ACB scheme based on reinforcement learning in M2M communication

Machine-to-machine (M2M) communication is a networked application and service taking the intelligent interaction of machine terminals as the core. However, the existing access control scheme cannot adapt to the rapidly changing access conditions during the long-term evolution process, resulting in a sharp decline in the success rate of massive M2M devices including delay tolerant devices (DTDs) and delay sensitive devices (DSDs). Quality of Experience (QoE) refers to the subjective experience of the quality and performance of devices, networks and systems, applications or services. In this paper, QoE with delay mapping is used as the simulation target. To solve the problem of adaptability, the paper proposes a access class barring (ACB) scheme based on reinforcement learning, which can be recorded as RL-ACB. The DTD, DSD barring factors and DSD scaling factor were adjusted to ensure the random access of massive M2M devices including DTDs and DSDs, and reduce the delay as much as possible. The simulation results verify the effectiveness of the scheme.