Realization of simultaneous perturbation stochastic approximation using FPGA for on-line PID parameter tuning

Simultaneous perturbation stochastic approximation (SPSA) algorithm has attracted considerable attention for challenging optimization problems. SPSA is a highly efficient gradient approximation that relies on two measurements of the objective function. On the other hand, field-programmable gate arrays (FPGAs) are glue logic and prototyping devices jumping to product. In recent years, FPGAs have become essential in the design of digital systems, complementary to the role played by microprocessors. This paper addresses realization of SPSA using FPGA for on-line parameter tuning for the PID controller. The voice coil motor (VCM) is used as a plant for illustrating the position control performance. Three stages are presented to develop a FPGA chip of the specific function, namely, Simulink simulation, hardware simulation and FPGA with hardware implementation. Experimental results show that realization of SPSA using the FPGA is very efficient for on-line application.