Joint Latency and Energy Consumption Optimization with Deep Reinforcement Learning for Proximity Detection in Road Networks

Artificial intelligence and the automobile business have grown significantly in recent years, and autonomous driving has progressively been the industry's focus. The problem of proximity detection in road networks refers to determining if two moving objects are close to one other in real time. However, in the real world, mobile devices' battery life and computing capabilities are restricted, resulting in excessive latency and energy usage. As a result, determining the proximity relationship between mobile users with low latency and energy usage is a difficult task. We formalize the joint latency and energy consumption optimization problem for proximity detection in road networks into a constrained optimization problem (COP) and solve it using a deep Q network (DQN). The simulation results show that DQN can ensure the desired sum cost under a variety of important factors, effectively reducing latency and energy usage.