| Servo turntable is an essential basis of radar, radio telescopes and test turntable. It plays an important role in the areas of aviation, military and communication. The preparation technology and control level have great effect for the performance of the whole machine. The traditional control method is confined to the linear model and PID control scheme, which unable to meet the new requirement from the rapid development in various fields. So it has theoretical and practical value to research on nonlinear modeling methods and nonlinear control algorithm.In this thesis, two types of radar antenna servo turntable are treated as research objects. The main work is as follows:At first, the fuzzy control method is combined with PID controller, and the feed-forward control algorithm is introduced to improve the system performance. The obtained dynamic control algorithm is applied to XX1type turntable. Good over-top tracking effects illustrate the effectiveness and simplicity of the control algorithm.Secondly, the mechanism modeling and nonlinear influencing factor of XX2type turntable are analyzed, and the Hammerstein model is introduced to characterize the input-output relationships. The system identification experiments are designed to get the useful field data. We use the Particle Swarm Optimization (PSO) method to deal with the experimental data, the Recursive Instrumental Variables (RIV) is applied to get the initial value and search space of the PSO. The modified RTV-PSO has the optimization ability, which is proved by the simulation and calculation of QF%.Then the Nonlinear Predictive Functional Control (NPFC) is proposed to the Hammerstein model. Based on the traditional predictive functional control, treat the nonlinear.subsystem as media subsystem and the media variable as nonlinear optimal object. With the overall calculation of Genetic Algorithm (GA), the global optimized controller is formed. This control scheme has little computational burden, also has the ability to reduce the error in normal two-step method without constrain on the structure of nonlinear mathematical model of Hammerstein system.Finally, the switching control algorithm is introduced to the motor driving servo turntable because of its multi working situation, such as, zero crossing distortion, limit cycle, dead zone, etc. Here, we choose several different subsystems to describe the input-output relationships in different working states, it will make basis of multi-model switched controller. In the parameters identification, the subsystems modeling problems are the multi objectives, besides, the smoothness of switching are constrains. The whole model is described as the Constrained Multi-Objective Problem (CMOP). Two methods are proposed to the solution the CMOP. The first one is single object Comprehensive Learning Particle Swarm Optimization (CLPSO), in which, all of the particle positions are learned to renew the particle. The second one is the heuristic multi-objective PSO method based on the Pareto optimization. The simulation and experimental results demonstrate the effectiveness of both identification algorithms.The work of the thesis is funded by the Technology Project from a military institute,’Simulation and modeling for tracking radar servo system’, and the Science Foundation of Anhui University,’The dynamic intelligent control algorithm of radar servo system’. |
