Abstract

This thesis investigates the decentralized receding horizon control (DRHC) for a network of cooperative vehicles where each vehicle in the group plans its future trajectory over a finite prediction horizon time. The vehicles exchange their predicted paths with the neighbouring vehicles through a communication channel in order to maintain the cooperation objectives. In this framework, more frequent communication provides improved performance and stability properties. The main focus of this thesis is on situations where large inter-vehicle communication delays are present. Such large delays may occur due to fault conditions with the communication devices or limited communication bandwidth. Large communication delays can potentially lead to poor performance, unsafe behaviour and even instability for the existing DRHC methods. The main objective of this thesis is to develop new DRHC methods that provide improved performance and stability properties in the presence of large communication delays. Fault conditions are defined and diagnosis algorithms are developed for situations with large communication delays. A fault tolerant DRHC architecture is then proposed which is capable of effectively using the delayed information. The main idea with the proposed approach is to estimate the path of the neighbouring faulty vehicles, when they are unavailable due to large delays, by adding extra decision variables to the cost function. It is demonstrated that this approach can result in significant improvements in performance and stability. Furthermore, the concept of the tube DRHC is proposed to provide the safety of the fleet against collisions during faulty conditions. In this approach, a tube shaped trajectory is assumed in the region around the delayed trajectory of the faulty vehicle instead of a line shaped trajectory. The neighbouring vehicles calculate the tube and are not allowed to enter that region. Feasibility, stability, and performance of the proposed fault tolerant DRHC are also investigated. Finally, a bandwidth allocation algorithm is proposed in order to optimize the communication periods so that the overall teaming performance is optimized. Together, these results form a new and effective framework for decentralized receding horizon control with communication faults and large communication delays.