Summary
In recent years, coordination among a group of autonomous aerial vehicles have been a topic of longstanding interest. These autonomous aerial vehicles in a multi-agent system setting, may have to accomplish predefined missions and operate in hostile and dynamically changing environments, without continual human guidance. Moreover, these aerial vehicles, also known as agents, need to operate as a team and perform coordinated tasks to accomplish the joint mission. Some of the applications of the cooperative tasks for autonomous aerial vehicles include surveillance, reconnaissance, and battle damage assessment. Each agent decides on a proper action to achieve the group objective using multiple inputs, e.g. history of its own actions, interactions with its neighboring agents, and its goal. In contrast with a single agent system, multi-agent systems are expected to benefit accuracy, efficiency, robustness, flexibility, and reliability. However, these agents need to share the information among each other through wireless communication networks which exposes them to an additional layer of faults and uncertainties. This demands for the design of cooperative control strategies which are robust to uncertainties in the communication network, unknown input disturbances, communication delays, parametric uncertainties, unmodeled dynamics, among others. In general, this project develops technical approaches to overcome these challenges. Specifically, mathematical and computational frameworks are developed to characterize the stability and robust performance measures of the cooperative multi-agent system, in terms of stability margins and time delay margin. Further, this work designs a distributed control protocol for the linear multi-agent system subject to non-uniform time-varying delays to achieve consensus. Finally, state feedback control strategies are designed for the cooperative autonomous systems with probabilistic uncertainties to minimize expectation performance indices.
