Optimized Tracking Control of AAV-AGV Swarm Under Mismatched Disturbances and Byzantine Attacks

作者：Shixun Xiong, Xiangpeng Xie, Guo‐Ping Jiang, Yong Ren, Yang Liu · 发表于：IEEE Internet of Things Journal · 年份：2025 · DOI：10.1109/jiot.2025.3555660 · 被引用次数：3 · 研究领域：Distributed Control Multi-Agent Systems、UAV Applications and Optimization

This article explores the optimized tracking control problem of autonomous aerial vehicle (AAV) and autonomous ground vehicle (AGV) swarm with mismatched disturbances and Byzantine attacks. Unlike traditional AGV-AAV models with different orders, multiple internal and external disturbances within the physical structure are considered in a novel second-order AAV-AGV swarm, which constructs the mismatched disturbances. Since the external environment may induce the swarm to generate and propagate false signals (called Byzantine attacks) to its neighbors, this article reinterprets it as the management of unknown variables in the control inputs. Then, a reinforcement-learning-based approach is introduced to address the unknown control inputs generated by Byzantine attacks. In conjunction with the Hamilton-Jacobi–Bellman (HJB) equation, an adaptive actor-critic–identifier (ACI) structure is designed using the gradient descent method. With the unknown terms estimated by ACI, an optimal robust control strategy under mismatched disturbances is proposed, and a pivotal scaling technique is employed for formation stability analysis. Finally, simulations and experimental results are performed to demonstrate the efficacy of the proposed approach.