Intelligent RACH Access Techniques to Support M2M Traffic in Cellular Networks

Lawal Mohammed Bello, Paul D. Mitchell, David Grace
2018 IEEE Transactions on Vehicular Technology  
Intelligent RACH access techniques to support M2M traffic in cellular networks. IEEE Transactions on Vehicular Technology. ISSN 0018-9545 eprints@whiterose.ac.uk https://eprints.whiterose.ac.uk/ Abstract -This paper provides a thorough investigation into the use of Q-learning as a means of supporting machine-tomachine (M2M) traffic over cellular networks through the random access channel (RACH). A new back-off scheme is proposed for RACH access, which provides separate frames for M2M and
more » ... ional cellular (H2H) retransmissions , and is capable of dynamically adapting the frame size in order to maximise channel throughput. Analytical model s are developed to examine the interaction of H2H and M2 M traffic on the RACH channel, and to evaluate the throughput performance of both slotted ALOHA and Qlearning-based access schemes. It is shown that Q-learning can be effectively applied for M2M traffic, significantly increasing the throughput capability of the channel with respect to conventional slotted ALOHA access. Dynami c adaptation of the back-off frames is shown to offer further improvements relative to a fixed-frame scheme. ). M2M communication will enable interaction between various devices with or without human intervention. M2M devices may be sensors, actuators, embedded processors, radio frequency identification (RFID) tags, smart meters, etc., [7] . The devices may be connected using wired or wireless access networks. Although wired networks are considered to be reliable and secure, they are very expensive to roll out and are not very flexible. As a result, many standards are not considering wired networks as an option for M2M communication. On the other hand, a wireless network is capable of providing excellent coverage, flexibility, mobility , and roaming capability. Hence, wireless access networks, which may be short range or long range (e.g. cellular), are considered as the most suitable option to deploy M2M communication [6] . To realise cellular M2M communication, different wireless communication standardisation bodies , including 3GPP, are actively involved in research to provide global standards. Initial access to a cellular network is through the random access channel (RACH) which has a limited capacity. One of the majo r challenges identified by 3GPP in supporting M2M communication is the potential for RACH overload, due to the significant increase in traffic load that will arise from large numbers of M2M devices. A number of solutions to this problem have been suggested by the 3GPP but they involve significant changes to the standards . In this paper, a relatively simple approach is presented that can enhance the capacity of the access channel through the use of Q-learning for M2M traffic. A key benefit of this approach is that it does not require any changes to the existing cellular network standards. We propose a solution that avoids RACH overload in supporting M2M traffic over existing cellular networks. The primary aim of this paper is to demonstrate how H2H and M2M traffic can effectively share the RACH of a cellular system by using Q-learning to control M2M traffic, through the D. Grace is with the
doi:10.1109/tvt.2018.2852952 fatcat:ex756nwytrggbd2qnvnxvhoyk4