Performance analysis and optimization for coverage enhancement strategy of Narrow-band Internet of Things

摘要

Narrowband Internet of Things (NB-IoT) is a clean-slate wireless protocol proposed by the 3rd Generation Partnership Project intending for massive machine type communications. 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 20dB enhanced coverage of NB-IoT, the concept of narrow-band modulation, coverage classes updating and adaptive repetition were introduced. To evaluate the performance of these new techniques, a Markov chain model with coverage classes as state variables was proposed to describe the dynamics of the coverage classes updating mechanism of NB-IoT. In addition, an optimization model minimizing average probability of access failure as well as average power consumption was formulated, with which the effects of preamble repetition number, system load and global maximum transmission number on the optimal configuration of maximum transmission number of each coverage class was analyzed. To solve the aforementioned optimization model, two algorithms, namely exhaustive search method combining constraint transform and particle swarm optimization (PSO) algorithm with self-adaptive stochastic inertia weight were proposed. Numerical results show that the maximum transmission number of normal coverage and extended coverage have a great influence on the system performance and their value ranges should be set within [3,10] and 11,61 respectively; while the maximum transmission number of extreme coverage has little influence on the system performance, and its recommended value is 1 for smaller power consumption. The average power consumption of coverage classes updating mechanism with coverage classes rollback is about 61% lower than that of 3GPP proposed model. All these researches together provide good reference for scale deployment of NB-IoT. (C) 2019 Elsevier B.V. All rights reserved.