The Research Of The Tracking System Of Space Object Trajectory Based On ANN

The combination of the intelligent control theory and the trajectory tracking technology is an emerging research direction, and this paper has made a new attempt under the background. Based on the predecessors research, the subject applies BP neural network control method that is relatively mature in the intelligent control theory to the tracking system of space target trajectory to implement the control of this system.Firstly, the paper introduces the tracking system of space target trajectory which is implemented by the three mutually perpendicular of coordinates target trajectory that is decomposed of the space target trajectory and this makes the question simple. It utilizes the UMAC controller that can operate quadric algorithm design to control servo motor, and depends upon three servo motors representing three-dimensional space coordinates axis to drive the ball bearing lead screw organization to realize the specific target trajectory tracking. The servo amplifier of the experiment platform works under the pattern of torque, namely, the input value of its mathematical model's space state equation is the value of torque, but the output of UMAC is the pulse and orientation and it cannot cause the servo amplifier to obtain the input of torque directly. So the torque can not be controlled directly in the experiment while the velocity can be controlled. Therefore, data synthesis is established between the torque and the velocity via the BPNN in order to learn and train it. An approximate functional relation is obtained through the data synthesis after using three layers of BPNN, which provide the important basis for the following experiments.Secondly, it introduces the M function of training BPNN that is made the COM module by the COM generator of MATLAB, which is transferred by VB as the upper formation server. MATLAB, which is as the engineering calculation's core, realizes the data analysis and simulation. MATLAB and VB are unified by the COM module by way of enhancing its strong points and avoid its weaknesses to realize their seamless integration. The superior management software is composed of the module of initialization, the module of the trajectory real-time movement, the display, the module of data storage and analysis that is developed by VB. The inferior real-time control module is composed of the module of the interpolation, the module of the servo driving, the module of the procedure explanation and the module of the data high speed gathering that is developed by UMAC. The communication between the superior PC and the inferior UMAC is through dynamic linking storehouse PCOMM 32.DLL of UMAC s own.Finally, the concrete steps are given in the experiment according to the experiment principle, such as the choice of movement pattern, the setup of parameter and data gathering. This paper establishes the experiment program to run, gathers experiment data and analyzes it, elaborates the reason of the error and put forward how to enhance the system's precision.The system has realized the control function of self-contained software and hardware and has a strong character of real time and linkage. In addition, it has a high openness, which enhances its universality and transplant ability, and also makes for its function expansion and upgrade.