A Lyapunov-Based Approach to Balancing AoI and Management Costs in Cooperative Perception for V2I Networks

This paper proposes a Lyapunov-based approach to jointly optimize the age of information (AoI) and management cost in cooperative perception within vehicle-to-infrastructure (V2I) networks. In a hierarchical network with a base station (BS) and multiple roadside units (RSUs), the BS aggregates sensor data to generate cooperative perception content (CPC), which is then cached at RSUs and provided to vehicles upon request. We formulate an optimization problem to minimize both content generation and update costs while ensuring that the AoI does not exceed its maximum threshold to maintain real-time road conditions. The problem is reformulated using the Lyapunov drift-plus-penalty method and modeled as a Markov decision process (MDP). We employ a reinforcement learning algorithm, specifically proximal policy optimization (PPO), <br/>to find optimal policies for content management. The PPO-based approach maintains the AoI within acceptable bounds while reducing the management cost, thereby providing a scalable solution for dynamic vehicular networks. Simulation results demonstrate that our proposed method outperforms baseline approaches in balancing AoI and cost.