Research On Non-cooperative Target Tracking Technology Based On Inertial Prediction

Spaceborne laser ranging system for non-cooperative target has the advantages of high accuracy and long range,but due to the long distance and small size of the target,the imaging frame rate of the target is limited,which leads to the limitation of target tracking ability.Furtherly,the laser echo signal can be very weak,which affects the ranging performance.In order to enhance the system's active vibration suppression ability and improve the tracking accuracy of the system,the following innovative work has been made based on the benefits of improving tracking accuracy in laser ranging and the necessity of using adaptive filtering algorithms to achieve long-distance tracking:1.A tracking method based on adaptive filtering is proposed.The adaptive filtering algorithm is introduced,and the principle of data fusion based on adaptive filtering is analyzed.Aiming at the problem that the sampling rate of the tracking camera and the inertial sensors do not match,a down-sampling filtering method is proposed to realize the adaptive fusion of low frame rate(50Hz)camera data and high frame rate(1k Hz)inertial sensor data.The Affine Projection algorithm is used to solve the defects of slow convergence speed and poor steady-state accuracy of the classic adaptive filtering algorithm.2.The tracking algorithm of spaceborne long-distance non-cooperative target is implemented and a variable order adaptive filtering algorithm is proposed.Based on the working principle of the tracking system,predictive compensation is made for the delay existing in the tracking system,and a prediction accuracy of 1?rad is achieved when the delay is 5ms.In order to reduce unnecessary calculations,balance the convergence speed and steady-state accuracy,a variable order adaptive filtering algorithm is proposed.Compared with the fixed-order LMS algorithm,the convergence speed is doubled,and the steady-state accuracy is improved by 3 times.Fiber optic gyroscope data of the tracking system is filtered,and in order to eliminate the zero drift of the fiber optic gyroscope,the accelerometer is used to eliminate the zero drift of the fiber optic gyroscope,achieving a 20 d B suppression effect.3.Data simulation and system comprehensive experiments are carried out on the method proposed in this paper.Through data simulation,the micro-vibration model of the satellite were used to verify the feasibility of the method to realize the fusion of inertial sensor data and camera data;And a comprehensive system experiment was carried out on the actual tracking system to which the algorithm is applied,the experimental results show that the tracking accuracy is better than 5?rad under the vibration of the main interference sources present in the satellite,which meets the tracking accuracy requirement in the field of laser ranging of long-distance noncooperative targets.