Task Allocation Technology Of Unmanned Aerial Vehicles For Emergency Relief In An Urban Terrain

Unmanned Aerial vehicle(UAV) has strong ability to perform multi tasks with low cost and high performance. Many natural calamities and emergencies exist in the complex urban terrain. The application of UAV in emergency relief has broad prospects. UAV collaborative task allocation can increase the efficiency of cooperative mission for UAV group. This thesis has studied the UAVs task allocation technology for emergency rescue in an urban terrain.First of all, a simple mathematical model of urban terrain is established. The requirements of emergence rescue and the characteristics of applied load are analyzed. Furthermore, the multi tasks including search, classification, disposition and monitor are introduced. The constraint relations between rescue tasks are analyzed.Secondly, a hierarchical task allocation of Mother Unmanned Aerial vehicle(MUAV) and Small Unmanned Aerial vehicle(SUAV) is proposed to solve the problem of multi-UAV applications in an urban terrain. The influence factor on task success rate of SUAV is added to the objective function to build the task allocation model inside the detection range. The particle swarm optimization(PSO) algorithm is presented to solve the problem. Concave function decreasing strategy of inertia weight and search operator of artificial bee colony(ABC) are added to the iterative process of ABC-PSO algorithm. The simulation results show that the model and the algorithm can effectively solve the hierarchical task allocation for heterogeneous multi-UAV in an urban terrain.Thirdly, multi base UAV task allocation model and task balance constraints are analyzed and established based on the urban environment division. The ABC-PSO algorithm is used to solve the task allocation problem. The simulation results verify the rationality of the model under the designed constraints.Finally, in order to solve the dynamic task allocation problem, a dynamic hierarchical task allocation system is designed based on the auction algorithm. The simulation results from different emergency conditions show that the algorithm is effective and easy to implement.