Microprocessor Based Near Minimum Time DC Motor Position Control

摘要

This article presents the design procedure, discrete time simulation and digital signal processor (DSP) implementation of a near time-optimal control of a DC motor positioning system. A general design procedure for state feedback pole placement is presented to achieve an almost time-optimal response constrained by a system with forward loop saturation. The control scheme uses discrete time state feedback in a DSP. A nonlinear algorithm to prevent integrator windup is given. The device to be controlled is a small permanent magnet DC motor with a spring load; however the design procedure is applicable to a general class of motor positioning control problems. The sample rates required (400 KHz-1 MHz) for the chosen problem are within the capabilities of many DSPs. The system is compared to true time-optimal bang-bang control where much higher sample rates would be required. Other nonlinear effects, including current limits and quantization of feedback signals, are incorporated. Lookup tables for feedback gains in the DSP are used to maintain near optimal response for a variety of reference position input values. Simulation results are presented as well as control implementation and timing results for a TMS320C40 DSP.