Research On Trajectory Prediction Technology Of Mission-oriented Cooperating Unmanned Aircraft

This thesis takes the mission-oriented cooperating UA as the object,and investigates the real-time prediction algorithm of the future short-term trajectory for UA by using the historical trajectory data and the real-time flight state data under typical mission scenarios.In this thesis,the proposed prediction algorithm is gradually improved,and the work can be summarized as follows:(1)In order to remain real-time prediction performance while taking full use of the historical trajectory data of UA,a short-term trajectory prediction algorithm based on RVFLNN model is proposed.After training by UA historical trajectory data,RVFLNN can grasp the flight trend of UA and make prediction for the position information of the next sampling moment quickly in real time.(2)For the performance of RVFLNN decreases rapidly along with the prediction step,based on the Kalman filter,RVFLNN is combined with the basic motion model dynamically as a combination prediction model.Comparing with RVFLNN,the combination prediction model has a better multi-step prediction performance when the motion model of combination prediction model matches the flight mode at next predicted time of the UA.(3)In order to match the multiple flight modes through whole flight process of the UA with prediction algorithm,the RVFLNN-IMM trajectory prediction algorithm is proposed based on IMM.RVFLNN-IMM can adaptively match different flight modes through the whole flight process.Then,a Softmax classifier is added to RVFLNN-IMM and the RVFLNN-IMM trajectory prediction algorithm based on Softmax classifier is proposed.By predicting the flight mode of the UA,a matching subset of motion models is selected and make prediction,which effectively improves the multi-step prediction performance of the algorithm for the whole flight process.The simulation results show that the RVFLNN-IMM trajectory prediction algorithm based on Softmax classifier can accurately predict the short-term flight trajectory of cooperating UA under typical mission scenarios in real time.This technology can be a support for delicacy management to UA in the future.