Optimal and suboptimal performance of stochastic linear tracking systems

Abstract

The derivation of performance indices for both optimal and suboptimal linear tracking systems is organised so that the expressions for the indices consist of terms that may be calculated from the system input and its parameters and which represent separately various identifiable components of the overall cost. The derivation of the performance index for optimal linear tracking systems is necessary to give a complete solution to the optimal linear tracking problem. A knowledge of the various cost terms in the performance index of suboptimal linear tracking systems means that the performance of existing or approximate designs, or of optimal designs either modified or having different noise inputs or performance indices, may be calculated. Further, the improvement of existing or approximate designs may be carried out by a systematic trial-and-error approach, which aims at reducing perhaps only the largest cost terms for each iteration.