convolutionally encoded and rate matched before being repeated three times on the center 72 subcarriers, excluding the DC subcarrier, on the available OS in subframes n and n+1 where n   =   0 or 1 depending on the CCA procedure (see Section 15.2.2.2.3). This is depicted in Fig. 15.7, where the CRS mapping for simplicity is not depicted. Each repetition covers six OS:

1. • The first repetition is mapped on OS 4, 7, 8, 9, 10, 11 in subframe n. This differs from the mapping of the LTE-M core part and instead follows the PBCH mapping specified in MulteFire Release 1.0.
2. • The second repetition is mapped on OS 12, 13 in subframe n, and on OS 0, 1, 4, 7 in subframe n+1.
3. • The third repetition is mapped on OS 8, 9, 10, 11, 12 and 13 in subframe n+1.

Fig. 15.7 uPBCH mapping on the anchor channel resource grid with subframe n   =   0 or 1 depending on the CCA procedure. CRS is not depicted.

The in total 18 uPBCH OS results in a similar code rate as the LTE-M PBCH core part which is mapped over 16 OS in 40   ms. For LTE-M an optional configuration of 4 repetitions of the core part is available for system bandwidths larger than 1.4   MHz. LTE-M-U does not offer any configurability in this aspect and the uPBCH is always mapped over 3·6   =   18 OS.

15.2.2.3.6. uPDSCH

Relative subframe	2 - 4 for SIB-A. Any between 6 and 65 for other transmissions.
Basic TTI	1   ms
Repetitions	1, 2, 4, 8, 16, 32
Subcarrier spacing	15   kHz
Bandwidth	1 - 6 PRBs
Frequency location	Within the narrowband

The uPDSCH supports unicast data and system information transmissions, including SIB-A and SIB1-BR (known as SIB1-BR-MF in the MFA specifications). The uPDSCH design closely follows the Release 13 LTE-M PDSCH CE mode A design. It supports up to 32 repetitions using 4 redundancy versions. The maximum transport block size equals 1000 bits. Transmission modes 1, 2, 6 and 9 are supported. More information about the PDSCH physical layer definition is found in Section 5.2.4.7.

One noticeable restriction imposed upon the uPDSCH is that a transmission must be confined within the subframes of a single mframe. It is, due to the CCA procedure performed in the start of each data segment, not allowed to start a downlink transmission in a first mframe n and then follow the frequency hopping pattern and continue the transmission in a second mframe n+1. This not only impacts the scheduling flexibility but also limits the downlink frequency hopping gain.

The uPDSCH SIB-A is a new message but is similar to the MIB in that it contains critical system information parameters. It is described in Section 15.2.3.1 together with a SIB1-BR and the other system information blocks. It can already here be noted that SIB-A is mapped on the last subframes of the anchor segment, while SIB1-BR occupies the first subframe after the PDRS transmission.

15.2.2.3.7. uMPDCCH

Relative subframe	Any between 7 and 65
Basic TTI	1   ms
Repetitions	1, 2, 4, 8, 16
Subcarrier spacing	15   kHz
Bandwidth	2, 4, 6 PRBs
Frequency location	Within the narrowband

The uMPDCCH supports scheduling of the uPDSCH, uPUSCH and the uPUCCH transmissions. Its physical layer definitions are very similar to the LTE-M MPDCCH described in Section 5.2.4.6. In the frequency domain the MPDCCH can be transmitted over 2, 4 or 6 PRBs using an aggregation level up to 24. The maximum supported number of repetitions have been limited to 16 to support similar coverage as CE mode A, but not CE mode B. This gives a transmission time that spans between 1 and 16 subframes.

The uMPDCCH transmission can occupy several downlink portions of an uplink-downlink configuration, but it must due to the CCA be confined within a single mframe. This is the same restriction as applies to the uPDSCH.

The scheduling of the uPDSCH and uPUSCH is controlled by means of Downlink Control Information (DCI) messages. LTE-M-U supports transmission of DCI formats 6-0A and 6-1A for operation in CE mode A. This is in further detail discussed in Section 5.3.2.1.

The resource used for uPUCCH, providing the HARQ acknowledgment to a uPDSCH transmission, is determined based on the lowest enhanced control channel element (ECCE) that is part of the uMPDCCH transmission scheduling the uPDSCH. The channelization and resource identification of the uPUCCH is discussed in Section 15.2.2.4.4.

It should be noted that the ECCE numbering is not device-specific. The MPDCCH DCI therefore signals a HARQ ACK/NACK resource offset indication that is combined with the lowest ECCE number to make the uPUCCH resource mapping device-specific. This mechanism is intended to support multiplexing of uPUCCH transmissions from multiple devices.

15.2.2.4. Uplink physical channels and signals

LTE-M-U supports the same set of physical channels and signals as LTE-M Release 13 with close to identical design as used by LTE-M. The LTE-M-U uplink transmission does not rely on the CCA procedure, so the differences compared to LTE-M are in short limited to the mappings on the physical resources, the information content of the transmitted messages and that the CE operation is limited to CE mode A.

In LTE-M a bitmap is broadcasted for indicating which subframes are valid for uplink transmissions (see Section 5.2.5.1). In LTE-M-U all uplink subframes being part of the uplink-downlink configuration are available for transmissions.

15.2.2.4.1. uPRACH

Relative subframe	According to higher layer configuration
Basic TTI	1   ms
Repetitions	1, 2
Subcarrier spacing	1.25   kHz
Bandwidth	839 subcarriers
Frequency location	Within a narrowband

The uPRACH is identical to the LTE PRACH (see Section 5.2.5.2). Its defined by a length 839 Zadoff-Chu sequence, which is used to generate 64 unique preamble sequences per cell, based one or more configured root indexes complemented by a set of cyclic shifts.

A difference compared to LTE-M is that only PRACH Format 0 is supported for LTE-M-U. Format 0 is designed with a cyclic prefix length of 103 μs while the Zadoff-Chu sequence corresponds to a length of 800   μs after the DFT-precoded OFDM modulation. When mapped to a single subframe a guard period of 97 μs is obtained, which corresponds to a maximum supported cell size of ∼15   km if collisions with transmissions in the next subframe are to be avoided.

Another difference compared to LTE-M is that only a repetition factor of 1 or 2 can be configured for the uPRACH. This was deemed sufficient for the coverage range needed to support industrial small cell deployments. The uPRACH transmissions are mapped onto the available uplink subframes according to 1 out of 29 specified configurations. The applicable mapping is indicated to the devices in a cell by means of system information broadcast signaling.

15.2.2.4.2. Uplink reference signals

LTE-M-U supports the uplink Demodulation Reference Signal (DMRS) and the Sounding Reference Signal (SRS). Their purpose and design follow LTE-M described in Section 5.2.5.3 with some small deviations. Since the uPUSCH transmission bandwidth is restricted to one narrowband, so is the DMRS bandwidth for the support of coherent uPUSCH demodulation. DMRS for uPUCCH formats 1, 1a, 2 and 2a are supported, just as for LTE-M operating as a FDD system. uPUCCH format 3 was added to the LTE-M-U design baseline to support ACK/NACK of up to 10 HARQ processes. For uPUCCH Format 3 the DMRS is mapped on symbols 1 and 5 in each slot as depicted in Fig. 15.8.

Fig. 15.8 DMRS symbol mapping for uPUCCH Format 3.

The SRS is mapped over the center four PRBs in subframes fulfilling n s f r e l m o d T S F C ∈ Δ S F C in mframes satisfying n m f r a m e m o d T m f r a m e = 0 . n s f r e l is the relative subframe numbering, T S F C is the SRS periodicity within the applicable mframes, Δ S F C defines the subframe for SRS transmissions within the period T S F C . T m f r a m e is finally the SRS mframe periodicity. 32 different SRS configurations are supported. In its most dense configuration the SRS is transmitted once every five uplink subframes. In the least dense configuration two SRS subframes are transmitted every four mframes.

15.2.2.4.3. uPUSCH

Relative subframe	Any between 21 and 65
Basic TTI	1   ms
Repetitions	1, 2, 4, 8, 16, 32
Subcarrier spacing	15   kHz
Bandwidth	1 – 6 PRBs
Frequency location	Within the narrowband

The uPUSCH supports unicast data transmission and aperiodic channel state information reporting. Its design largely follows the LTE-M design described in Section 5.2.5.4. The transmission bandwidth is limited to at most 6 PRBs. The number of repetitions is limited to 32 as only CE mode A is supported.

The LTE-M-U uplink is designed as a non-adaptive frequency hopping system to meet the ETSI regulations. This means that the uplink transmissions need to meet the MU requirement described in Section 15.2.3.4. In addition the uPUSCH may not be transmitted for more than 5 consecutive subframes before resting at least 5   ms, and in total not transmit for more than 15   ms during any given data segment. To facilitate this transmission scheme the uplink part of the data segment is divided in 2 subframe sets 0 and 1. Fig. 15.9 illustrates the concept for uplink-downlink configuration 1. A device may only make use of one of the subframe sets for its uPUSCH transmissions. A uPUSCH transmission of a length exceeding 5   ms, starting in the nth subframe set 0 will be postponed during the nth subframe set 1 before it continues in the next subframe set 0.

Fig. 15.9 Uplink-downlink configuration 1 with the uplink part portioned in subframe sets 0 and 1.

While the downlink transmissions need to be confined within a single mframe due to the channel sensing regulation the uplink transmissions may continue across mframes. This since the uplink operates as a non-adaptive frequency hopping system. This may give repeated uplink transmissions a frequency hopping gain, which is not present for the downlink transmissions.

15.2.2.4.4. uPUCCH

Relative subframe	Any between 21 and 65
Basic TTI	1   ms
Repetitions	1, 2, 4, 8
Subcarrier spacing	15   kHz
Bandwidth	1 PRB
Frequency location	Within the narrowband

The uPUCCH follows the LTE-M PUCCH design described in Section 5.2.5.5. LTE-M supports PUCCH Formats 1, 1a, 2 and 2a for FDD and Formats 1b and 2b for TDD. LTE-M-U uses LTE-M FDD from 3GPP Release 13 as its design baseline and implements support for PUCCH Formats 1, 1a, 2 and 2a. It also supports the LTE PUCCH Format 3 to enable HARQ ACK/NACK for up to the 10 supported HARQ processes to be transmitted in a single uPUCCH message.

The uPUCCH is mapped to uplink subframes according to the uplink-downlink configuration and the uplink subframe set concept. In the frequency domain it is mapped to the outermost PRBs of the narrowband. While the LTE PUCCH is frequency hopping across the system bandwidth the uPUCCH is transmitted only on the lower edge of the narrowband. To support efficient multiplexing of the uPUCCH and uPUSCH on a single narrowband it is important to minimize the frequency resources used by the different PUCCH formats.

PUCCH Format 1 and 1a are based on a set of 30 base sequences r ¯ u ( k ) . These are the same sequences that defines the single-PRB DMRS transmissions. Each base sequence is uniquely defined by a set of tabulated phases φ u ( k ) :

r ¯ u ( k ) = e j φ u ( k ) π 4 , k = 0 , . , 11 , u ∈ { 0 , . , 29 }

(15.4)

When a single HARQ ACK/NACK bit d is transmitted it is mapped to a complex-valued symbol d ˜ which is modulated by a α_m phase rotated version of r ¯ u ( k ) and multiplied with a block spreading sequence w n o c . This generates the set of complex symbols z ( k , l ) mapped to resource elements ( k , l ) in a PUCCH Format 1a resource block:

z ( k , l ) = w n o c ( l ) · e j α m k · r ¯ u ( k ) · d ˜ , k = 0 , .. , 11 , l = 0 , 1 , 5 , 6 , m ∈ { 0 , .. , 11 }

(15.5)

Note that index k identifies the subcarrier, while index l identifies the OFDM symbol associated to a certain resource element. A DMRS sequence, supporting coherent demodulation of the uPUCCH, is in similarity to Eq. (15.5) phase rotated by α_m and spread over symbols l   =   2, 3, 4. The 12 available phase shifts combined with the length-3 block spreading sequence applied on the DMRS allows in theory up to 3 · 12 = 36 different PUCCH transmissions to be multiplexed on a single uplink resource block. As a rule of thumb only every second phase shift is useful if orthogonality between users is to be kept under time dispersive channel conditions. This reduces the PUCCH Format 1 multiplexing capacity to 3 ⋅ 6 = 18 users per resource block.

PUCCH Format 2 contains up to 11 bits per subframe. Each of the first 10 bits is modulated by a α_m phase rotated version of r ¯ u ( k ) and is mapped to the resource elements of data symbols 0, 2, 3, 4 and 6 on each of the two slots in the subframe. The use of the same set of base sequences r ¯ u ( k ) allows PUCCH Format 1 and 2 to be multiplexed on the same PRB using unique subsets of the 12 phase shifts α_m, which gives orthogonality between the multiplexed transmissions.

PUCCH Format 3 follows a design different compared to PUCCH Formats 1 and 2. It carries 22 information bits that are block coded into 48 bits mapped to 2 sets of 12 QPSK symbols. Each set of 12 QPSK symbols is multiplied with a length-5 block spreading sequence and mapped to the 12 resource elements of each of the 5 data symbols in each slot. As a result, PUCCH Format 3 cannot be multiplexed on the same resource blocks as the PUCCH Format 1 and 2 transmissions. Fig. 15.10 illustrates an example where PUCCH Formats 1 and 2 are multiplexed on PRB 0, and PUCCH Format 3 is mapped on PRB 1. This leaves at least 4 PRBs for PUSCH transmission. The uPUCCH resource blocks may be used by uPUSCH transmissions when no uPUCCH is scheduled for transmission.

15.2.3. Idle and connected mode procedures

In this section we describe the LTE-M-U idle and connected mode procedures. The design principle of LTE-M-U was to maximize the reuse of the LTE-M procedures described in Section 5.3. Here we therefore focus on the procedures that deviate from the LTE-M baseline, including new procedures required by ETSI and FCC regulations.

Fig. 15.10 uPUCCH and uPUSCH multiplexing on a narrowband.

15.2.3.1. Cell selection and system information acquisition

The first task a device performs when turned on is to search for, evaluate the suitability of, and select a cell. While an LTE-M device scans all its supported frequency bands in the search for a suitable cell, an LTE-M-U device only scans the 3 anchor channels in the 2.4   GHz frequency band, depicted in Fig. 15.1, in its search for an LTE-M-U cell. The rationale for limiting the set of anchor channels is partly due to the frequency hopping nature of the LTE-M-U system, which leads to a sparse transmission of the anchor channel including its synchronization and MIB blocks. If multiple attempts to synchronize are needed the system acquisition time quickly grows as the uPSS is transmitted only every 80   ms or even only every 160   ms. With only 3 potential anchor channels this is no longer an issue of significance.

After acquiring synchronization to the downlink frame structure and the physical-cell identity via the uPSS and uSSS, the device reads the uPBCH MIB and the uPDSCH SIB-A. The MIB contains the system frame number information, which gives the mframe index within a hyper frame, the SIB-A and the SIB1-BR scheduling information.

15.2.3.1.1. SIB-A

SIB-A is a new system information message, which is similar to the MIB in that it contains critical system information parameters. SIB-A is mapped on the last 2 or 3 subframes on the anchor channel. If the channel is cleared during the first 2 OS the uPBCH transmission takes place during subframes 0 and 1 and the SIB-A is transmitted in subframes 2, 3 and 4. If the enhanced CCA needs to be performed during subframe 0, and is successful, the uPBCH is transmitted on subframes 1, 2 and the SIB-A is only mapped over subframes 3 and 4. Fig. 15.11 illustrates the two possible SIB-A configurations.

In total, SIB-A carries 28 bits including for example the Channel Set information element that signals the channel groups that defines the set of data channels used to frequency hop over. For 16 data channels, 4 channel groups will be selected and for 32 data channels, 8 channel groups will be selected. This signaling allows LTE-M-U to configure a set of 16 or 32 data channels that are exposed to as little interference as possible by other systems. The SIB-A information elements (IE) are summarized in Table 15.3. The Paging Indication IE allows devices to skip paging monitoring to save power. The System Information Value Tag IE has been moved to SIB-A to enable early detection of the need to acquire a fresh copy of the other system information messages.

Fig. 15.11 SIB-A mapping on the anchor channel resource grid.

15.2.3.1.2. SIB1-BR

SIB1-BR informs the devices in a cell about the LTE-M-U access mode of operation. A MulteFire cell can operate either using the PLMN access mode or using the Neutral Host Network (NHN) access mode. In case of the PLMN access mode, LTE-M-U is connected to an Evolved Packet Core network in standalone mode or as an additional radio access network, e.g. complementing an LTE-M deployment. In the NHN access mode, LTE-M-U operates as a standalone system providing local access for a set of service providers. In PLMN mode, the PLMN selection is based on the established 3GPP procedures. In case of NHN, the NHN identifier identifies the network, while the set of service providers utilizing the network are identified by a Participating Service Provider identity, and the PLMN information element broadcasted in SIB1-BR is set to a single reserved PLMN identity which is common for all MulteFire NHNs.

Table 15.3

SIB-A information elements.

Parameter	Description
Channel set	Identifies the set of 16 or 32 data channels configured for data transmission.
Hyper system frame number	The three least significant bits of the hyper SFN. This gives, combined with the master information frame number information, the complete mframe index.
mframe configuration	Indicates the configured uplink-downlink configuration.
Paging indication	Indicates if there will be a paging during the next data segment.
PDRS window size	Indicates the subframe window in which the PDRS may be transmitted if the channel is cleared.
System information value tag	Informs the devices if a change in the system information has occurred.

The LTE-M-U SIB1-BR is transmitted with a periodicity ranging between 80 and 640   ms using 8 or 4 repetitions. The SIB1-BR scheduling is presented in Table 15.4 under the assumption that the PDRS is transmitted in subframe 5. The first subframe in the data segment after the PDRS transmission always contains a SIB1-BR transmission.

The content and scheduling mechanism for the remaining SIB information elements follow the LTE-M specification as described in Section 5.3.1.2.

15.2.3.2. Paging

The LTE-M-U paging is based on the procedure described for LTE-M in Section 5.3.1.4. Both DRX and eDRX are supported. In LTE-M, the start of the common search space for the MPDCCH containing the P-RNTI is determined by a paging frame, and a subframe within the paging frame known as the paging occasion. In the extreme case of high paging loads up to 4 subframes in every radio frame can be configured as paging occasions. In LTE-M-U, a new concept called a Paging Occasion Window (POW) is defined. A POW spans 1 to 4 mframes. There is only one paging occasion per mframe, which starts at the first valid downlink subframe within the mframe, i.e. it starts at a subframe not containing system information. If a paging transmission in the first paging occasion fails, then the network may attempt to page the device in the next paging occasion, which is sent in the next mframe within the POW.

To reduce device power consumption from monitoring paging occasions for the case there is no page, LTE-M-U supports a paging indication bit sent in SIB-A. It informs the devices in a cell if they can expect a page or not during the next anchor segment period.

15.2.3.3. Power control

The LTE-M-U uplink power control follows the definitions and procedures specified for LTE-M CE mode A as described in Section 5.3.2.4. An important difference is the device power control accuracy. While 3GPP allows a band dependent maximum output power tolerance of 2–2.5   dB, the ETSI and FCC specifications impose strict limitations on the maximum output power. For ETSI, following the design of a non-adaptive FHSS system implies a maximum EIRP of 20 dBm. In case of FCC, the limit for a hybrid system is 8 dBm/3   kHz which equates to 25.8   dBm over 1 PRB. The first LTE-M-U device category is expected to support power class 5 operating at a max power of 20   dBm, due to the ETSI requirements.

Table 15.4

LTE-M-U SIB1-BR scheduling.

Period	Repetitions	SIB1-BR subframe index
80   ms	4	6, 7, 8, 9
80   ms	8	6, 7, 8, 9, 10, 11, 12, 13
160   ms	4	6, 7
320   ms	4	6
640   ms	4	6

The situation is similar for the downlink. 3GPP allows tolerances in the maximum basestation output power of at most 2.5   dB. ETSI has a strict limit of 20   dBm EIRP for adaptive FHSS systems, while FCC allows up to 30   dBm for a digitally modulated system. There is no acceptance for exceeding these maximum output power levels.

15.2.3.4. Medium utilization

Non-adaptive equipment operating according to the ETSI regulations, i.e. LTE-M-U devices, with an EIRP exceeding 10   dBm must comply with a medium utilization (MU) factor of at most 10%. The MU factor balances the power and time resources used by a certain transmitter and is defined as:

MU = P 100 mW · DC

(15.6)

Where P corresponds to the output EIRP, and the DC is the duty cycle. A device operating at the maximum output power of 100   mW (20   dBm) is limited by a duty cycle of 10%. The duty cycle allowance increases with decreasing output power.

15.3. NB-IoT-U

The development of EC-GSM-IoT, LTE-M and NB-IoT was partly triggered by the competition from various low power wide area network (LPWAN) technologies operating in unlicensed spectrum. 3GPP provides through EC-GSM-IoT, LTE-M and NB-IoT a highly competitive offering for close to all types of licensed spectrum arrangements. But for enterprises lacking access to licensed spectrum a LPWAN solution operating in free, unlicensed spectrum is still attractive. The MFA therefore developed NB-IoT-U to complement LTE-M-U and make a 3GPP based LPWAN available in unlicensed spectrum.

The next three sections introduce the design principles followed by MFA when developing NB-IoT-U, the NB-IoT-U physical layer functionality and the procedures used in RRC idle and connected mode.

15.3.1. Radio access design principles

The NB-IoT-U system design is in the first release targeting operation in the FCC ISM band defined by the frequency range 902–928   MHz and the ETSI SRD Band 54 defined by the 250   kHz band located between 869.4 and 869.65   MHz [10]. EU, through the European Commission, has recently decided to make Band 47b containing the frequency ranges 865.6–865.8, 866.2–866.4, 866.8–867 and 867.4–867.6   MHz available for LPWAN [11]. The MFA focused on ETSI band 54 during its work, and this is reflected in the descriptions in this book. Band 54 and 47b are subject to similar regulations, which is expected to make an extension from operation in Band 54 to Band 47b straightforward. These sub-GHz frequency bands are all attractive for low-power wide-area technologies due to their favorable radio propagation characteristics.

NB-IoT-U uses the NB-IoT stand-alone design as its baseline. This is a natural choice, since the system is intended to operate in non-3GPP spectrum. Operation within or in the guard-band of an LTE system is not foreseen in unlicensed spectrum. The intention of NB-IoT-U is to build on the NB-IoT eco-system, to allow a reuse of both device and network implementations. This motivates a reuse of the NB-IoT design to the maximum extent possible. This book distinguishes between the NB-IoT physical channels and signals and the corresponding NB-IoT-U counterparts, by a prefix ‘u’ that is added to the names of the latter.

NB-IoT-U is designed as a time division duplex type of system. Despite this, it is based on the 3GPP Release 13 NB-IoT FDD specifications, and not the 3GPP Release 15 specifications that supports NB-IoT in TDD operation. The explanation for this is that MFA Release 1.1 specifying NB-IoT-U and 3GPP Release 15 were developed in parallel.

NB-IoT-U shares similarities with LTE-M-U in that it follows a TDD design with a frame structure that is divided into a fixed anchor and frequency hopping data segments. The anchor channel is carrying the synchronization signals and the broadcast channel. The data channels support both uplink and downlink unicast control and data transmissions.

The next two sections introduce the FCC and ETSI regulations. Detailed descriptions of parts of the regulations is also provided later in the text. Section 15.3.2.2.2 presents the frequency hopping rules, while Section 15.3.3.2 describes the power control regulations.

15.3.1.1. FCC regulations

In US the NB-IoT-U anchor segment transmissions are following the FCC rules [2] for a digitally modulated system and is mapped on 3 PRBs. Synchronization and broadcast channels are located at the 2 outer PRBs to secure that the 500   kHz bandwidth requirement for digital modulation is always met. The anchor segment is transmitted on a cell specific frequency which allows devices to sequentially scan the supported frequency bands when performing cell selection. The base stations’ digitally modulated transmissions are not permitted to exceed an average conducted power of 30   dBm, and a PSD of 8 dBm/3   kHz. The maximum allowed average radiated power is defined by an EIRP of 36   dBm.

The data segment follows the frequency hopping regulations. It is modulated on a single PRB that frequency hops over 64 frequencies. The NB-IoT-U downlink is a good example of a hybrid system with the anchor part being digitally modulated while the data part is frequency hopping. Due to the required random location of the frequency hopping channels, the anchor segment does not qualify as a frequency hopping system. The frequency hopping PRB may be transmitted using a conducted peak output power of 30   dBm, and a maximum peak EIRP of 36   dBm. The frequency hopping regulations defines the power in terms of peak power, and not average power as in the case of the digital modulation regulations. This may seem like a minor detail but is an important aspect for systems such as NB-IoT-U with a non-zero PAPR. The PAPR of the transmitted OFDM waveform needs to be taken into account when determining the maximum allowed average output power.

The anchor part is reserved for downlink operation, meaning that the devices in the uplink are only permitted to operate over the frequency hopping data segments. The uplink is hence declared as a frequency hopping system.

15.3.1.2. ETSI regulations

In Europe NB-IoT-U is designed according to the regulations defined in EN 300 220 for Non-specific Short Range Devices (SRD) [10]. The standard is also applicable to LPWANs although its name may suggest that it does not cover LPWANs. The use of Band 54 allows a peak, not average, output power of 29.2 dBm EIRP, in combination with a 10% duty cycle.

The duty cycle rule is limiting the transmit activity of NB-IoT-U base stations and devices. The impact is most pronounced on the ability of the base stations to provide the needed downlink capacity for serving all devices in a cell. Due to the expected infrequent and small uplink data transmissions the impact on the devices is assumed to be lower.

Most ETSI frequency bands in the range 25   MHz to 1   GHz are associated with a duty cycle budget well below 10%. This makes Band 54 an attractive choice despite its limitations. Table 15.5 summarizes the NB-IoT-U design choices made to comply with the FCC and ETSI regulations.

15.3.2. Physical layer

The description of the physical layer is distributed in four parts. The Physical resources section 15.3.2.1 presents the basic time and frequency resources used transmitting the NB-IoT-U signal in the FCC ISM and ETSI SRD bands. The Transmission schemes section 15.3.2.2 presents generic aspects and functionality that applies across large parts of the physical layer design. The last two sections 15.3.2.3 and 15.3.2.4 present the details of the downlink and uplink physical channels and signals.

15.3.2.1. Physical resources

15.3.2.1.1. Channel raster

NB-IoT-U is in the first release targeting the FCC ISM band 902–928   MHz, and ETSI SRD Band 54 spanning the range 869.4–869.65   MHz. The specified channel raster is dependent on the channel bandwidth definition which differs between ETSI and FCC.

Table 15.5

NB-IoT-U radio access design principles.

Regulation	Link direction	Design basis
FCC	Downlink	Hybrid
	Uplink	FHSS
ETSI	Downlink	Non-specific Short Range
	Uplink	Non-specific Short Range

In ETSI the NB-IoT-U channel bandwidth is defined by the occupied bandwidth of the modulated waveform. Its definition aligns with 3GPPs occupied bandwidth concept and corresponds to the frequency range containing 99% of the total power of the studied signal. In 3GPP the NB-IoT occupied bandwidth equals 200   kHz which fits in the center of ETSI band 54.

In FCC the frequency range 902–928   MHz have been divided in 3 anchor channels and 64 data channels. The anchor channel baseband signal spans 3 PRBs while the data channel baseband waveform is defined by a single PRB. For the channel bandwidth, the FCC regulations define a 20-dB bandwidth requirement. This corresponds to the bandwidth where the signal emissions are attenuated by 20   dB relative to the power measured in the center of the carrier. The 20-dB bandwidth definition is more stringent than the 3GPP and ETSI occupied bandwidth requirement which motivates a wider channel bandwidth declaration in FCC compared to ETSI and 3GPP. The single PRB data channel 20   dB bandwidth was agreed to 345   kHz after careful study of the 3GPP spectrum emission mask. The intent of MFA was to make sure that the NB-IoT-U RF requirements are not more stringent than the 3GPP RF requirements. The 345   kHz channel bandwidth for simplicity also defines the data channel frequency raster.

15.3.2.1.2. Frame structure

NB-IoT-U introduces frame structure Type 3N. It is a TDD type of frame structure which is a natural choice since ETSI band 54 and FCC band 902–928   MHz are both unpaired bands. Frame structure Type 3N builds on the 3GPP frame structure with the addition of the new nframe concept. A nframe spans N _Frame   =   2, 4, or 8 radio frames. The nframe index wraps around every 4 hyperframes as illustrated in Fig. 15.12.

The nframe defines the lengths of the TDD uplink-downlink configurations presented in Section 15.3.2.2. The shorter nframes supports more frequent uplink to downlink switching, while the longer nframes are intended to lower the switching frequency to reduce its associated overhead.

Two versions of the frame structure Type 3N are specified. Frame structure Type 3N1 supports operation in the US according to the FCC regulations. In this frame structure the nframe length is fixed to 20   ms. Every fourth nframe contains the 20   ms anchor segment. The anchor segment concept is inherited from LTE-M-U, and supports uNPSS, uNSSS and uNPBCH transmissions. The anchor definition facilitates a frequency hopping design of the data channels. The nframe length determines the dwell time on each RF during the frequency hopping cycle. In Frame structure Type 3N1 the anchor is 3 PRB wide as described in Section 15.3.2.1.3. Besides this exception NB-IoT-U follows a single PRB system design. A device is expected to make use of a single PRB receiver, even on the 3 PRB anchor segment.

Frame structure Type 3N2 supports operation according to the ETSI regulations defined for Band 54. It has a 10% duty cycle restriction which motivates less frequent uplink-downlink switching for optimizing the system capacity. It is also beneficial to group the few available downlink subframes for the support of cross subframe channel estimation, described in Section 8.2.3.2. This further motivates a reduced uplink-downlink switching rate. A reduced overhead from uNPSS, uNSSS and uNPBCH transmissions is also attractive. This was achieved by extending the anchor periodicity to 1280   ms.

Fig. 15.12 NB-IoT-U frame structure Type 3N including the nframe concept.

Which of the two frame structures that is configured is implicitly indicated by the NB-IoT-U operating band, or more exactly by the frequency location of the anchor carrier configured in a cell.

The subframes in a nframe are numbered between 0 and 10·N _frame -1 for N _frame   =   2, 4 or 8, and are referred to as the relative subframe number. Fig. 15.13 illustrates the two frame structures.

15.3.2.1.3. Resource grid

The basic resource grid follows the same definitions as the LTE-M-U resource grid described in Section 15.2.2.1.2. In ETSI NB-IoT-U follows a single PRB design. In FCC the anchor is digitally modulated and spans 3 PRBs during 20   ms, while the data segment is mapped on a single PRB.

15.3.2.2. Transmission schemes

15.3.2.2.1. Anchor and data segment

NB-IoT-U reuses the NB-IoT transmission schemes presented in Section 7.2.2 with one significant exception: While the NB-IoT TDD uplink-downlink switching period is associated to the 10   ms radio frame, NB-IoT-U uplink-downlink switching periods are defined by the nframe uplink-downlink configurations. Table 15.6 presents the uplink-downlink configurations used for data transmission in frame structure Type 3N1 intended for operation in the US.

Fig. 15.13 NB-IoT-U nframe and anchor periodicity.

In frame structure Type 3N1, 3 types of special subframes are specified. The first and second type supports downlink transmissions followed by a guard-period of either 4 or 7 OS. The third type, indicated as S_G in Table 15.6, is defined as a 1   ms guard-period in case of the first nframe that follows after an anchor segment. In case the nframe follows a nframe used for uplink data transmission then also this special subframe is used for uplink transmissions.

Table 15.6

Frame structure Type 3N1 uplink (U) - downlink (D) configurations, including special (S) subframes.

Configuration	Nframe length	Relative subframe number
		0	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19
0	2	D	D	D	D	D	S	U	U	U	U	U	U	U	U	U	U	U	U	U	U
1	2	D	D	D	D	D	D	D	S	U	U	U	U	U	U	U	U	U	U	U	U
2	2	D	D	D	D	D	D	D	D	D	S	U	U	U	U	U	U	U	U	U	U
3	2	D	D	D	D	D	D	D	D	D	D	D	S	U	U	U	U	U	U	U	U
4	2	SG	U	U	U	U	U	U	U	U	U	U	U	U	U	U	U	U	U	U	U
5	2	D	D	D	S	U	U	U	U	U	U	D	D	D	S	U	U	U	U	U	U
6	2	D	D	S	U	U	U	U	U	U	U	D	D	S	U	U	U	U	U	U	U
7	2	D	S	U	U	U	U	U	U	U	U	D	S	U	U	U	U	U	U	U	U

Table 15.7 shows the uplink-downlink configurations for frame structure Type 3N2 including nframe lengths of 2, 4 and 8 radio frames. Configurations 0, 2, 4 and 7 reserves 10% of the subframes for downlink transmissions to guarantee that the ETSI 10% duty cycle requirement is never violated by the NB-IoT-U base stations. Configurations 1, 3, 5 and 6 specifies 20, 30 or 40% of the subframes as downlink subframes. Since the duty cycle requirement is measured over a sliding measurement interval of 1   hour, these configurations allow the network to overprovision the downlink during short periods. The total number of transmitted downlink subframes should not exceed 360.000 in 1   h, but it is up to the network to decide when to spend the subframe budget. Downlink overprovisioning allows the network to meet short demands for increased system capacity or improved link level latency. In frame structure type 3N2 the special subframe S_G is dedicated to serve as guard-period between uplink and downlink dwells.

Table 15.8 summarizes the 3 special subframe types defined for NB-IoT-U, including their definitions in terms of downlink (DwPTS), uplink (UpPTS) and guard-period (GP) parts. The guard-period length sets an upper limit on the supported cell size. To avoid that the uplink transmissions of devices near the cell edge interfere the base stations downlink transmissions the cell radius should not exceed ∼42.5   km or ∼75   km when a guard-period of 4 or 7 OS, respectively, is configured. In the case of the full subframe being used as guard-period it is no longer the special subframe configuration that limits the cell size. Instead other factors, such as the uNPRACH format will limit the cell size.

Table 15.7

Frame structure Type 3N2 uplink (U) - downlink (D) configurations, including special (S_G) subframes.

Configuration	Nframe length	Relative subframe number
		0	1	2	3	4	5	6	7	8	9	10	11	12–15	16	17–19	20–39	40–79
0	8	D	D	D	D	D	D	D	D	SG	U	U	U	U	U	U	U	U
1	8	D	D	D	D	D	D	D	D	D	D	D	D	D	SG	U	U	U
2	4	D	D	D	D	SG	U	U	U	U	U	U	U	U	U	U	U	–
3	4	D	D	D	D	D	D	D	D	SG	U	U	U	U	U	U	U	–
4	2	D	D	SG	U	U	U	U	U	U	U	U	U	U	U	U	–	–
5	2	D	D	D	D	SG	U	U	U	U	U	U	U	U	U	U	–	–
6	2	D	D	D	D	D	D	D	D	SG	U	U	U	U	U	U	–	–
7	2	D	D	SG	U	U	U	U	U	U	U	D	D	U	U	U	–	–

Table 15.8

NB-IoT-U special subframe definitions in terms of OFDM symbols (OS).

Frame structure	Special subframe notation	DwPTS	GP	UpPTS	Cell radius supported by GP
Type 3N1	S	10 OS	4 OS	–	∼42.5   km
Type 3N1	S	7 OS	7 OS	–	∼75   km
Type 3N1, Type 3N2	SG	–	14 OS	–	∼150   km

15.3.2.2.2. Frequency hopping

NB-IoT-U frame structure Type 3N1 is required to follow the FCC frequency hopping regulations. The FCC regulations couples the permitted output power to the number of frequency hopping channels. To be able to transmit the data segments with a peak EIRP of 36 dBm the system is required to frequency hop over at least 50 channels. The hopping is required to be performed in a pseudo random pattern.

NB-IoT-U meets this requirement using a frequency hopping algorithm that is based on the LTE-M-U frequency hopping functionality described in Section 15.2.2.2.4. The frequency hopping algorithm takes the PCID and a counter, that counts nframes carrying data segments, as input and generates a cell specific hopping sequence across the full set of 64 supported frequency channels. This pseudo random hopping pattern repeats itself every 4 hyperframes.

The hopping across the full set of 64 data channels optimizes the frequency diversity. This approach can be compared with that taken for LTE-M-U that ideally makes use of the least interfered subset of specified RF channels at any given moment. In contrast to LTE-M-U, it is not possible for NB-IoT-U to select a subset to hop over as all specified frequency channels (excluding the three anchor channels) are included in a single large frequency hopping set.

Fig. 15.14 illustrates the frequency hopping applied for frame structure Type 3N1. It also illustrates the mapping of the different downlink and uplink physical channels and signals, which in detail is discussed in the next two sections.

15.3.2.3. Downlink physical channels and signals

This section describes the NB-IoT-U downlink physical channels and signals including the Narrowband Primary and Secondary Synchronization Signals (uNPSS, uNSSS), the Narrowband Physical Broadcast Channel (uNPBCH), the Narrowband Physical Downlink Shared Channel (uNPDSCH) and the Narrowband Physical Downlink Control Channel (uNPDCCH). NB-IoT-U strives to reuse the 3GPP Release 13 NB-IoT downlink design described in Section 7.2.4. The focus of this section is to identify and describe the few differences between the NB-IoT-U and the NB-IoT downlink designs.

Fig. 15.14 NB-IoT-U frame structure Type 3N1 frequency hopping.

In NB-IoT a bitmap signals the downlink subframes that are available for NB-IoT transmissions (see Section 7.2.4.1). For NB-IoT-U all subframes indicated as downlink subframes in the uplink-downlink configuration are available for downlink transmissions.

15.3.2.3.1. Synchronization signals

To support frequency hopping data transmission, it was decided to follow the LTE-M-U design, described in Section 15.2.2.3.1, and group the uNPSS, uNSSS and uNPBCH transmissions together in an anchor segment. The anchor periodicity is specified to 80   ms in FCC and 1280   ms in ETSI.

With such infrequent anchor segment transmissions, it is important that the design of the synchronization signals facilitates a single shot detection of the anchor segment synchronization signals followed by a quick acquisition of the uPBCH MIB. This was accomplished by a densification of the synchronization signals and the uNPBCH transmissions within the anchor segment. In this new design the uNPSS is mapped over 8 consecutive subframes, followed by 2 uNSSS subframes and 10 uPBCH subframes.

Fig. 15.15 illustrates the mapping of the uNPSS and uNSSS for frame structure Type 3N1 and 3N2. In case of frame structure Type 3N1 used in FCC the synchronization signals are mapped on the first of the 3 PRBs.

Fig. 15.15 NB-IoT-U synchronization signal mapping on the anchor segment.

15.3.2.3.2. uNPSS

Relative subframe	0–7
Periodicity	80   ms, 1280   ms
Subcarrier spacing	15   kHz
Bandwidth	165   kHz
Frequency location	PRB 0

The uNPSS serves the same purpose as the NPSS and supports cell detection and synchronization to the downlink frame structure in time and frequency. The basic waveform is defined by the same Zadoff-Zhu base sequence that defines the NPSS waveform (see Section 7.2.4.2.1):

p ( n ) = e − j 5 π n ( n + 1 ) 11 , n = 0,1 , … , 10

(15.7)

In the frequency domain p(n) is mapped over the first 11 sub-carriers of PRB 0, leaving the 12th sub-carrier empty. The NPSS is transmitted in subframe 5 in every radio frame. To leave room for the LTE control region, in case of inband mode of operation, its base sequence p(n) is mapped on the last 11 OS in a subframe. LTE inband operation in unlicensed spectrum is not an expected use case and NB-IoT-U is only designed for stand-alone operation. Consequently, the uNPSS base sequence is mapped over all 14 OS on each of the first 8 subframes in the nframe carrying the anchor segment. A computer-generated cover code of length 8·14   =   112 is applied on top of the 112 repeated copies of the base sequence.

The device can assume that the base station applies the same mapping of the signal to the antenna ports over the first 4 subframes, and over the last 4 subframes. This allows a base station transmitter configured with multiple antenna ports to switch mapping in the middle of the NPSS transmission which introduces spatial diversity in the transmission. This is especially advantageous for devices suffering from limited fading diversity.

The device receiver is expected to detect the uNPSS by performing a sliding auto-correlation across the repeated set of base sequences, while correcting for the cover code. The long cover code supports a high receiver processing gain which facilitates efficient detection of the uNPSS at low SINR levels.

15.3.2.3.3. uNSSS

Relative subframe	8, 9
Periodicity	80   ms, 1280   ms
Subcarrier spacing	15   kHz
Bandwidth	180   kHz
Frequency location	PRB 0

The uNSSS signals the PCID of a cell. 504 PCIDs are, just as for NB-IoT, supported. The uNSSS is as the uNPSS mapped across all 14 symbols of a subframe. The uNSSS spans all 12 sub-carriers of the PRB which means that it in total is mapped over 12·14   =   168 resource elements.

The length 168 uNPSS sequence s(n) is generated based on a length 167 Zadoff-Chu sequence, that is multiplied with a length 160 Hadamard sequence b q ( n ″ ) and rotated by a phase θ l :

s ( n ) = b q ( n ″ ) · e − j 2 π θ l · e − j u π n ′ ( n ′ + 1 ) 167 n ′ = n m o d 167 n ″ = n m o d 160 n = 0,1 , … , 167

(15.9)

This design is close to identical to the NSSS design. The only differences are that the Zadoff-Chu and Hadamard sequences have been extended in length to support the mapping across all 168 resource elements of a subframe.

The uNSSS is, just as the NSSS, made cell specific thanks to the configured Zadoff-Zhu root u, and through the selection of 1 out of 4 available scrambling sequences b _q (n) determined as:

u = ( P C I D m o d 126 ) + 3

q = ⌊ P C I D 126 ⌋

For NB-IoT the phase shift θ l signals the timing of the NSSS transmission.

In case of NB-IoT-U this indication is no longer needed, so the uNSSS phase shift θ l has for simplicity been fixed to 42 168 .

To support robust performance the uNSSS sequence s(n) is repeated twice on relative subframes 8 and 9 in the nframe containing the anchor segment.

15.3.2.3.4. uNRS

Relative subframe	Any DL subframes according to the UL-DL configuration
Periodicity	10   ms
Basic TTI	1   ms
Subcarrier spacing	15   kHz
Bandwidth	180   kHz
Frequency location	Any

The uNRS supports coherent demodulation of the uNPBCH, uNPDCCH and uNPDSCH, and radio resource management measurements. Its design is identical to the NRS, described in Section 7.2.4.3. The only practical difference is where a device can assume to find the uNRS transmissions.

In case of frame structure Type 3N1 and before reading the narrowband version of system information block 1 (SIB1-NB) a device can assume that uNRSs are always present in PRB 0 and PRB 2 in the anchor segment. Relative subframes 10–19 in PRB 0, carries the uNPBCH including NRSs. Relative subframes 0–19 in PRB 2 continuously transmit a uNPDSCH, including uNRSs, carrying system information messages as described in Section 15.3.2.3.5. With the synchronization signals and the uNPBCH mapped to PRB 0 and the system information message always transmitted on PRB2 the Type 3N1 anchor segment can be said to fulfill the FCC digital modulation 500   kHz bandwidth requirement.

In case of frame structure Type 3N2, and before reading SIB1-NB a device can assume that NRSs are always present in the subframes that carry the uNPBCH and in subframes that may contain SIB1-NB transmissions according to the densest SIB1-NB configuration (see Section 15.3.2.3.5 to understand the SIB1-NB mapping).

These default minimal sets of uNRS transmissions are supporting idle mode NRSRP and NRSRQ measurements for the purpose of idle mode radio resource management, including cell selection and reselection.

Besides the default uNRS transmissions the devices can assume that uNRSs are present in any type of scheduled uNPDCCH and uNPDSCH transmissions during idle and connected mode operation.

Compared to the NB-IoT anchor channel the amount of default uNRS transmissions is significantly reduced. This should be seen as an attempt to better cope with the ETSI duty cycle requirement and limit the interference to other systems operating in the same unlicensed spectrum.

15.3.2.3.5. uNPBCH

Relative subframe	10–19
Periodicity	80   ms, 1280   ms
Basic TTI	640   ms, 10.24   s
Subcarrier spacing	15   kHz
Bandwidth	180   kHz
Frequency location	PRB 0

The uNPBCH carries the 34-bit MIB. A 16-bit CRC is appended to the information bits. The resulting 50 bits are convolutional encoded, and rate matched into 2432 bits. The encoded bits are divided into 8 code sub-blocks each of 304 bits. This procedure follows the NPBCH design (see Section 7.2.4.4). It is only the total number of encoded bits that has increased from 1600 for the NPBCH to 2432 for the uNPBCH.

The 304 bits of each code sub-block is modulated into 152 QPSK symbols. These are mapped to the available resource of a subframe where only resource elements for uNRS antenna ports 1 and 2 are reserved. In case of the NPBCH, resources are also reserved for the LTE control region and CRS transmissions. Fig. 15.16 shows the code sub-block mapping on a uNPBCH subframe. It can be compared to the NPBCH mapping illustrated in Section 7.2.4.4.

Fig. 15.16 uNPBCH resource element mapping.

The uNPBCH code sub-block subframe is repeated 10 times on relative subframes 10 to 19. Fig. 15.17 shows the complete mapping of one uNPBCH code sub-block onto the anchor segment for frame structures Type 3N1 and Type 3N2. For Type 3N1 the anchor segment and uNPBCH code sub-block periodicity equals 80   ms. This gives a total transmission time of 640   ms, just as for the NPBCH. For Type 3N2 each anchor segment, including the uNPBCH code sub-block, is transmitted with a periodicity of 1280   ms. This gives a total transmission time of 1.28   s ⋅ 8   =   10.24   s.

The transmission scheme use for uNPBCH is the same as for NPBCH, meaning that at most two antenna ports are supported using Space-Frequency Block Code (see Section 7.2.4.4) based transmit diversity.

Fig. 15.17 uNPBCH resource element mapping.

15.3.2.3.6. uNPDCCH

Relative subframe	Any DL subframes according to the UL-DL configuration
Basic TTI	1   ms
Repetitions	1, 2, 4, 8, 16, 32, 64, 128, 192, 256, 512, 1024, 2048
Subcarrier spacing	15   kHz
Bandwidth	90 or 180   kHz
Frequency location	Any

The uNPDCCH follows the NPDCCH specification described in Section 7.2.4.5. It carries uplink and downlink scheduling information by means of DCI formats N0 and N1, which are fully reused from NB-IoT (see Section 7.3.2.2). The uNPDCCH transmission opportunities are configured by means of user specific and common search spaces described in detail in Section 7.3.2.1. In frame structure Type 3N1 uNPDCCH transmissions are mapped onto the single PRB nframes, and on the center PRB in the anchor segment. In frame structure Type 3N2 uNPDCCH transmissions can only be mapped on nframe parts not reserved by the anchor segment. The uNPDCCH transmissions must also align with the applicable uplink-downlink configuration. It is, just as the NPDCCH, transmitted from the first valid subframe equal to or after its starting subframe. So in case the starting subframe of a search space is located in an uplink dwell, the actual start of the search space and its uNPDCCH transmission is postponed to the start of the next downlink dwell.

Thanks to the frequency hopping design in frame structure Type 3N1, the uNPDCCH will frequency hop in case its transmission is mapped across one or more nframe borders.

15.3.2.3.7. uNPDSCH

Relative subframe	Any DL subframes according to the UL-DL configuration
Basic TTI	1, 2, 3, 4, 5, 6, 8, 10   ms
Repetitions	1, 2, 4, 8, 16, 32, 64, 128, 192, 256, 384, 512, 768, 1024, 1536, 2048
Subcarrier spacing	15   kHz
Bandwidth	180   kHz
Frequency location	Any

The uNPDSCH supports unicast data and system information message transmissions along the lines of the 3GPP Release 13 NPDSCH design described in Section 7.2.4.6. This means e.g. that the largest supported transport block size is 680 bits. Single Cell Point to Multipoint (SC-PTM) feature specified in 3GPP Release 14 is not supported in the first NB-IoT-U release.

In frame structure type 3N1 uNPDSCH unicast data transmissions are mapped onto the single PRB nframes, and on the center PRB in the anchor segment. In frame structure type 3N2 uNPDSCH transmissions can only be mapped on nframe parts not reserved by the anchor segment. The dynamic scheduling follows that specified for the NPDSCH with the difference that the uNPDSCH transmissions must align with the applicable uplink-downlink configuration, and that transmissions spanning multiple nframes will frequency hop as described in Section 15.2.2.2.

In case of system information transmissions, the uNPDSCH scheduling is described in Section 15.3.3.1.

15.3.2.4. Uplink physical channels and signals

The design for the uplink channels and signals closely follow the NB-IoT Release 13 design described in Section 7.2.5. The main change is that the uplink transmissions follow the NB-IoT-U specific uplink-downlink configurations and, for frame structure type 3N1, is frequency hopping. Uplink transmissions are not supported in the anchor segment. In case of a potential overlap between a scheduled uplink transmission and the anchor segment, the applicable part of the uplink transmission is postponed for the duration of the anchor segment.

15.3.2.4.1. uNPRACH

Relative subframe	According to higher layer configuration
Basic TTI	5.6 or 6.4   ms
Repetitions	1, 2, 4, 8, 16, 32, 64, 128
Subcarrier spacing	3.75   kHz
Bandwidth	3.75   kHz
Frequency location	According to higher layer configuration

The uNPRACH supports the random-access procedure, using a design that follows the NB-IoT Release 13 specification described in Section 7.2.5.1. Both preamble formats 0 and 1 are supported. The use of the longer preamble format, with a basic transmission time of 6.4   ms, is not supported for frame structure Type 3N1 uplink-downlink configuration 5 which contains only 6 consecutive uplink subframes. All other uplink-downlink configurations are designed with at least 7 consecutive uplink subframes to support both preamble formats.

15.3.2.4.2. DMRS

Subframe	Any UL subframes according to the UL-DL configuration
TTI	Same as associated uNPUSCH
Repetitions	Same as associated uNPUSCH
Sub-carrier spacing	Same as associated uNPUSCH
Bandwidth	Same as associated uNPUSCH
Carrier	Any

The DMRS supports coherent demodulation of the uNPUSCH. Its design follows the NB-IoT Release 13 specifications. The only new aspects to take into consideration are the uplink frequency hopping design, and the uplink-downlink configuration.

15.3.2.4.3. uNPUSCH

Subframe	Any UL subframes according to the UL-DL configuration
Basic TTI	1, 2, 4, 8, 32   ms
Repetitions	1, 2, 4, 8, 16, 32, 64, 128, 256
Sub-carrier spacing	3.75, 15   kHz
Bandwidth	3.75, 15, 45, 90, 180   kHz
Carrier	Any

Two uNPUSCH formats are specified. uNPUSCH Format 1 (F1) supports uplink unicast transmissions, while uNPUSCH Format 2 (F2) is used for asynchronous HARQ feedback of uNPDSCH transmissions. The design of both uNPUSCH formats are close to identical to the Release 13 NPUSCH formats. New aspects to take into consideration are the uplink frequency hopping and the uplink-downlink configurations. For long NB-IoT NPUSCH transmissions, transmission gaps are specified to allow a device to perform NRS based frequency and time tracking. For the uNPUSCH such tracking is inherently supported thanks to the uplink-downlink configuration that introduces the needed uplink transmission gaps.

15.3.3. Idle and connected mode procedures

The NB-IoT-U physical layer design closely follows that of Release 13 NB-IoT. The same applies for the higher layer and physical layer procedures. This section presents idle and connected mode procedures and focuses on the few cases where significant differences between the NB-IoT and the NB-IoT-U procedures are identified.

15.3.3.1. Cell selection and system information acquisition

When performing initial selection, a NB-IoT-U device needs to search for anchor segment transmissions in the FCC ISM band 902–928   MHz and in ETSI SRD band 54 covering the frequency range 869.4–869.65   MHz. ETSI SRD band 47b covering the frequency ranges 865.6–865.8, 866.2–866.4, 866.8–867 and 867.4–867.6   MHz may also eventually be available. In the FCC ISM band a sparse anchor raster is specified which supports no more than three anchor carrier frequencies. This design is following the same principles as described for LTE-M-U in Section 15.2.3.1.

After synchronizing to an anchor carrier, a device can determine based on the frequency location of the anchor carrier which of the two frame structures Type 3N1 and Type 3N2 that is configured. Based on this information it may acquire the uNPBCH MIB, SIB1-NB and the other system information blocks.

15.3.3.1.1. System information acquisition for frame structure type 3N1

According to frame structure type 3N1 SIB1-NB is encoded into a codeword that spans over 20 subframes. This can be compared with the 8 subframes used for NB-IoT. The codeword is divided in two code sub-blocks of equal length, i.e. 10 subframes, which are mapped on the first half of PRB 2 on the anchor segment.

In case 16 repetitions are configured the two SIB1-NB code sub-blocks are mapped on the 10 first subframes of PRB 2 in two consecutive anchor segments leading to a code sub-block TTI of 160   ms. In case of 8 repetitions the code sub-block TTI is extended to 320   ms, with the first code sub-block being mapped onto the first anchor segment, and the second code sub-block being mapped on the third anchor segment. For the lowest repetitions count of 4 the code sub-block TTI is further scaled up to 640   ms. The first code sub-block is mapped onto the first anchor segment, and the second code sub-block being mapped on the fifth anchor segment. In all three cases of 4, 8 or 16 repetitions the SIB1-NB total transmission time equals 2560   ms for frame structure type 3N1.

In 3GPP the supported system information blocks 2, 3, 4, 5, 14 and 16 (see Section 7.3.1.2) are transmitted over 2 or 8 subframes. The subframe length is dependent on the transport block size. For NB-IoT-U the transmission time is extended from 2 or 8 to 10 subframes. In case of the 3GPP 2 subframe mapping the NB-IoT-U extension to 10 subframes is done by means of repetitions. In case of the 3GPP 8 subframe mapping the NB-IoT-U extension is achieved by means of rate matching. The 10 subframe system information code blocks are mapped on available subframes in PRB2 not occupied by SIB1-NB transmissions.

The scheduling of the system information blocks is arranged by means of non-overlapping and periodically reoccurring system information windows (SI-windows), like the concept illustrated for NB-IoT in Section 7.3.1.2.3 but each SI message can be configured with its individual SI-window length. In each SI-window only one system information message type is transmitted to fill the resources on PRB2 not consumed by SIB1-NB transmissions.

Fig. 15.18 illustrates the system information scheduling on the anchor segment in frame structure Type 3N1.

15.3.3.1.2. System information acquisition for frame structure type 3N2

In frame structure Type 3N2 the SIB1-NB message is encoded and mapped on 8 subframes. These 8 subframes are spread over the first half of the downlink dwells in a set of consecutive nframes. The second half of the dwell is reserved for control and data channel transmissions. This procedure is repeated 4, 8 or 16 times for sets of nframes that are spread evenly in time across a total SIB1-NB transmission interval of 5120   ms.

Fig. 15.19 shows an example where uplink-downlink configuration 5 (see Table 15.7) and 16 SIB1-NB repetitions are configured. In uplink-downlink configuration 5 the downlink dwell consist of 4 subframes. The SIB1-NB code block is mapped over the 2 first subframes in 4 consecutive nframes, starting from the second subframe after the anchor. The block of 4 nframes is repeated in total 16 times, with the repetitions spread evenly across the SIB1-NB transmission time interval of 5120   ms. This concept shares the principle behind the NB-IoT SIB1-NB mapping described in Section 7.3.1.2.1.

The scheduling of system information blocks 2, 3, 4, 5, 14 and 16 in case of frame structure Type 3N2 reuses the NB-IoT scheduling principles, described and illustrated in Section 7.3.1.2.3, with some minor adaptations to cater for the unlicensed spectrum operation.

Fig. 15.18 NB-IoT-U frame structure Type 3N1 system information scheduling.

Fig. 15.19 SIB1-NB scheduling example for frame structure Type 3N2.

The content of the MIB, and the system information messages are to large parts aligned with the NB-IoT Release 13 specification. Since operation in inband and guard-band modes are no longer relevant the MIB signaling related to these operations are excluded from the NB-IoT-U signaling. New information elements related to the two new frame Types 3N1 and 3N2, and the uplink-downlink TDD configurations have instead been added. NB-IoT-U supports in similarity to LTE-M-U both PLMN and NHN access modes (see Section 15.2.3.1.1). The needed signaling to support these two options are added to the SIB1-NB information elements.

15.3.3.2. Power control

The NB-IoT-U uplink reuses the NB-IoT open loop power control described in Section 7.3.2.3. The ETSI regulations mandates the use of an adaptive power control, and it was assumed that the existing functionality is sufficient to satisfy this requirement.

Both the ETSI and FCC regulations specify strict power control limits that needs to be met. The radiated peak power in ETSI band 54 equals 29.2 dBm EIRP. EIRP and its part in the path loss calculation performed in Section 15.4.2 is illustrated in Fig. 15.20. In FCC, a frequency hopping system is allowed to deliver a conducted peak power of 30 dBm and a radiated peak power of 36 dBm EIRP.

NB-IoT supports the device power classes of 14, 20 and 23 dBm. The 3GPP power classes are defined in terms of average conducted power. In case NB-IoT-U specifies the same three power classes, then a 23 dBm device operating in EU needs to guarantee that its PAPR does not exceed 29.2 – 23   =   6   dB. This should be possible for a NB-IoT-U device using a DFT pre-coded uplink with QPSK as highest modulation order.

In case of downlink single PRB transmissions, an efficient compression of the OFDM PAPR is important to maximize the average base station transmit power. In addition, the FCC requirements for a digitally modulated system needs to be met for the anchor transmissions in frame structure Type 3N1. For a digitally modulated system, the average conducted power may not exceed 30   dBm, and the average radiated power may not exceed 36   dBm EIRP. For the 3   PRB anchor this implies a maximum average EIRP of 31.2   dBm per PRB.

Fig. 15.20 Coupling loss and path loss.

15.4. Performance

The performance of systems operating in unlicensed spectrum is not as straightforward to evaluate as the performance of systems operating in controlled licensed spectrum. Uncoordinated systems using the same frequencies may expose each other to high uncontrolled interference levels. Duty cycle limitations do not only put an upper bound on the capacity that can be provided by a system in unlicensed spectrum but may also impact link level key performance indicators such as latency and throughput. Listen-before-talk regulations makes the operation unpredictable as the transmission is dependent on the channels availability.

In this section the NB-IoT-U link level performance is evaluated under the assumption of operation in a controlled radio environment not exposed to uncoordinated interference. This allows a comparison to the link level performance presented for EC-GSM-IoT, LTE-M and NB-IoT in Chapters 4, 6, and 8. When reading this section it should be kept in mind that a controlled radio environment in unlicensed spectrum may be rare in reality.

When it comes to LTE-M-U, limited performance evaluations were available at the time of writing this book. Consequently, performance of LTE-M-U is here only shortly described in general terms. LTE-M-U is intended to serve industry and enterprise networks with relatively small cell deployments. Compared to LTE-M the maximum supported cell size and coverage requirements have been significantly relaxed. Operation in CE mode B is for example not supported. The LPWAN coverage, latency and battery life requirements presented in the next sections are therefore not fully representative for the use cases envisioned for LTE-M-U.

15.4.1. Performance objectives

The NB-IoT-U performance is evaluated in terms of its supported path loss, and the achievable data rates, latency and battery life at the maximum path loss (MPL). The basic framework used in the evaluations is reused from to that presented in Chapter 8. The performance is evaluated both for operation in US using frame structure Type 3N1 and in ETSI using frame structure Type 3N2. The relevant regulations are when motivated taken into consideration.

15.4.2. Coverage and data rates

NB-IoT-U coverage performance is in this section presented for operation according to the ETSI and FCC regulations. Compared to the results presented for NB-IoT in Chapter 8 where coverage is evaluated in terms of maximum coupling loss (MCL) here the results are presented in terms of maximum path loss (MPL). The sum of the transmitting and the receiving nodes antenna gains defines the difference between coupling loss and path loss as can be seen in Fig. 15.20. MPL is a suitable metric for the ETSI and FCC unlicensed regulations since they specify output power limits in terms of EIRP, and not in terms of conducted power levels as traditionally done in 3GPP.

The base station EIRP is for ETSI specified in terms of peak power. The same applies for FCC for the frequency hopping single PRB downlink transmissions. To determine the average EIRP, which is relevant for MPL calculations, the PAPR of the transmissions needs to be assessed. Fig. 15.21 shows PAPR statistics applicable for single PRB uNPBCH, uNPDCCH or uNPDSCH transmission. The maximum PAPR is ∼9   dB. The uNPSS has thanks to its Zadoff-Chu base sequence a PAPR around 4   dB. The uNSSS PAPR is dependent on the configured PCID which defines the uNSSS waveform. In ETSI and FCC the allowed downlink peak EIRP equals 29.2 and 36 dBm, respectively. With a PAPR of 9   dB the average EIRP needs to be drastically reduced to 20 and 27   dBm to comply with the ETSI and FCC regulations. To lower the impact, it is possible to compress the uNPDCCH and uNPDSCH waveforms to achieve a reduced PAPR. Fig. 15.21 also presents the PAPR of uNPDCCH and uNPDSCH when the modulated signal amplitude is limited, or compressed, before the base station transmit filter. In this basic compression scheme a PAPR below 5   dB is achieved, which allows a downlink power of 24 and 31 dBm to be configured in ETSI and FCC, respectively.

Fig. 15.21 NB-IoT-U uncompressed and compressed PAPR characteristics.

The FCC digitally modulated anchor segment transmissions may use an average EIRP of 36 dBm. Since the anchor segment is transmitted over 3 PRBs this gives a maximum average EIRP of 31.2 dBm per PRB. Note that signal compression is not motivated by the FCC regulations for the case of digitally modulated waveforms.

The device output power is both in ETSI and FCC defined in terms of peak EIRPs of 29.2 and 36   dBm, respectively. Here we consider a device power class of 20   dBm, and a device antenna gain of 0   dB, which leaves more than sufficient margin to the allowed peak EIRPs to support the full PAPR of the SC-FDMA modulated NB-IoT-U uplink.

Table 15.9 presents the simulation assumptions used in the evaluation of the NB-IoT-U MPL. Besides the EIRP assumptions most parameters can be recognized from the NB-IoT evaluations presented in Chapter 8. In ETSI the uNPDCCH and uNPDSCH performance was evaluated using cross-subframe channel estimation over 4 subframes. This choice matches the number of consecutive downlink subframes in frame structure Type 3N2 uplink-downlink configuration 5 with an nframe duration of 20   ms. This configuration was chosen to reflect that the base station over time needs to respect the 10% duty cycle requirement. In frame structure Type 3N1 uplink-downlink configuration 1 supports uNPDCCH and uNPDSCH cross-subframe channel estimation over 7 subframes (not taking the special subframe into account), which was also the assumption used for the evaluations of the FCC performance. The uNPBCH was in both regulations evaluated using 10 subframes in the cross-subframe channel estimation. For the uNPUSCH F1 and F2 simulations 8 subframes cross-subframe channel estimation was assumed. The device and base station noise figures were reused from the 3GPP Release 13 performance evaluations on NB-IoT presented in Chapter 8.

Table 15.10 presents the performance resulting from link level evaluations performed for NB-IoT-U operation in ETSI. The MCL and MPL calculations follow the illustration in Fig. 15.20. Note that the presented uNPDCCH, uNPDSCH, uNPUSCH F1/F2 and uNPRACH transmission times are not accounting for the spreading in time that the uplink-downlink configurations cause. Some background to the declared BLER targets is given in Section 8.2.1. The choice of a 6 dBi base station antenna gain was inspired by the FCC regulations that explicitly mentions this antenna gain. The uNPBCH MIB was correctly encoded after reading 4 out of the 8 code sub-blocks. Synchronization to the uNPSS and uNSSS was achieved after 2 synchronization attempts.

Table 15.9

Assumptions made in the evaluations of NB-IoT-U MPL.

Parameter	ETSI value	FCC value
Frequency band	869.4–869.5   MHz	902–928   MHz
Propagation condition	Typical Urban (TU)
Fading	Rayleigh, 1   Hz
Frequency error	uNPSS/uNSSS: ±20   ppm uNPBCH, uNPDSCH, uNPDCCH, uNPUSCH, uNPRACH: ±50   Hz
Timing error	±2.5 μs
Device NF	5   dB
Device antenna configuration	1 transmit antenna, 1 receive antenna
Device EIRP	20 dBm
BS NF	3   dB
BS antenna configuration	2 transmit antennas, 2 receive antennas
BS EIRP	24 dBm	uNPSS, uNSSS, uNPBCH: 31.2 dBm uNPDSCH, uNPDCCH: 31 dBm
PAPR compression target	uNSSS, uNPBCH, uNPDCCH, uNPDSCH: 5   dB	uNPDCCH, uNPDSCH: 5   dB
Frame structure	Type 3N2	Type 3N1

The achievable MPL in ETSI equals 151   dB. The downlink performance can be said to be the limiting factor. Operation at a lower downlink SINR in ETSI will due to the long anchor segment periodicity result in very long uNPSS, uNSSS and uNPBCH acquisition times. The downlink data rate is also at risk to become unacceptably low.

It can be seen that the MCL corresponding to the declared MPL is limited to 145   dB. This serves as a reference for comparison with the 164   dB MCL achieved by EC-GSM-IoT, NB-IoT and LTE-M (see Chapters 4, 6, 8).

To calculate the uNPUSCH F1 and uNPDSCH MAC-layer data rates at the ETSI MPL the overhead from the uNPDCCH scheduling cycle must be taken into consideration. It depends on the transmission times declared in Table 15.10 weighted by the applied uplink-downlink configuration. For this purpose let's assume that the uplink-downlink configuration 6 (8 DL:1   S: 11 UL subframes) is configured.

In case of NPUSCH F1 scheduling, the UE specific search space can be assumed to be configured using RRC parameters R _max   =   256 and G   =   4 which determines the search space periodicity to R _max·G. This gives a uNPDCCH scheduling cycle and uMPUSCH transmission opportunity every 4·256   =   1024   ms (see Section 7.3.2.1) and an uplink MAC-layer data rate of 879 bps.

Table 15.10

NB-IoT-U MPL in ETSI for frame structure type 3N2.

Performance/Parameters	Downlink coverage				Uplink coverage
	NPSS/NSSS	NPBCH	NPDCCH	NPDSCH	NPRACH	NPUSCH F1	NPUSCH F2
TBS [bits]	–	34	23	680	–	1000	1
Bandwidth [kHz]	180	180	180	180	3.75	15	15
EIRP [dBm]	24	24	24	24	20	20	20
NF [dB]	5	5	5	5	3	3	3
#TX/#RX	2/1	2/1	2/1	2/1	1/2	1/2	1/2
Transmission/acquisition time [ms]	2560	3850	256	1024	25.6	80	2
BLER	10%	10%	1%	10%	1%	10%	1%
SNR [dB]	-11	-11.4	-11	-11	9.3	4.3	-2
MCL [dB]	145.4	145.8	145.4	145.4	146.0	145	151.2
TX antenna gain [dB]	6	6	6	6	0	0	0
RX antenna gain [dB]	0	0	0	0	6	6	6
MPL [dB]	151.5	151.9	151.5	151.5	152.0	151	157.2

THP = ( 1 − BLER ) · TBS MPDCCH Period = 0.90 · 1000 1.024 = 879 bps

(15.10)

For the uMPDSCH data rate calculations G   =   1.5 can be configured to obtain a scheduling cycle of 1.5·256   =   384   ms. This configuration allows a uMPDSCH transmission every ninth scheduling cycle, i.e. every 3456   ms and gives a downlink MAC-layer data rate of 177   bps.

Based on these configurations downlink duty cycles of 37% and 22% are calculated over the uMPDCCH scheduling cycle for the cases of downlink and uplink data transmissions. Such levels can only be tolerated during short durations to not violate the 10% duty cycle requirement measured over 1   hour. Data rates that can be sustained over a longer period of time while simultaneously meeting the duty cycle requirement will therefore be significantly lower than the just derived data rates.

Table 15.11 presents the performance achievable in FCC for frame structure type 3N1. Compared to ETSI the achievable MPL is increased by 10   dB to 161   dB. In the downlink this is facilitated by a higher permitted EIRP in combination with an 80   ms anchor segment periodicity which allows operation at a lower SINR with a rather limited impact on system acquisition time. In the uplink the SINR operating point is reduced by 10   dB compared to ETSI.

Table 15.11

NB-IoT-U MPL in FCC for frame structure type 3N1.

Performance/Parameters	Downlink coverage				Uplink coverage
	NPSS/NSSS	NPBCH	NPDCCH	NPDSCH	NPRACH	NPUSCH F1	NPUSCH F2
TBS [bits]	–	34	23	680	–	1000	1
Bandwidth [kHz]	180	180	180	180	3.75	15	15
EIRP [dBm]	31.2	31.2	31	31	20	20	20
NF [dB]	5	5	5	5	3	3	3
#TX/#RX	2/1	2/1	2/1	2/1	1/2	1/2	1/2
Transmission/acquisition time [ms]	340	570	512	2048	102.4	1024	16
BLER	10%	10%	1%	10%	1%	10%	1%
SNR [dB]	−13.3	−13.7	−14.7	−14.3	−0.7	−6.7	−8
MCL [dB]	155	155.3	156.1	155.7	156.0	155.9	157.2
TX antenna gain [dB]	6	6	6	6	0	0	0
RX antenna gain [dB]	0	0	0	0	6	6	6
MPL [dB]	161	161.4	162.2	161.8	162.0	161.9	163.2

The achieved MCL is again clearly inferior to the 164   dB MCL that can be achieved in licensed spectrum by EC-GSM-IoT, NB-IoT and LTE-M (see Chapter 4, 6, 8).

The reduced SINR operating point in FCC results in significantly reduced MAC-layer data rates. Again, a slightly uplink heavy uplink-downlink configuration was used for evaluating the uNPUSCH and uNPDSCH data rates. In this example frame structure type 3N1 uplink-downlink configuration 2 (7 DL: 1   S: 12 UL subframes) is assumed. In the case of the uNPUSCH the UE specific search space was configured by means of R _max   =   512 and G   =   8, leading to a uNPUSCH transmission every 512·8   =   4096   ms and a MAC-layer data rate of 220   bps.

For the uNPDSCH, G   =   1.5 can be assumed which leads to a scheduling cycle length of 512·1.5   =   768   ms, and a uNPDSCH transmission every 10th scheduling cycle. This gives a downlink MAC-layer data rate of 80   bps.

Tables 15.12 and 15.13 summarized the uplink and downlink MAC-layer throughput estimates for ETSI and for FCC. It is seen that the uplink data rate for FCC does not meet the 3GPP Release 13 target of 160   bps (see e.g. section 8.1). A more downlink centric uplink-downlink configuration would allow for a slightly improved FCC downlink data rate, but the 160-bps target is still not within reach. In ETSI the estimated data rates only apply for a base station duty cycle significantly exceeding the 10% regulation. If honoring the 10% duty cycle the ETSI downlink data rate would sink below the 160-bps requirement.

Table 15.12

NB-IoT-U downlink MAC-layer data rates at the MPL.

Regulation	UL-DL configuration	MPL	DL MAC-layer data rate	BS duty cycle
ETSI	Config. 6	151   dB	177 bps	37%
FCC	Config. 2	161   dB	220 bps	–

Table 15.13

NB-IoT-U uplink MAC-layer data rates at the MPL.

Regulation	UL-DL configuration	MPL	UL MAC-layer data rate	BS duty cycle
ETSI	Config. 6	151   dB	879 bps	22%
FCC	Config. 2	161   dB	80 bps	–

15.4.3. Latency

Table 15.14 presents the latency that is achieved when a device's higher layers initiates a transmission of a single uplink packet while the device is in idle mode. The modeled packet size equals 85 bytes on top of the PDCP layer, and 90 bytes when including overhead from PDCP, RLC and MAC as presented in Table 8.23. The overall evaluation framework including the latency definition is aligned with that used for NB-IoT and described in Section 8.4. The RRC Resume procedure is used (see Section 7.3.1.7.1), with the 90-byte packet included in Message 5 carrying the RRC Connection Resume Complete message. The link level performance used for deriving the overall latency corresponds to that presented in Section 15.4.2.

The applied uplink-downlink configurations are for simplicity aligned with those assumed in Section 15.4.2 in which the data rates at the MPL were evaluated. It can be seen in Table 15.13 that the 3GPP target of 10   seconds (see e.g. Section 8.4) is within reach for both regulations. In ETSI this comes at the cost of a base station duty cycle much higher than 10%, which means that latencies around 10   seconds can only be made available in exceptional cases. For frame structure type 3N2 uplink-downlink configuration 2, which is better aligned with a sustained duty cycle of 10%, the latency equals 19.3   seconds, which is far from the 10   seconds requirement.

Table 15.14

NB-IoT-U latency at the MPL.

Regulation	UL-DL configuration	MPL	Latency	BS duty cycle
ETSI	Config. 6	151   dB	10.2   s	31%
ETSI	Config. 2	151   dB	19.3   s	10%
FCC	Config. 2	161   dB	9.2   s	–

Given the higher SINR operating point in ETSI, it may seem contradictive that the latency is slightly better in FCC even when we overprovision the ETSI downlink duty cycle. The actual uplink report transmission is indeed taking longer time in FCC, but this is more than compensated for by the shorter system acquisition time and the fact that the FCC 3 PRB anchor segment supports uNPDCCH and uNPDSCH transmissions.

15.4.4. Battery life

The device power efficiency and ability to operate over many years on limited battery power is crucial for the massive IoT use case. To examine the NB-IoT-U battery life the evaluation performed for NB-IoT in Section 8.5 was repeated for NB-IoT-U. The NB-IoT assumptions and evaluation methods were reused with one exception: The lower device power class of 20 dBm assumed for NB-IoT-U motivates a reduction in the transmit power consumption from 500   mW (see Table 8.26) to 400   mW as shown in Table 15.15. The link level performance corresponds to that presented in Section 15.4.2.

The actual battery life achieved under the assumption of a 200-byte uplink transmission followed by a 65-byte downlink response is presented for FCC and ETSI in Table 15.16. The reporting interval is as seen following a 2-hour or 24-hour periodicity. In case of a diurnal reporting the 10-year target set in 3GPP Release 13 (see Section 8.5) is comfortably met. In case of reporting every 2   hour this is not the case. These results follow the same pattern as presented for NB-IoT in Table 8.27. As the power consumption is dominated by the 200-byte uplink transmission it is no surprise that the battery life in ETSI, thanks to the lower MPL, shows better performance than for FCC.

Table 15.15

NB-IoT-U device power consumption.

TX, 20 dBm	RX	Light sleep	Deep sleep
400   mW	80   mW	3   mW	0.015   mW

Table 15.16

NB-IoT-U battery life at the maximum path loss.

Regulation	Reporting interval	DL packet size	UL packet size	MPL	Battery life
ETSI	2   h	65 bytes	200 bytes	151   dB	4.9   years
	24   h				24.5   years
FCC	2   h	65 bytes	200 bytes	161   dB	2.2   years
	24   h				16.4   years

References

[1]. European Telecom Standards Institute, .  EN 300 328, V2.0.20. Wideband transmission systems; Data transmission equipment operating in the 2,4 GHz ISM band and using wide band modulation techniques; Harmonised Standard covering the essential requirements of article 3.2 of the Directive 2014/53/EU . March 2016.

[2]. U.S. Government Publishing Office, .  Electronic Code of Federal Regulations, Article 15.247 . 7 July 2016.

[3]. MulteFire Alliance.  Technical specification 36.300, v1.1.1. Evolved Universal Terrestrial Radio Access (E-UTRA) and Evolved Universal Terrestrial Radio Access Network (E-UTRAN); Overall Description; Stage 2 . 2019.

[4]. MulteFire Alliance.  Technical Specification 36.311, v1.1.1. Evolved Universal Terrestrial Radio Access (E-UTRA); Physical Channels and Modulation . 2019.

[5]. MulteFire Alliance.  Technical Specification 36.312, v1.1.1. Evolved Universal Terrestrial Radio Access (E-UTRA) . Multiplexing and Channel Coding; 2019.

[6]. MulteFire Alliance.  Technical Specification 36.313, v1.1.1. Evolved Universal Terrestrial Radio Access (E-UTRA); Physical layer procedures . 2019.

[7]. MulteFire Alliance.  Technical Specification 36.304, v1.1.1. Evolved Universal Terrestrial Radio Access (E-UTRA); User Equipment (UE) Procedures in Idle Mode . 2019.

[8]. MulteFire Alliance.  Technical Specification 36.321, v1.1.1. Evolved Universal Terrestrial Radio Access (E-UTRA); medium access control (MAC) Protocol Specification . 2019.

[9]. MulteFire Alliance.  Technical Specification 36.331, v1.1.1. Evolved Universal Terrestrial Radio Access (E-UTRA) . Radio Resource Control (RRC); 2019.

[10]. European Telecom Standards Institute, .  EN 300 220-1, V3.1.1. Short Range Devices (SRD) Operating in the Frequency Range 25 MHz to 1 000 MHz; Part 1: Technical Characteristics and Methods of Measurement . February 2017.

[11]. European Commission. Commission Decision of 9 November 2006 on harmonisation of the radio spectrum for use by short-range device (2006/771/EC), Edition of 18.8.2017.