VIETNAM NATIONAL UNIVERSITY – HOCHIMINH CITY

INTERNATIONAL UNIVERSITY
SCHOOL OF ELECTRICAL ENGINEERING

STUDY ON
MACHINE LONG-TERM EVOLUTION (LTE-M)

BY
NGUYỄN DUY QUANG

A SENIOR PROJECT SUBMITTED TO THE SCHOOL OF ELECTRICAL

ENGINEERING IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF BACHELOR OF ELECTRICAL ENGINEERING

HO CHI MINH CITY, VIET NAM

2021
 STUDY ON MACHINE LONG-TERM EVOLUTION (LTE-M)

BY

NGUYEN DUY QUANG

Under the guidance and approval of the committee, and approved by its members, this

thesis has been accepted in partial fulfillment of the requirements for the degree.

Approved:

________________________________
Chairperson

_______________________________
Committee member

________________________________
Committee member

________________________________
Committee member

________________________________
Committee member

i
 HONESTY DECLARATION

My name is Nguyen Duy Quang. I would like to declare that, apart from the acknowledged

references, this thesis either does not use language, ideas, or other original material from

anyone; or has not been previously submitted to any other educational and research

programs or institutions. I fully understand that any writings in this thesis contradicted to

the above statement will automatically lead to the rejection from the EE program at the

International University – Vietnam National University, Ho Chi Minh City.

Date: 27/01/2021

Student’s Signature

Nguyen Duy Quang

ii
 TURNITIN DECLARATION

Name of Student: Nguyễn Duy Quang

Date: 27/01/2021

Advisor Signature Student Signature

Nguyen Duy Quang

iii
 ACKNOWLEGMENT

It is with deep gratitude and appreciation that I acknowledge the professional guidance of

Dr. Ta Quang Hien. His constant encouragement and support helped me to achieve my

goal. He supported me with his deep knowledge about wireless communication and

provided me the materials in order to accomplish the project.

iv
 TABLE OF CONTENTS

HONESTY DECLARATION ......................................................................................... ii

TURNITIN DECLARATION ....................................................................................... iii

ACKNOWLEGMENT .................................................................................................. iv

TABLE OF CONTENTS.................................................................................................v

LIST OF TABLES ...................................................................................................... viii

LIST OF FIGURES ....................................................................................................... ix

ABBREVIATIONS AND NOTATIONS.........................................................................x

ABSTRACT ................................................................................................................. xii

CHAPTER I BRIEF STUDY..........................................................................................1

1.1. Background and Radio Access Design Principles .............................................1

1.2. The differences between LTE-M and other IoT networks................................2

1.3. Physical layer ......................................................................................................2

1.3.1. General principles .......................................................................................2

1.3.2. Numerology .................................................................................................3

1.3.3. Transmission schemes .................................................................................4

1.3.4. Downlink and Uplink physical layer ..........................................................5

CHAPTER II ACCESS PROCEDURES ........................................................................8

2.1. Idle Mode ............................................................................................................8

2.1.1. Cell selection................................................................................................8

v
 2.1.2. SI Acquisition ..............................................................................................9

2.1.3. Cell Reselection; Paging and eDRX; PSM ............................................... 10

2.1.4. RA and Access Control ............................................................................. 12

2.2. Connected Mode ............................................................................................... 13

2.2.1. Scheduling ................................................................................................. 13

2.2.2. Random Accessing .................................................................................... 16

2.2.3. Power Control and Mobility Support ...................................................... 16

2.3. MPDCCH Search Spaces and Frequency Hopping ........................................ 17

2.4. Improvements in Release 14 ............................................................................ 17

2.4.1. Support for Higher Data Rate .................................................................. 18

2.4.2. Improved Positioning; Multicast communication; Voice and Mobility

Enhancement ...................................................................................................... 19

CHAPTER III PARALLELIZATION OF RANDOM ACCESS BASED ON PREAMBLE

VARIETY ..................................................................................................................... 20

3.1. Introduction ...................................................................................................... 20

3.2. System Model.................................................................................................... 20

3.3. Parallelization of RA Based on Preamble Variety .......................................... 21

3.3.1. Key Feature ............................................................................................... 21

3.3.2. Overall Procedure ..................................................................................... 21

3.3.3. Performance Analysis ............................................................................... 22

vi
CHAPTER IV CONCLUSION AND FUTURE WORK ................................................ 23

REFERENCES .............................................................................................................. 24

APPENDICES............................................................................................................... 25

vii
 LIST OF TABLES

Table 2.1: SIB types in LTE-M ................................................................................... 10

Table 2.2: DCI 6-1A, 6-1B Layout ............................................................................. 15

Table 2.3: DCI 6-0A, 6-0B Layout ............................................................................. 15

viii
 LIST OF FIGURES

Figure 1.1: Difference between LTE-M & NB-IoT ....................................................... 2

Figure 1.2: LTE-M Frame Structure ............................................................................. 3

Figure 1.3: LTE-M Time Structure ............................................................................... 3

Figure 1.4: LTE-M Resource Grid ................................................................................ 4

Figure 1.5: LTE-M Narrowband ................................................................................... 4

Figure 1.6: LTE-M DL Subframe Structure .................................................................. 5

Figure 1.7: LTE-M PBCH Mapping.............................................................................. 6

Figure 1.8: LTE-M PRACH Preamble Structure. .......................................................... 7

Figure 1.9: LTE-M UL Reference Signal. ..................................................................... 7

Figure 2.1: LTE-M Protocol Stack ................................................................................ 8

Figure 2.2: LTE-M Cell Selection and Reselection ....................................................... 8

Figure 2.3: LTE-M Paging and eDRX cycle ............................................................... 11

Figure 2.4: LTE-M Random Access Procedures ......................................................... 12

Figure 2.5: LTE-M DL Scheduling example ............................................................... 14

Figure 2.6: LTE-M UL Scheduling example ............................................................... 14

Figure 2.7: LTE-M Frequency Hopping example ........................................................ 17

Figure 3.1: The System Model .................................................................................... 20

Figure 3.2: The procedure with N = 2 and k = 2 example ............................................ 22

ix
 ABBREVIATIONS AND NOTATIONS

IoT: Internet of Things

LPWAN: Low Power Wide Area Network

MTC: Machine-Type Communications

MAC: Medium Access Control

BS: Base Station

DL: Downlink

UL: Uplink

CP: Cyclic Prefix

PRBs: Physical Resource Blocks

PSS: Primary Synchronization Signals

SSS: Secondary Synchronization Signals

FDD: Frequency-Division Duplex

TDD: Time-Division Duplex

PBCH: Physical Broadcast Channel

MPDCCH: MTC Physical DL Control Channel

PDSCH: Physical DL Shared Channel

PRACH: Physical Random Access Channel

PUSCH: Physical UL Shared Channel

PUCCH: Physical UL Control Channel

UCI: Uplink Control Information

SIB: System Information Block

RSRP: Reference Signal Received Power

x
DRX: Discontinuous Reception

eDRX: extended-Discontinuous Reception

RA: Random Access

RAR: RA response

DCI: DL Control Information

C-RNTI: Cell Radio Network Temporary Identifier

HARQ: Hybrid Automatic Repeat Request

MBM: Multimedia Broadcast Multicast

SC PTM: Single-Cell Point-to-Multipoint

VoIP: Voice over Internet Protocol

ACK: Acknowledgement

xi
 ABSTRACT

LTE, referred to as 4G Cellular Network, is a standard and up-to-date generation of

wireless data transmission. A remarkable increase of number of devices are connected in

the future wireless network from different applications, which urges an emerging technique

enabling devices’ massive access. Especially for low-cost low-power devices such as smart

metering, sensors or alarms for industry, 3GPPP project has released a novel technology,

machine-LTE (LTE-M), for their access to LTE network at a narrower bandwidth.

Therefore, the project aims to study these robust performances of LTE-M, especially for

massive access. The project also considers improving LTE-M random access protocol,

which has not been optimized in the conventional wireless network, to support the demand

of massive access.

xii
 CHAPTER I

BRIEF STUDY

This chapter will introduce a brief study overview of the LTE-M technology, its advantages

over other Cellular IoT technologies and the physical layer of the technology.

1.1. Background and Radio Access Design Principles

Machine Long-Term Evolution, so called LTE-M, was introduced in September 2011 by

the 3GPP study. It is a LPWAN radio technology developed from the LTE network to meet the

low-end massive MTC market needs such as low device cost and complexity. Additionally, the

LTE-M, referred as Cat-M device category series, supports features for both IoT and MTC.

LTE-M targets three main objectives: Lower the device cost and complexity, enhance the

coverage and extend the battery lifetime of the device. First is how to lower the cost and

complexity of the device, which can be done by many techniques, such as Reducing the peak rate,

Single receive antenna implementation, Reducing the bandwidth, Reducing the maximum transmit

power, or Half-Duplex operation. To accomplish the objective, the manufacturing cost of LTE-M

devices must be of 1/3 of the LTE devices. The second objective is the CE techniques to Enhance

the Coverage of LTE-M, which targets to enable 20 dB wider in coverage operation compared to

LTE. However, after considering other aspects, the target was changed to 15 dB. There are 2 CE

modes: mode A and mode B. In CE mode B, by using the repetitions techniques, the original target

of 20 dB of coverage can be reached. The third objective is the Long battery lifetime. Release 12

and Release 13 has introduced the support to extend the battery lifetime of the device to ten or

more years. The LTE-M devices can consume less power during the activated time by reducing

the receive and transmit bandwidths. Moreover, these features also supported other 3GPP radio

1
access technologies. Additionally, there are other objectives that LTE-M also targets such as

Massive Number of Devices Propose, Deployment Flexibility, etc.

1.2. The differences between LTE-M and other IoT networks

LTE-M were created by the cellular industry in response to the rise of LPWAN

technologies. LTE was way too complicated for IoT, therefore, LTE-M was created. It was

designed to support device communication via carrier networks in a way that is less expensive and

more power-efficient than traditional LTE. First, LTE-M developed specifically for device

communication using simpler, cheaper chips that require less battery power than the ordinary LTE

network, which means LTE-M is a great option if the use case requires low power. The LTE-M is

also less expensive and less complex compared to LTE , so it is very promising for the low-end

MTC market. Despite of being more complicated and more costly than NB-IoT, LTE-M serves an

overly broad set of use cases, from water meters to agricultural monitors and beyond. In short,

LTE-M provides a wide area of coverage, which NB-IoT is currently lacking. Due to its pros and

cons, LTE-M network are already in place in the U.S., Mexico, Romania, some countries in

Europe, and properly being developed to cover even more countries in the modern world.

Figure 1.1. Difference between LTE-M & NB-IoT

1.3. Physical layer

1.3.1. General principles

The physical layer of LTE-M is built on the solution already available on LTE but with

improved support for MTC features. The basic transmission schemes, numerologies such as

2
channel raster, CP lengths, frame structure, resource grid, etc., are the same as LTE. Most changes

in physical layer design in LTE-M are designed to achieve the three main objectives cited above.

1.3.2. Numerology

LTE-M physical signals center frequency is transmitted within the narrow band and

mapped over 6 PRBs for each signal, supported with a 100 kHz Channel Raster.

Figure 1.2. LTE-M Frame Structure

Figure 1.3. LTE-M Time Structure

3
 The Frame Structure is demonstrated in Figure 1.2. As in the figure, there are 1024 Hyper

frame in total. Each Hyper frame consists of 1024 frames, 1 frame is divided into 10 subframes,

each subframes contain 2 slots of 0.5 ms. The number of OFDM symbols in each slot is

demonstrated in Figure 1.3 for different case of CP length. In practical, the normal CP length is

more common than the extended CP length. The Resource Grid is demonstrated in Figure 1.4.

There are 12 subcarriers in each PRB, each subcarrier is 15 KHz, which leads to 180 KHz in total.

Figure 1.3. LTE-M Resource Grid

1.3.3. Transmission schemes

There are 3 transmission schemes: Duplex Modes, Narrowband Operation and Coverage

Enhancement Modes. Both FDD and TDD operation Modes are supported in LTE-M.

Figure 1.5. LTE-M Narrowband

4
 Next is the Narrowband Operation demonstrated in Figure 1.5, which supports a

maximum of 6 PRBs channel bandwidth. To guarantee excellent frequency range even for the

reduced bandwidths, frequency hopping technique is proposed. The last scheme is the Coverage

Enhancement Mode. LTE-M supports many CE enhancing techniques, but the most important

one is the multi-subframe repetition technique, which means that a single transport block is

transmitted over multiple subframe, therefore providing higher transmit energy per information

bit. In LTE-M, CE mode A targets fairly modest CE to balance the simplified implements of LTE-

M low-cost devices and some extra coverage, achieved via a small number of repetitions. CE

mode B targets more extensive coverage for more challenging coverage conditions, achieved via

a large number of repetitions. In general, CE modes in LTE-M trade off performance (data rate,

latency, etc.) for better coverage. Additionally, in usual coverage, the LTE network will prefer the

ordinary operation mode to the CE mode to have the available better performance in LTE.

1.3.4. Downlink and Uplink physical layer

The physical layer transport data to higher layers via the MAC layer, in the form of

transport channels. There are 6 DL transport channels that LTE-M supports: the DL subframes,

the synchronization signals, the DL reference signals, PBCH, MPDCCH and PDSCH.

Figure 1.6. LTE-M DL subframe structure

5
 First is the Downlink Subframes, which continues to use the subframe REs in LTE, except

for MPDCCH, a new control channel of the LTE-M. Unlike being mapped to the control region

like LTE, the channel is aligned in the data region to prevent any conflicts between LTE-M new

channel and LTE ordinary channel. Therefore, in LTE-M, the LTE data region consist of both

control and data channel (MPDCCH and PDSCH). Next, LTE-M Synchronization Signals is

constructed on LTE’s PSS and SSS. PSS is based on a frequency-domain Zadoff-Chu sequence,

whereas SSS is based on maximum length sequences (m-sequences). Third is the predefined

Downlink Reference Signals, delivered by the BS to permit the device for approximation of the

DL propagation channel, hence operate the demodulation of the physical channels and the

measurements of DL quality. The fourth DL physical channel considered is the PBCH, which

provides essential information to operate in the network for the device by delivering the MIB,

positioned in the center 72 subcarriers of the bandwidth of the system. Following is MPDCCH,

which carries Downlink Control Information (DCI). Last is PDSCH, which transmits unicast data

and broadcasts information like SI, paging messages, random access related messages.

Figure 1.7. LTE-M PBCH Mapping

In LTE-M UL schemes, there are five physical channels and signals: the UL subframes,

PRACH, the UL reference signals, PUSCH, PUCCH. First is the Uplink Subframes, which are

6
all configured as acceptable for LTE-M transmission typically. Second is PRACH, which

initializes connection, allowing the serving BS to estimate the arrival time (round-trip

propagation delay) of the UL transmission. The basic channel preamble structure is demonstrated

in Figure 1.3. The PRACH sequences in the structure are basically Zadoff-Chu sequences.

Figure 1.8. LTE-M PRACH Preamble Structure.

Next is the predefined Uplink Reference Signals, transmitted from the device to the BS

to allow it to approximate the UL propagation channel, therefore demodulate UL physical

channels, perform issue timing advance commands and UL quality measurements. The last

channels are the PUSCH, which transmits unicast data; and the PUCCH, which carries the UCI.

Figure 1.8. LTE-M UL Reference Signal

7
 CHAPTER II

ACCESS PROCEDURES

Figure 2.1. LTE-M Protocol Stack

2.1. Idle Mode

2.1.1. Cell selection

Figure 2.2. LTE-M Cell Selection and Reselection

8
 The first step of the procedure is the cell selection, which main purpose is to detect,

synchronized to, and consider whether a cell is suitable or not. First, the device needs to

synchronize in time with PSS, acquiring a carrier frequency estimation of error, then acquires the

frame timing using SSS. Sequences of PSS and SSS are applied to acquire the PCID together,

allowing the device to get the CRS placement, hence demodulate PBCH and then decode it. PBCH

carries the Master Information Block (MIB). Because the subblock transmitted in a particular

frame is unknown, the device must perform blind decoding to decode the MIB by forming four

hypotheses. To correctly decode, the device must hypothesize the amount of antenna port used for

CRS transmission at the BS. Indicating a correct CRC will lead to a successful MIB decoding.

Which by then, the device has achieved the SIB1, information needed for scheduling The device

can detect and decrypt SIB1 based on the MIB and its scheduling information. The SIB1 includes

H-SFN, tracking area, PLMN identity and cell identity, therefore, the device can completely

synchronize to the framing structure by acquiring SIB1.

2.1.2. SI Acquisition

A device needs to obtain the complete System Information (SI) set once it have chosen

an appropriate camping cell. The device can choose the camping cell and gain access to the cell

using SIB1 and 2, which include the most essential SI. Even though LTE-M uses LTE definitions

of SIB, the LTE-M and the ordinary LTE devices SI are transmitted independently. This happens

because the SI transmission of LTE uses a too large channel bandwidth, hence the LTE-M devices

cannot receive the SI transmissions. SIB1 carries information such as cell barring and cell

reservation information, the tracking area identity, the PLMN identity, and the minimal RSRP,

RSRQ required to camp on and connect to the cell, which is essential for DL scheduling

transmissions. The SIB1 information is the same during a modification period of 5.12s at least,

9
but in practical it is usually far longer. The other SIBs are carried in each SI memo, and the SI

memo content is the same throughout an editable adjustment phase of 2, 4, 8, 16, or 64 paging

cycle of the cell but the time is usually far longer in practice. The other SIBs scheduling

information is described in Table 2.1. When modifying the cell SI, the network is able to imply it

to the devices within a 5-bit field, changing every time the SI content changes, referred as the SI

value tag. This allows the device to acquire the changed SI messages again without acquiring all

of them again, therefore consuming less energy. An LTE-M device valid time to store the SI is 3

or 24 hours (depends on the network configuration) and the device can acquire the SI again before

accessing the network when the SI becomes invalid.

Table 2.1. SIB types in LTE-M

2.1.3. Cell Reselection; Paging and eDRX; PSM

The cell reselection procedure is initiated when the device detects a RSRP stronger

neighbor cell, which can significantly improve the device battery lifetime. This procedure is one
10
of the core mechanisms for the mobility of idle mode. For idle mode operation, a paging

transmission is transmitted to the devices to check if its identity is matched with the paging

records, and replies by operating an RA procedure to connect to the cell if there is a match. This

checking has significant impacts on extending the battery lifetime of the device and improving the

latency of the data transmission to the device of DL. The paging occasions are determined by the

device identity and its DRX/eDRX design. This process is monitored in MPDCCH. Due to DRX

and eDRX mechanisms which provide the minimum power consumption for IoT applications, the

device can reach the Mobile Terminated (MT) traffic via paging in a short time. For applications

that cannot reach the device within that time, the Power Saving Mode can support further power

saving. Therefore, the MO transmission in UL device can still be performed without interruption.

The PSM is a separate feature that can be applied to every 3GPP radio access technology.

Figure 2.3. LTE-M Paging and eDRX cycle

11
2.1.4. RA and Access Control

Figure 2.4. LTE-M Random Access Procedures

The RA procedures of LTE-M is quite similar to LTE network, which is demonstrated in

Figure 1.4. First, after synchronizing to the network and analyzing the SIB2 information of

PRACH, the device transmits a PRACH preamble in order to connect to the network in Message

1. If the BS can detect the preamble, it will send back a RAR in Message 2, then the device will

send Message 3 which contains a Contention Request to connect. And finally, in Message 4 the

network delivers an access setup or resume and data of conflicted resolution to solves the conflicts

that the first step causes because of the transmission of the same preamble of multiple devices at

the same time. To improve the data transfer latency, the device can attach UL data to the MAC

layer. LTE-M supports either RRC resume process or connection operation. In addition, the device

must determine a suitable PRACH resource based on its coverage level estimation: If the CE level

increases, the repetitions must as well increase. If a RAR message is not received, it will make

12
further attempts until the RAR is received, or the maximum number of attempted is reached. In

LTE-M, the device can apply both PRACH preamble power ramping and CE level ramping

(leveling up the CE after a few unsuccessful attempts). In general, the RA procedure uses 5

physical channels: PRACH, MPDCCH, PDSCH, PUSCH and PUCCH.

Both ACB and EAB are supported in LTE-M. EAB information is carried in the scheduling

information for SIB14, contained by SIB1. If SIB14 is present, barring is triggered, otherwise it is

not triggered. The barring parameters can be changed any time.

2.2. Connected Mode

2.2.1. Scheduling

The BS sends a DCI contain in MPDCCH, transmitted through an MPDCCH search space

to actively schedule a device. The DCI contains the modulation and coding scheme, resource

distribution in both frequency and time domains, and data required to support the HARQ

retransmission scheme. There is also a CRC masked with a C-RNTI attached to the DCI, which

makes just only the DCI intended device can be decoded successfully. The other devices will

discard the CRC because it does not pass for them. There are three scheduling in this procedure:

The Dynamic Downlink and Uplink Scheduling and the Semipersistent Scheduling. First is the

Dynamic Downlink Scheduling. LTE-M network supports the scheduling principles such as

Cross-subframe scheduling and HD-FDD operation to lower the complexity of the device

implementation. The Cross-subframe scheduling in LTE-M is a delay of 2 ms from the DCI carried

in MPDCCH to the scheduled PDSCH.

13
 Figure 2.5. LTE-M DL Scheduling example

Next, in the Dynamic Uplink Scheduling, a DCI schedules a transmission in PUSCH 4ms

after that. The uplink scheduling in LTE-M is identical to LTE except for the asynchronous instead

of synchronous HARQ layer. Therefore, LTE-M retransmissions of HARQ are constantly and

openly scheduled by DCI. If a device needs to transmit data in connected mode but does not have

PUSCH resource, it can transmit a scheduling request for the resource on PUCCH. And in case of

no acceptable resource neither, it will apply the RA process. In general, DCI 6-1A and 6-1B layout

is used in CE A and B mode to dynamically schedules DL communication on PDSCH, similarly,

DCI 6-0A and 6-0B layout is used to dynamically schedules UL communication on PUSCH.

Figure 2.6. LTE-M UL Scheduling example

14
 Table 2.2. DCI 6-1A, 6-1B layout

Table 2.3. DCI 6-0A, 6-0B layout

Last is the Semipersistent Scheduling (SPS), which is supported for both LTE-M

downlink and uplink in CE A mode only. It is mainly driven by the VoIP services in LTE, but

periodic sensor reporting is a more possible use case in LTE-M. The SPS operation can be toggle

by a SPS-C-RNTI DCI. When activated, the DCI specifies what MCS, frequency resources, or

15
repetitions number, etc. that is needed at the periodic persistent resources. However, it is noted

that dynamic scheduling can override SPS whenever it needs to.

2.2.2. Random Accessing

For connected procedure, the device is able to start the RA procedure when an UL TA

control or grant request is needed. Then, a contention-based RA is executed with the same idle

mode transmissions of RAR and message 3. However, an RRC message is not included in message

3 and in step four, a C-RNTI is preferred over TC-RNTI to perform the contention resolution.

The BS can also order a device to initiate RA by sending a PDCCH order in either CE Mode A

or B. The PDCCH order can specify a dedicated PRACH preamble index by then to initiate a

contention-free RA, so that there is no necessary specific contention resolution phase, and the

procedure is stopped when RAR is received. If there is no indicated preamble index, the device

performs a contention-based RA normally.

2.2.3. Power Control and Mobility Support

Power control performance in LTE-M is identical to LTE. Nevertheless, CE mode B of

LTE-M devices is in poor coverage, which will constantly use maximum power transmission.

Additionally, whenever the PRACH CE of a device reaches level 3 in RA procedure, which is the

highest level, it will also constantly use maximum power during PRACH transmission.

Other than the Cell Selection and Reselection procedure, mobility mechanism of

connected mode, for example, RRC re-establishment and assembly release with redirection,

handover, measurement reportage, etc., is also supported in LTE-M, similar in LTE. Radio

Resource Management measurements can also be performed while the device has its receiver

retuned. The device compares CRS-based measurements with Qin and Qout thresholds to

determine if it is in sync with the serving cell or not. This evaluation is called Radio Link

16
Monitoring and if the outcomes do not sync for a specific number of times, over a specific period

of time, the device will declare Radio Link Failure and stop the transmission to prevent causing

undesirable interference. When this happens, the device can search for another cell via cell

selection and create the connection again.

2.3. MPDCCH Search Spaces and Frequency Hopping

There are two common procedures for LTE-M idle and connected mode: MPDCCH

Search Spaces, which is the opportunities of transmission for MPDCCH, including blind

decoding candidates with different MPDCCH repetitions and Frequency Hopping, a technique

that LTE-M use to offers means for frequency diversity for LTE-M transport channels apart from

PBCH and PSS/SSS. The time when the hops occur is indicated by the time intervals and the size

that the hops ought to be is indicated by the offsets in the frequency domain. The transmission of

distinctive devices frequency hops can occur at once due to the synchronized time intervals.

Figure 2.7. LTE-M Frequency Hopping example

2.4. Improvements in Release 14

In order to improve the performance of LTE-M, Release 14 has developed these features:

Support for higher data rate; Improved positioning; Multicast transmission; Voice enhancements

and Mobility enhancements. These are non-compulsory for LTE-M devices and network.

17
2.4.1. Support for Higher Data Rate

In Release 13, the Cat M1 devices is suitable for low data rate MTC applications, however,

if the LTE-M supports higher data rate, closer to LTE devices, it would be an attractive solution

for low cost device and extensive battery life IoT applications. There are several solutions

developed to support this feature. First is the New device category Cat-M2, which has a new

improvement of 5 MHz bandwidths instead of 1.4 MHz bandwidths in Release 13. These wider

bandwidths allow the DL and UL data transmission with 24 PRBs wideband instead of 6 PRBs

narrowband. Cat-M2 physical layer has rapid data peak rates of approximately 4 Mbps and 7 Mbps

in DL and UL, given by the maximum 4008 bits TBS and 6968 bits in DL and UL, respectively.

Nevertheless, the maximum channel BW is still 6 PRBs for control channels since the data rates

increment are not necessary in these channels. Therefore, the implementation efforts to update the

network to Cat-M2 for higher data rates will be relatively small. Additionally, a Cat-M2 device

is also able to serve as an ordinary LTE-M Cat-M1 and only initiates the improved features when

designed to do so by a BS. Next feature is the Wider Bandwidth in CE Modes. Release 14

establishes supported signaling that let the device indicate to the BS whether it would choose to

be designed with a certain maximum bandwidth in CE mode, and after that the BS can configure

the device based on this information. The third feature is the Higher Uplink TBS for Cat-M1,

which Release 14 supports a 2984 bits TBS instead of 1000 bits. The higher UL TBS is an elective

improvement that can be proposed in every duplex mode. Release 14 also supports 10 Downlink

HARQ Processes in FDD, which helps the device DL scheme to obtain the maximum TBS in

every subframe, without increasing the soft channel bits number saved in the device decoder.

Hence, increase the FD-FDD peak data rate of DL to 1 Mbps. The next feature is the HARQ-ACK

Package in HD-FDD, which transfer up to 3 serial HARQ-ACK package. This feature along with

18
the 10 DL HARQ processes helps the device to obtain information in 10 of 17 subframes in HD-

FDD DL, therefore increase its peak rate to 588 kbps. Last feature is the Faster Frequency

Retuning, which helps the transmitted signal getting a better performance of communication.

2.4.2. Improved Positioning; Multicast communication; Voice and Mobility Enhancement

Release 14 proposes OTDOA improvements regarding to the time and frequency domains

PRS designs. Due to the limited receive bandwidth, the LTE-M device will have advantage from

a longer duration PRS than a wide bandwidth PRS. Thus, the new PSS design in Release 14 is

delivered with a slower and more frequently PRS occasion, which allows the devices to reach the

same positioning accuracy as in LTE. Moreover, Release 14 also propose features to support

multicast communication constructed on the MBM Service framework, referred as the SC PTM

transmission. This transmission was supported in both Machine LTE and Narrow Band IoT idle

mode. This feature is beneficial for some MTC applications with the requirement of effective

software allocation or firmware enhancements to an enormous size of devices.

Real-time applications of delay sensitive like VoLTE are already supported in Release 13,

but the delay-tolerant MTC applications was more preferred. Therefore, features like Dynamic

HARQ-ACK delays; Modulation structure restraint and SRS coverage enhancement; New PUSCH

repetition aspects are proposed in Release 14 to optimize the coverage of VoLTE, specifically for

TDD and HD-FDD. These features can also be used for other applications besides VoLTE. For

Mobility Enhancement, Release 14 connected mode have a full mobility support in the shape of

interfrequency RSRP/RSRQ and intrafrequency RSRQ estimations. These enhancements are

preferred for the more transportable LTE-M use cases such as wearables and VoLTE.

19
 CHAPTER III

PARALLELIZATION OF RANDOM ACCESS BASED ON PREAMBLE

VARIETY

3.1. Introduction

This section introduces a new RA parallelization technique created on a concept of

preamble variety to efficiently minimize the confrontation problem in the LTE-M network. This

scheme allows immediate transference of various preambles at an RA slot, allows each device to

attempt multiple RAs in parallel. Therefore, the RA failure is able to be considerably decreased

due to the prevention of the fully collision of the entire preambles with the preambles chosen by

other devices. Additionally, the scheme also seeks to minimize the RA failure probability.

3.2. System Model

Figure 3.1. The System Model

The system model is described in Figure 3.1. This is a single cell network with devices

that are evenly positioned inside the cell, assuming the network traffic is not overwhelmed. Each

device attempts the RA process at the next-accessible RA slot (PRACH) for uplink transmissions.

This system considers one single RA slot for analytical tractability and each device selects the

same preambles number. M is the available RA preambles number, N is the active devices

20
simultaenously attempt Ras number and k is the simultaneously transmitted preambles number (k

= 1 in the existing LTE/LTE-A systems). The RA success or failure is primarily affected by the

preamble collisions only. Nevertheless, a single preamble collision does not cause the RA failure,

it only occurs once all k preambles of a device are used by other devices all at once.

3.3. Parallelization of RA Based on Preamble Variety

3.3.1. Key Feature

In this scheme, each device transmits more than one preamble, enable it to initiate multiple

RA procedures in parallel. The scheme decreases the RA failures due to its efficiency to avoid the

entire preambles are selected by one single device. However, the performance is not always

improved by increasing k because it may lead to an overloaded in PRACH. Moreover, more

signaling is required, thereby extend the energy use and reduce the efficiency of the radio resource.

3.3.2. Overall Procedure

The procedure includes four steps: Preamble transmissions; Random access responses;

Scheduled transmissions; Acknowledgment. In Step 1, each device selects different k preambles

and transmits them to the BS at the same time through the PRACH. After detecting the preambles

transmitted, the BS delivers RARs in Step 2. Each includes a preamble index, an uplink grant, a

temporary identifier, and a timing alignment value. Next, the device evaluates the Step 1 k

preamble indices with those comprised in each RAR. Then it progresses k subsequent steps in

parallel in Step 3. Each device makes less than k multiple replicas of its original packet and delivers

each packet on each of the assigned uplink resources indicated by the UG value comprised in RAR.

Finaly, in Step 4, after the packets are decoded succesfully by the BS, the ACK messages will be

delivers to the devices. Each device RA attempt is considered successful if it receives one or more

ACK messages. Otherwise it will attempt the RA process again at the next-accessible RA slot.

21
 Figure 3.2. The procedure with N = 2 and k = 2 example

3.3.3. Performance Analysis

This analysis is considered from a single device of interested d0 perspective. The RA falure

probability is the probability that other devices use all k preambles chosen by d0 . If there is at

least one exclusively utilized by d0 preamble exists, it can successfully complete its RA procedure.

The probability is denoted as pf(k), which is derived by the formula: pf(k) =

(𝑀−𝑘)!(𝑀−𝑚)! 𝑁−1 𝑘
∑𝑘𝑚=0(−1)𝑚 [ ] ( ). When k = 1 for LTE/LTE-A system, the formula is reduced
(𝑀−𝑚−𝑘)!𝑀! 𝑚

1 𝑁−1
to: pf(1) = ∑𝑘𝑚=0 1 − [1 − 𝑀] . It can be seen from the formula that the failure probability is

highly dependent on k value. To minimize the probability, an optimazation problem can be

formulated to find the optimal k’: k’ = argmin pf(k) (1≤ 𝑘 ≤M). The value number can be found

when N is availabe to the BS. In conclusion, the scheme introduced in this section is very

ssufficient for cellular IoT, in this case LTE-M in terms of minimizing the confrontation problem

compared to other schemes. The scheme can also extend the contending resources amount, such

as the M value, or modify the traffic load sufficiently through several access control mechanism,

in order to take advantage of the performance gain effectively.

22
 CHAPTER IV

CONCLUSION AND FUTURE WORK

The global cellular IoT market is growing rapidly and intensively due to the constantly

developing technologies. Therefore, highly efficient network such as LTE-M is continuously

invested to meet the future demands of 5G networks. However, despite of being developed based

on LTE technology but with new and improved features, there still exists some problems in the

LTE-M network that need to be solved in order to develop a better performance.

The LTE-M network has many benefits such as low cost, low complexity, long battery

lifetime, extended coverage, etc. as mentioned in the above sections. These benefits allow the

network to be implemented in a wide range of use cases, for example, Low-Density Sensor,

Automatic Meter Reading, Asset Tracking, etc. Nevertheless, to meet the future demands of 5G

evolution of IoT, the network still has a lot of aspects to be improved and optimized. First is the

deployment cost, although being much cheaper than the ordinary LTE network, the LTE-M market

is still not widely used like LTE because of the high deployment cost. Second is the need to

improve the Random Access Procedure due to the increasement of the IoT devices. Because the

number of IoT devices is massive and still increasing as technology is growing, there is not enough

space for all devices to access, therefore causing many more issues such as latency, interferences,

etc. In Chapter III, a proposal to solve a problem in RA procedure was mentioned. Hence, future

work may involve further improving and practical simulation to solve the RA problems in LTE-

M network, which will significantly improve the network performance.

23
 REFERENCES

[1] Olof Liberg, Marten Sundberg, Y.-P. Eric Wang, Johan Bergman, Joachim Sachs, “Cellular
Internet of Things Technologies, Standards, and Performance” (2018, Academic Press)
[2] Taehoon Kim, Inkyu Bang, “Random Access Parallelization Based on Preamble Diversity
for Cellular IoT Networks” (IEEE COMMUNICATIONS LETTERS, VOL. 24, NO. 1,
JANUARY 2020)
[3] Erik Dahlman, Stefan Parkvall, Johan Skld, “4G, LTE-Advanced Pro and The Road to
5G” (2018, Third Edition, Academic Press)
[4] Link: https://www.iotacommunications.com/blog/lte-m-vs-nb-iot/
[5] Link: https://www.rfwireless-world.com/Terminology/LTE-M-Frame-Structure.html
[6] Link: https://www.mathworks.com/help/lte/ug/synchronization-signals-pss-and-sss.html

24
APPENDICES

25