kumoh national institute of technology
Networked Systems Lab.

Sanjay Bhardwaj, Dong-Seong Kim ,"Deep Q-Learning Based JT-CoMP for Resource Allocation Enabling eURLLC in Industrial Wireless Networks", Wireless Networks,2021, (Pr).
By :
Date : 2021-02-03
Views : 125

Ultra reliable low latency communication (URLLC) have received considerable attention as a promising candidate of 5G. Enhanced URLLC (eURLLC) is the enhancement of URLLC in terms of reliability, introduced in Release 16 of 3rd generation partnership project (3GPP). Joint transmission coordinated multipoint (JT-CoMP) can provide high reliability for interference limited cell-edge user equipment (UE), where UE is simultaneously connected to multiple base stations. In industrial wireless networks for the enhancement of reliability in URLLC, a decentralized deep Q-learning based JT-CoMP resource allocation scheme is proposed, called as eURLLC. In the proposed method, agent corresponds to the link between base station (BS) and UE, B2U and each agent makes its own decision for the optimal allocation of the resources. Simulation results show that each agent successfully learns within the reliability and latency constraints about the dynamic environment, thereby minimizing the effect of interference resulting in better error probability and throughput for eURLLC in comparison to traditional communication system (CS) and URLLC. The fairness index further confirm that the proposed method achieves better fairness in comparison with random allocation.