Handover-Aware Cross-Layer Aided TCP With Deep Reinforcement Learning for High-Speed Railway Networks
Jianpeng Xu, Bo Ai, Lina Wu, Liangyu Chen · IEEE Networking Letters · 2020
Realizing high data rate transmission over the high-speed railway (HSR) networks is the goal that we are striving toward. Unfortunately, existing TCP fails to distinguish between negative effects (such as packet loss and round-trip time increase) induced by congestion and those induced by handover, leading to severe performance degradation in the HSR scenario. To tackle this problem, this letter develops a new cross-layer aided congestion control method based on deep reinforcement learning, namely RTCP, to alleviate performance degradation problems caused by handover. Simulation results illustrate that the RTCP can achieve higher goodput than existing TCP schemes in HSR networks.