关键词：传感器；自主导航；控制系统；计算机仿真
摘 要：Military conflicts are shifting from jungles and deserts to cities. This is because terrorists, insurgents, and guerrillas find that these areas provide a rich target environment and good hideouts. With the use of UAVs, urban threats can be effectively tracked and targeted. However, in urban environments where there is little or no GPS signal and many obstacles, navigation of UAVs is a major challenge. Multiple UAVs can be employed to share sensor information to counter these challenges and to perform Intelligence, Surveillance, and Reconnaissance (ISR) missions with greater ground coverage and better success rates. This thesis explores the various types of UAVs deployed for urban operations, and investigates trends in the design of such UAVs in terms of their weight, altitude, speed, and sensor suite. The thesis discusses the challenges and requirements for interoperability of multi-UAVs in urban environments, and proposes a direct-method-based control system for multiple UAV collaboration and obstacle collision avoidance. A dynamic model was developed for the simulation testing of the control system algorithm, which was followed by a physical experiment in an indoor environment using Quanser QBall-X4 UAVs to evaluate the results. Results show that the UAVs were able to share and integrate their sensors' information for joint cooperation.