Scheduling and Resource Allocation for Multi - Hop URLLC Network in 5G Sidelink

5G New Radio (NR) is envisioned to efficiently support ultra-reliable low-latency communication (URLLC) for new services and applications with high reliability, availability and low latency such as factory automation and autonomous vehicles. Multi-hop Device-to-device (D2D) communication is one such means that expands D2D coverage and achieves lower latency in the mobile edge and NR sidelink. In this paper, we first analyze the URLLC requirements in 5G and the multihop D2D communication problem with perfect knowledge of the network. Subsequently, we investigate the deep reinforcement learning (DRL) algorithm for the scheduling and resource allocation problem with only local information for each node. A simulation is employed to evaluate the performance of the related algorithms. Numerical results show that the proposed DRL algorithm outperforms the greedy algorithm in terms of different relay nodes between the source and destination, and is robust to the coming or leaving of relay nodes.