Optimum Sampled-Data Systems with Quantized Control Signals

Optimum design of sampleddata systems with quantized control signals is considered. The control signals mi(k) are members of a set [-CL, ... -C2, -C1, 0, C1, C2, ... CL] where ci≫ 0, i= 1, 2,... L, and Ck≫Cj for k≫j. By selecting a proper set [ci], a design compromise between the extremes of proportional control systems and relay control systems is achieved. Dynamic programming techniques are used in carrying out the optimum design. Both minimum summed-square-error systems and minimum -N systems are discussed and illustrated for dynamic performance.