Situation Estimation Method Of UAV To Ground With Multiple Maneuvering Targets

In recent years,drones have developed rapidly in the fields of logistics inspection,aerial photography,and plant protection,relying on a series of features such as easy control,low cost and low risk.In many cases,the behavior of the drone to the ground is used.In a specific task,it is necessary to analyze the situation of the whole system to ensure that the drone is more effective in manipulating the ground target.This article will take the situation estimation.analysis.Firstly,the requirements of the situation estimation system are analyzed and the system design is carried out.The composition and operation environment are introduced respectively.On the basis of the completion of the whole system design,the situation model is established,the situation elements are selected,and finally the ground target is selected.Trajectory prediction is carried out.Since effective target tracking and obstacle avoidance are the key to the success of UAV missions,the extended Kalman filter(EKF)is used to estimate the motion state of the object detected by the drone from the air.The estimated target state is estimated using a Kalman filter motion model to predict the optimal target trajectory.Secondly,for the situation estimation system,multiple objective functions are mutually constrained,and a multi-objective particle swarm optimization algorithm is proposed.Because of the slow convergence and easy to fall into local optimum in solving multi-objective optimization problems A multi-objective particle swarm optimization algorithm based on Gaussian mutation and adaptive reference point fusion is proposed.The Gaussian mutation location update method is used to improve the premature phenomenon of the solution,and the multi-objective particle swarm optimization algorithm is used to search for solutions in the optimization process.Diversity,using an external file maintenance strategy with adaptive reference points,eliminates poorly convergent particles and improves the convergence of the algorithm.Finally,taking one-to-many as an example,based on the experimental measurement,the situation of the whole system is classified by the situation,and the fuzzy C-means clustering algorithm is used to cluster the selected samples and the clustering result data.Based on the semi-supervised naive Bayesian classification,an improved algorithm based on data classification is proposed.The algorithm is used to classify the data of the control system and improve the classification performance.