kumoh national institute of technology
Networked Systems Lab.


NSL > Education> Correction
Deep-reinforcement Learning for Traffic Signal Optimization in Edge-based Traffic Control System
By : Riesa Krisna Astuti
Date : 2019-10-07
Views : 205

The road intersection problem is a challenging issue of intelligent transportation systems (ITSs), which needs to be addressed. The existing traffic control system (TCS) still cannot solve common problems that occur at intersections. For instance, road intersections are prone to traffic congestion. The traffic condition is worst when traffic accidents occur at road intersections. The current TCS also does not have the capability to prioritize emergency vehicles (e.g., ambulance, police car, fire truck) during urgent situations. To overcome these issues, this paper proposes a deep reinforcement learning (DRL) algorithm to optimize traffic-light signals under three traffic conditions (traffic congestion, emergency vehicle, and vehicle collision). Furthermore, the DRL algorithm is implemented in the edge computing-based TCS. The communication latency and computation process can be reduced by leveraging edge computing in the TCS. The results demonstrate that the proposed algorithm can reduce vehicle waiting time in intersections under the three mentioned conditions.
Attachment 1:   ÷ KakaoTalk_File_20191007_1038_06265.doc(55.2KB)