Direct pole placement without using observers

A new and direct approach to solve the pole placement problem for a given linear plant is reported in this paper. The conventional procedures involve complete identification of the plant and construction of either an Asymptotic or Kaiman Filter type observer which use the estimated plant parameters. The estimated state variables are then fed back to place the poles at the desired locations. In the new method suggested in this paper, a simple model is constructed based on specified pole locations. A minimum number of coefficients related to the zeros of the plant and state feedback are determined together by a composite identification algorithm. When the plant input/output are corrupted by noise, it is possible to obtain accurate estimates of the parameters by using a periodic probing signal. The new procedure discussed involves minimum amount of computation for solving the direct pole placement problem. The core of the model itself takes the role of an observer.