Application of independent component analysis in target trajectory prediction based on moving platform

Abstract

In Electro-Optical tracking systems, compound control is used to keep high-precision tracking of fast targets by predicting the trajectory of the target. Traditional ground based Electro-Optical tracking system uses encoder data and target missing quantity read from image sensors to achieve the target trajectory by using prediction filtering techniques. Compared with the traditional ground based systems, relative angle between the tracking system and the ground cannot be read directly from encoder data in an Electro -Optical tracking system based on moving platform. Thus the combination of inertial sensors’ data and target missing quantity is required to composite the trajectory of targets. However, the output of the inertial sensors contains not only the information of the target's motion, but also the residual error of vibration suppression. The existence of vibration suppression residual error affects the trajectory prediction accuracy, thereby reducing compensation precision and the stability of the compound control system. Independent component analysis (ICA) method that can effectively separate the source signals from the measurement signals is introduced to target trajectory prediction field in this paper. An experimental system based on the method is built by settling a small dual-axis disturbance platform, which is taken as the stable platform, on a large dual-axis disturbance platform, which is used to simulate the motion of the moving platform. The result shows that the vibration residual is separated and subtracted from the combined motion data. The target motion is therefore obtained and the feasibility of the method is proved .