| Summary: | 碩士 === 中華大學 === 資訊工程學系 === 105 === Unmanned Aerial Vehicle (UAV) was originated for the military use. With rapid improvements of UAV technology, UAV has many practical applications in the civilian areas. Because UAV can overcome harsh environmental conditions and the limitations of ground transportation, more and more new missions are taking advantages of UAV, such as mountain patrolling, sending packages, environmental monitoring, disaster investigation, and searching and locating the escaped criminals. As a result, planning paths of UAV has become one of the major task in UAV applications.
In the literature of UAV path planning, most of the approaches are path planning without any obstacles. Only a few approaches proposed path planning considered static obstacles by avoiding the no-fly zone. However, those approaches only focused on path planning but ignored the major goal in regional monitoring. Therefore, this research investigates the application of UAVs with adjustable imaging angle in monitoring a dynamic obstacles environment, considering the objectives of path planning, UAV load balancing, and monitoring requirements.
In this paper, we proposed a multi-objective evolutionary approach to solve the investigated problem. A real-world railway environment is adopted to simulate the dynamic obstacles environment, UAVs are deployed for the orbit and peripheral monitoring tasks. The experimental results show that the proposed approach can effectively obtain multiple non-dominated solutions, which optimized the objectives of path planning, UAV load balancing, and monitoring requirements
|
|---|
