Research On Perceptual Technology For Collaborative Detection Of Multiple UAVs

For multi-UAV cooperative detection,collaborative sensing can bring greater detection range,higher detection accuracy and better robustness.At the same time,it also brings challenges to information fusion and management of multi-sensors.How to efficiently utilize multi-sensor resources on a high-speed mobile platform such as drone,how to quickly integrate sensor measurement data to obtain more accurate state estimation.It is an important problem that needs to be solved to achieve multi-UAV collaborative detection.In this article,multi-UAV cooperative target tracking and target recognition are taken as research objects,and UAV is regarded as mobile aerial sensor carrier From the perspective of sensor operation mode,the research of multi-sensor information fusion technology and multi-sensor adaptive management technology for multi-UAV cooperative detection is carried out.Modeling and analysis of target tracking problem in the process of multi-UAV cooperative detection of ground maneuvering targets,a multi-UAV target tracking data fusion method based on Lyapunov navigation vector field guidance was proposed.The method uses Unscented Kalman Filter(UKF)and Unscented Rauch-Tung-Striebel(URTS)fusion estimation algorithm.Estimating the current position information of the ground maneuvering target by combining the direction information of the target obtained by the drones in different locations,and then guiding the control of the drones by combining the Lyapunov navigation vector field.And the phase control method is used to avoid the collision of the drones due to convergence in the same limit cycle.Thereby,a dynamic data fusion process is formed,and the independent cooperative tracking of the multi-UAV to the ground maneuvering target is completed,partially solved the problem of tracking instability of multi-UAVs on ground moving targets.Aiming at the problem of multi-sensor resource dynamic allocation in multi-objective cooperative multi-target process,multi-sensor adaptive management technology was developed.From the perspective of sensor working mode,a multi-UAV adaptive management method is proposed.From the perspective of sensor working mode,a multi-UAV adaptive management method is proposed.The method uses the Bayesian network for statistical analysis of historical detection data,and uses Bayesian decision theory to construct the target recognition model.The perceptual information entropy is used to measure the uncertainty of the target category,and the fuzzy discretization method is used to process different types of feature data,and the multi-UAV multi-target assignment problem is solved based on genetic algorithm.Therefore,a multi-UAV detection loop of "perception-learning-decision-action" is constructed,and finally the optimal allocation of sensor resources is realized.In order to verify the effectiveness of the two algorithms proposed,a multi-UAV cooperative detection simulation system is designed and implemented.The simulation system implements two application scenarios:target tracking and target recognition.Based on this system,the performance of the proposed two algorithms is analyzed,and the effectiveness of the proposed two algorithms is verified.