User Scheduling Based CLI Mitigation Scheme for NAFD Cell-Free Systems Toward mURLLC

In network-assisted full-duplex (NAFD) systems, where uplink (UL) and downlink (DL) communications occur simultaneously, the emergence of massive ultra-reliable low-latency communications (mURLLC) with high density of mobile users brings about severe cross-link interference (CLI) from UL users to DL users. To analyze and mitigate the impact of inter-user CLI, the closed-form expression of decoding error probability is derived and a novel user scheduling based CLI mitigation scheme is designed to jointly maximize the number of served users and system service rate, which determines how many users should be admitted by the network in a timeslot and how the users should be split into groups. To address this optimization problem, we explore an algorithm based on a Deep Q Network (DQN) with low complexity. Numerical results validate the effectiveness of the proposed solution. Moreover, the proposed scheme provides up to 33.9% better rate performance when compared to random user scheduling while guaranteeing the quality of service (QoS) requirements and power constraints.