Based On Neural Network Pid Controller Dsp

Turntable for 863-2 dynamic-simulation-scan-imaging experiment is specially designing practicality simulation turntable for sifting and general-checking combining type scout camera; it is a typical system of servo control.Former people already do some work in designing aspect about low-speed control system, but it can't completely meet turntable's control precision request of low-speed moving. By analyzed former system of classical PID control-scheme and experiment data, The finding show dry-friction is important factor on the influence of low-speed stabilization during turntable do a low-speed moving, it is difficult to have upper static and dynamic capability by adopt traditional control-scheme. Aim at such nonlinearity friction moment moving, Sufficient combining neural networks and feedback control respective advantage, a sort of PID controller structure basing on neural networks was being used to overcome the dry friction. Traditional PID controller commonly is fixation parameter; moreover PID controller structure basing on neural networks can automatically find optimization PID parameter combination aim at protean friction moment. It also adds systemic self-adaptability.This paper beginning with friction moment compensate, basing on former hardware, analyzed control-object's low-speed characteristic, presents and demonstrates a new PID control arithmetic basing on BP neural networks. And combining the debug procession of single-axis simulation turntable goes into particulars.Through theory analysis,computer simulation testing,this paper sufficiently proves effective of PID control arithmetic basing on neural networks, this is a value reference for the design of control system on aftertime.