Analysis and design of the receding horizon constrained optimization for class-D amplifier driving signals

This paper reports the analysis and design method of receding finite horizon constrained optimization approaches to generate driving signals for digital amplifiers. The performance of a digital amplifier that is based on power MOSFETs highly depends on the modulator that modulates the reference signal into binary sequences. The concept of receding finite horizon constrained optimization modulator is to generate the binary output in the sense that the power of filtered error is minimized. The stability analysis of this nonlinear feedback system is derived in detail and the method to estimate the bound of filtered error of the high dimensional quantization is proposed. A second order system with horizon length one and two is designed to produce a 1.5-bit constrained output. Simulation results demonstrate the accuracy of the analysis.