Intelligent Computation Offloading and Trajectory Planning for 3-D Target Search in Low-Altitude Economy Scenarios

Abstract

The low-altitude economy provides new opportunities for target search, as autonomous aerial vehicles (AAVs) improve search efficiency through aerial surveillance and data collection. AAVs support overhead views of the search area, while ground vehicles (GVs) offer side views, facilitating effective collaboration in wireless networks. However, the limited battery life restrict AAVs from performing computation-intensive and latency-sensitive tasks, and their wireless communication links are susceptible to jamming attacks. Therefore, this letter proposes a joint resource scheduling approach for an edge computing enabled multi-AAV multi-GV cooperative target search system under intelligent jamming attacks. Specifically, the approach aims to minimize the uncertainty of the search area by jointly optimizing ground base station (GBS) association, task offloading rate, and 3D trajectory control. Since the problem is non-convex and the intelligent jamming behavior is dynamic, a multi-agent softmax deep double deterministic policy gradients (MA-SD3) approach is proposed to effectively perform resource management for target search and resist jamming attacks. Simulation results demonstrate that the proposed approach outperforms the baseline approaches under different settings.