Completion Time Minimization for ISAC: Joint Trajectory Planning and Power Optimization

Abstract

While trajectory planning for integrated sensing and communication (ISAC) has been widely studied, existing frameworks often neglect time efficiency by assuming fixed mission intervals. To address this, we propose a deep reinforcement learning (DRL)-based joint trajectory and power allocation strategy designed to minimize the travel time to a designated endpoint. We define a novel cumulative sensing metric, named path exposure, derived from the probability of detection (PoD) under a majority-vote fusion rule with coordinated base stations. We formulate this time-optimal control problem as a Markov decision process (MDP) and employ a DRL-based solver to optimize mobility and resource allocation simultaneously. Simulations demonstrate that our approach achieves superior time efficiency over conventional baselines while ensuring all ISAC requirements are met.