Performance analysis and optimization for coverage enhancement strategy of Narrow-band Internet of Things
Introduction
To support communications among billions of miscellaneous innovative devices, Internet of things (IoT) has gained unprecedented momentum and commercial interest. That is, almost everything can be connected through IoT network. According to the stipulated protocol, IoT connects any things with the Internet to exchange information through radio frequency identification (RFID), infrared sensors, global positioning system, laser scanners and other information sensing devices, which aims to realize intelligent identification, location tracking, monitoring and management. As we know, IoT is a very broad concept. From the perspective of transmission rate, the communication services of IoT can be coarsely classified into three categories: high-data-rate services, medium-data-rate services and low-data-rate services. High-data-rate services are the services with data transmission rate up to 10 Mbps, such as video surveillance, auto-driving and healthcare system [1], [2], which are mainly realized by Long Term Evolution-Vehicle (LTE-V), Long Term Evolution-Advanced (LTE-A) technologies, Mobile cloud Computing (MCC) [3] and Edge Computing [4], [5]; Medium-data-rate services are the services with data transmission rate less than 1 Mbps, such as wearable equipment, intelligent home defense, elevator advertising, etc., which are mainly realized by enhanced Machine-Type communication (e-MTC), General Packet Radio Service (GPRS) technologies; Low-data-rate services are those with data transmission rates below 200 kbps, such as meter reading service, intelligent agriculture and intelligent parking service [6], which are mainly implemented by Narrow-band Internet of Things (NB-IoT), Long Range Radio (LoRa) and other technologies. Among the three categories, the low-data-rate services represent more than 67% of total IoT services [7]. In addition, NB-IoT supported by 3rd generation partnership project (3GPP) works in authorized spectrum [8], it can be directly deployed in GSM or LTE network, which will inevitably become the mainstream technology in the future low-power Wide Area Network (WAN) market.
NB-IoT is a new radio access technology developed by 3GPP to support the kind of applications aiming at massive lower rate data sensing and acquisition, such as smart meters, environmental monitoring and so on [9], [10]. The general objectives of the NB-IoT include supporting massive connections, enhanced coverage, reduced cost and complexity, ultra-low power consumption, and flexible delay characteristics. To achieve the objective of 20 dB enhanced coverage of NB-IoT for wide area outdoor coverage or deep indoor coverage, 3GPP proposed a special link adaptation technology, that is, NB-IoT device determines its coverage class (CC) according to its channel state and then implements the corresponding coverage enhancement (CE) mechanism, such as adaptive repetition [11]. Coverage class is a new concept proposed by 3GPP for NB-IoT. According to the maximum coupling loss (MCL), NB-IoT defines three coverage classes, namely, Normal, Extended, and Extreme coverage classes. Classes are differentiated through thresholds based on received signal strength, defined to introduce three levels of coverage extension w.r.t GSM/GPRS: 0 dB, 10 dB, and 20 dB for Normal, Extended, and Extreme, respectively [12]. The intention of this design is that usually the coverage radius of NB-IoT base station is large (30 km or larger) and NB-IoT devices may be deployed in challenging positions (such as garage, basement, etc.); If all NB-IoT devices adopt the same CE mechanism, the performance of NB-IoT devices experiencing unsatisfactory channel environment would deteriorate dramatically; thus a unified framework of CE mechanism (mainly relies on adaptive repetition) with dedicated parameters setup for different coverage classes are proposed; the key idea of NB-IoT CE mechanism is to trade more power consumption for higher access success probability, that is, to allow NB-IoT devices in poor channel environment to accumulate more transmission power via more repetitions; in term of this, access success probability could be guaranteed without increasing too much power consumption of NB-IoT devices.
At present, few theoretical researches focus on NB-IoT performance analysis. Most of existing references give only performance analysis of NB-IoT by use of simplified field tests or numerical simulations. Moreover, these results vary greatly because of different deployment scenarios or parameters setup. The concerns on optimal design or configuration of NB-IoT are rare less [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23]. In terms of NB-IoT coverage enhancement mechanism, researches are limited to static analysis by use of MCL prediction. References [17], [18], [19], [20], [21] tested coverage performance of NB-IoT network deployed by operators and gave suggestions on approaches of network construction. Reference [21] gave suggestions on how to configure the number of preamble repetitions. Reference [22] upgraded existing commercial LTE network to LTE-Machine to Machine (LTE-M) and NB-IoT networks, and measured their coverage and capacity performance in rural areas. Reference [23] analyzed coverage performance of NB-IoT main physical channels by using a typical urban channel propagation model. Reference [24] proposed an uplink adaptive scheme with deterministic transmission times to ensure transmission reliability and improve the throughput of NB-IoT. Reference [25] characterized a fundamental trade-off between ‘repetition’ and ‘retransmission’ schemes in NB-IoT random access channel (RACH). These works provide good reference for further studies on NB-IoT theoretical analysis. However, without taking into account NB-IoT coverage classes updating mechanism, all the above researches fail to describe the dynamic working process of NB-IoT coverage enhancement mechanism. This is because when channel state changes or the number of consecutive access success or failure reaches the maximum number specified in the current coverage class, NB-IoT device needs to dynamically adjust its current coverage class. Moreover, 3GPP does not explicitly give the coverage classes updating mechanism of NB-IoT which should specify how to determine the transition conditions of handover between different coverage classes.
To solve aforementioned problems, the main contributions of this paper can be summarized as follows:
- (1)
We modeled the coverage classes updating mechanism of NB-IoT as a Markov process with coverage classes as state variables, established the transition probability matrices between different coverage classes and calculated the steady state probability of each state.
- (2)
We formulated an optimization model minimizing average probability of access failure as well as average power consumption of NB-IoT devices to study the optimal configuration strategy of NB-IoT coverage classes updating mechanism. The effects of preamble repetition number, system load and global maximum transmission number was also analyzed.
- (3)
We provided improved theoretical analysis of NB-IoT coverage classes updating mechanism, which can provide strong technical support for potential commercial process of NB-IoT.
In the process of solving optimization model proposed in this paper, we have used the particle swarm optimization (PSO) algorithm with self-adaptive stochastic inertia weight. The process of standard PSO algorithm with fixed inertial weight is more like particle clustering, particles all fly to a current optimal direction, resulting in the diversity of particles decreasing and fitness stagnating, especially in late evolution period, the algorithm lacks the ability to improve solution. Aiming to solve the problems, the performance improvement of the PSO algorithm can be reported as: Reference [26] proposed guaranteed convergence PSO (GCPSO) algorithm which has local convergence performance. Reference [27] proposed a PSO algorithm with time-varying acceleration coefficients, which combines PSO algorithm with immune algorithm to improve the diversity of particles so that particles can avoid falling into local optimum. In this paper, we analyzed the effects of inertia weight and global optimal value on standard PSO algorithm, then we proposed using the PSO with a self-adaptive stochastic inertia weight which varies with the global optimal value, and its simulation results show that it can match our engineering expectations well.
The rest of this paper is organized as follows. Section 2 introduces NB-IoT uplink random access procedure and all the parameters used in this paper as well as their symbol representations and explanations. Section 3 presents two Markov chain models for standard NB-IoT coverage class updating mechanism proposed by 3GPP and the one with rollback mechanism proposed by us respectively. Section 4 presents the optimization model and its solution method. Section 5 gives the numerical analyses and their discussions. Finally, Section 6 concludes this paper.
Section snippets
NB-IoT uplink random access procedure
As shown in Fig. 1, NB-IoT device needs to determine its initial coverage class after downlink synchronization, and it also needs to update the coverage class in its RA procedure. Once random access is successful, targeted coverage enhancement can be adapted to subsequent data transmission. Since most of NB-IoT applications have long triggering cycles, that means most of them are time-insensitive, NB-IoT only supports contention-based random access, and adopts four hand-shake steps similar to
Updating model abstracted from 3GPP protocol
As can be seen from Section 2, NB-IoT device would switch to the next coverage class when the number of consecutive access failure reaches . If device succeeds in random access within , it could start its data transmission, and re-determine its initial coverage class when the next random access is initiated. However, due to dynamics of channel state, device may make decision mistakes and choose wrong initial coverage class. This implies the initial coverage class could be any of the
Optimization model of coverage class updating mechanism
NB-IoT RA procedure reduces the access failure probability by increasing the number of preamble transmission and repetition, i.e. by increasing power consumption and delay in exchange for coverage enhancement. However, NB-IoT devices are generally powered by batteries, in order to achieve the goal that 5 Wh batteries can have a life of up to 10 years, it is necessary to minimize the system power consumption under the requirement of a certain access failure probability. Therefore, how to
Numerical analysis
Based on Section 4.2, this section studies the influence of different parameter configurations on the optimal values of and the corresponding , E, and compares the model abstracted from the 3GPP with the model introducing rollback mechanism. This paper sets threshold ; Power consumption of sending preambles for one time .5610−7 Wh.
Conclusion
This paper aims to establish a theoretical model for NB-IoT performance analysis and its coverage class updating mechanism to analyze the influence of main system parameters on the performance of NB-IoT. Taking three coverage classes proposed by 3GPP as state variables, the RA procedure of NB-IoT is modeled as a Markov model. And this paper proposes a coverage class updating mechanism which can fully characterize the dynamic changes of coverage class in the RA procedure. Meanwhile, an
Acknowledgments
This work was funded by the National Natural Science Foundation of China (No. 61501065, No. 61571069, No. 61701054 and No. 61771080), the Fundamental Research Funds for the Central Universities, China (No. 2018CDXYTX0009, No. 2019CDXYTX0023), the Chongqing Research Program of Basic Research and Frontier Technology, China (No. CSTC2016JCYJA0021).
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work
Xiangming Wang is currently pursuing his B.S. degree with the College of Micro-Electronic and Communication Engineering, Chongqing University, China. His research interests include NB-IoT and wireless sensor networks.
References (36)
- et al.
Edge cognitive computing based smart healthcare system
Future Gener. Comput. Syst.
(2018) - et al.
Narrow band internet of things: simulation and modelling
IEEE Internet Things J.
(2017) - et al.
Multi-objective optimization of nanofluid flow in flat tubes using CFD, artificial neural networks and genetic algorithms
Adv. Powder Technol.
(2014) - et al.
An iterative solution approach to a multi-objective facility location problem
Appl. Soft Comput.
(2018) - et al.
5G-smart diabetes diagnosis with healthcare big data clouds
IEEE Commun.
(2018) - et al.
Opportunistic task scheduling over co-located clouds in mobile environment
IEEE Trans. Serv. Comput.
(2018) - et al.
Label-less learning for traffic control in an edge network
IEEE Netw.
(2018) - et al.
A dynamic service-migration mechanism in edge cognitive computing
ACM Trans. Internet Technol.
(2019) - et al.
Software-defined network function virtualization: A survey
IEEE Access
(2015) - et al.
Narrow band internet of things
IEEE Access
(2017)
Cognitive-LPWAN: Towards intelligent wireless services in hybrid low power wide area networks
IEEE Trans. Green Commun. Netw.
Overview of narrowband iot in LTE Rel-13
New 3GPP standard for IoT [mobile radio]
IEEE Veh. Technol. Mag.
RACH Preamble repetition in NB-IoT network
IEEE Commun. Lett.
M2M Massive access in LTE: RACH performance evaluation in a smart city scenario
Data channel design and performance for LTE narrowband IoT
NB-IoT Deployment study for low power wide area cellular IoT
NB-IoT System for M2M communication
Cited by (5)
Optimization of Computer Network Connection Enhancement Relying on Nonlinear Active Queue Management Algorithm
2024, Lecture Notes in Electrical EngineeringMonitoring and control of smart urban drainage systems using NB-IoT cellular sensor networks
2023, Water Science and TechnologyData allocation optimization for sensor information of internet of things
2021, International Journal of System Assurance Engineering and Management
Xiangming Wang is currently pursuing his B.S. degree with the College of Micro-Electronic and Communication Engineering, Chongqing University, China. His research interests include NB-IoT and wireless sensor networks.
Dong Yang received the master degree from University of Electronic Science and Technology of China in 2012. He is now a senior engineer of the 26th institute of China Electronic Technology Group Corporation. His main research interests include the Internet of things, intelligent signal processing and implementation of digital signal processing based on FPGA.
Xin Jian (corresponding author) received his B.E. and Ph.D. degree from Chongqing University, Chongqing, China in 2009 and 2014, respectively. He is an associate professor at the College of Micro-Electronic and Communication Engineering, Chongqing University, China. His interests include internet of things, next generation mobile communication, and wireless ad hoc network.
Min Chen has been a full professor in the School of Computer Science and Technology at HUST since February 2012. He is Chair of the IEEE Computer Society STC on big data. His Google Scholars Citations reached 13500+ with an h-index of 58. He received the IEEE Communications Society FredW. Ellersick Prize in 2017. His research focuses on cyber physical systems, IoT sensing, 5G networks, SDN, healthcare big data, etc.
Jože Guna is an Assistant Professor at the Faculty of Electrical Engineering, University of Ljubljana. His area of research focuses on Internet technologies, multimedia technologies and IPTV systems with special emphasis on user centered design, user interaction modalities and designing the user experience, VR/AR/MR technologies, including gamification and flow aspects. Currently he is involved in a number of projects focusing on the development of intuitive user interfaces for elderly users of eHealth application and interactive multimedia HBBTV and VR/AR/MR applications. He is an expert in Internet, ICT and IPTV technologies and holds several industrial certificates from CISCO, Comptia and Apple, including trainer licenses from Cisco and Apple. He is a senior member of the IEEE organization and IEEE Slovenia Section Secretary General.