A Disturbance Attenuation Approach to Missile Guidance and Control.

Abstract

Abstract : Solutions to stochastic control problems currently do not produce mechanizable missile control laws. To fill the gap without making excessive structure assumptions, it is suggested that the disturbance attenuation problem be extended to nonlinear guidance and control problems. The deterministic approach, which does not include all the properties of the stochastic control solution, does have characteristics which are quite appealing. For example, certain classes of partial information disturbance attenuation problems can be solved numerically essentially because variation methods are available. Over the three year period of the grant robust and adaptive guidance and control laws which are mechanizable with near future computer technology are developed which can meet system objectives in the presence of large uncertainties, system structural changes, and nonlinearities. Of particular importance emerging from our focus on disturbance attenuation is a new structure for adaptive control, new detection filters for detection and identifying structural changes, and methodologies for including system nonlinearites. Finally, a new multiple hypothesis adaptive estimator, using a single linear filter whose parameters are changing according to the on line computation of the probability of each hypothesis conditioned on the residual history, is shown to have better or equivalent performance to the current bank of filters concept with dramatic decrease In computation and ease of implementation.