A Modified Simulated Annealing Algorithm for SUAVs Path Planning

Small Unmanned Aerial Vehicles are becoming very popular for several applications including precision agriculture, road and building inspections and aerial photography. One possible scenario for these sUAV is to use them in a cooperative flights. In this case one UAV finds Points of Interest and the others inspect them based on each UAV's capabilities. This scenario requires that the path planning algorithm is executed by the embedded processors such that the UAVs can collaborate directly. This paper presents a path planning algorithm based on Simulated Annealing for small Unmanned Aerial Vehicles. The objective of the algorithm is to find flight paths to allow the UAVs to inspect numerous points of interest placed over a region, considering that these points may belong to different kinds and should be inspected by a specific type of UAV. This is considered as a multiple Traveler Salesman Problem, solved by the proposed SA algorithm where the Energy function takes into account the mismatches between POI type and UAV specialty. The algorithm is implemented on a Raspberry Pi and tested using a ArduCopter platform flight simulator, considering two inspector UAVs. The results show that the algorithm is capable of calculating paths matching POI and UAV types with an execution time that is suitable for on-the-fly mission generation.