Model Reference Adaptive Control - A Perspective

Three adaptive control structures are discussed and compared. Similarities between all three methods are shown. First, the plant control law is the product of an adaptable gain matrix and a vector of observations. Second, the time rate of change of the gain matrix is equal to a product of a constant gain matrix, the error vector, and the transpose of the observation vector. All these methods are based on the minimization of a quadratic error function, similar to Widrow algorithm used in adaptieve filtering. The robust application of adaptive control is also discussed.