Abstract:

In the case of limited measurement information, Kalman filter (KF) is difficult to deal with unmanned aerial vehicle tracking by using a single motion model. To solve this problem, a novel long short-term memory(LSTM)-KF algorithm combining LSTM and KF algorithm is proposed. First, LSTM is used to predict the average and instantaneous velocity of the target so that the problem of poor generalization ability of nonparametric model can be solved in position prediction task. Then, the prediction limitation of KF algorithm using motion model is analyzed, and the method of using LSTM prediction results to modify the prediction results of motion model is proposed to reduce the prediction error. The revised prediction results are combined with the measurement data to realize the state estimation of the target. Finally, the proposed algorithm is verified on the generated trajectory. The simulation results show that LSTM-KF algorithm has higher tracking accuracy and stronger robustness than the existing models.

Key words: unmanned aerial vehicle,  long short-term memory, Kalman filter, trajectory tracking