Collaborative Estimation And Decision Of Multiple UAVs For Ground Moving Target Tracking

In recent years,unmanned ground vehicles(UAV)have been increasingly used in various military and civil scenes,because of the rapid development of UAV technology.UAVs are gradually replacing human and playing an increasingly important role in attacks,tracking,search and rescue.For ground moving targets in urban environments,the collaborative estimation and decision algorithms of UAVs are studied in this thesis.An adaptive interacting multiple model algorithm based on the square root cubature Kalman filter is proposed to estimate and predict the states of non-cooperative ground moving targets.Considering the strong mobility and randomness of suspected targets,the interacting multiple model algorithm is used to estimate the states of targets,which improves the matching degree of the model.To reduce the computation,the square root cubature Kalman filter is combined with interacting multiple model algorithm.And the adaptive transition matrix is introduced to further enhance the accuracy of the hybrid model.Then,the predictive states of the targets are obtained by the hybrid model.And the correction strategy is designed to eliminate the interference of the accumulated error as much as possible.After that,the distributed consistency method with modified Metropolis weight is introduced,which is based on the distributed communication network.Finally,the effectiveness of the algorithm is verified in simulations.A multi-objective optimization model is designed for the three-dimensional path planning of multi-UAV-single-target tracking.The performance indicators including the maximum three-dimensional coverage of target,the minimum control cost,and the minimum auxiliary energy consumption of sensor is considered in this model.The fuzzy multi-objective function is established with the three indicators.And the relaxed order of satisfactory degrees is also established based on the different importance of each indicator.In addition,the obstacle avoidance constraints are specifically designed for obstacles and no-fly zones in the environment.Considered the synchronous distributed receding horizon control framework and the communication delay,the collision avoidance constraints are designed.Finally,the proposed path planning method was verified in the simulation.A dynamic integrated optimization model is designed for the task allocation of multi-UAV-multi-target tracking.According to the basic task allocation model,the priority of ground targets,and the event triggered principle specified to avoid redundant task allocation,the dynamic task allocation model is established.Based on the centralized communication network,the path planning model and the task allocation model are combined to form the dynamic integrated optimization model,which can effectively avoid the possible collisions caused by the coupling of path planning and task allocation.Then,binary hybrid particle swarm optimization is used to solve the above optimization problem.Finally,the proposed method is verified in the simulation,which proves that the proposed algorithm can greatly reduce the computation.