RELATIVE STABILITY MARGINS OF MULTIVARIABLE SYSTEMS - CHARACTERISTIC LOCUS APPROACH

Recent work has proven that characteristic locus plots form the natural medium for the generalization of the Nyquist approach. In tho present paper these plots arc used to extend classical scalar techniques of assessing relative stability margins to the multivariablc caso. Thus the estimation of closed-loop poles using curvilinear squares is first discussed and subsequently the use of constant dynamic magnification circles in predicting performance under feedback is considered. A new concept relevant to both techniques is introduced, namely that of interference. Interference relates to the loop distribution of eigenvalues, and complements the concept of interaction which relates to eigenvector distri bution. © 1979 Taylor & Francis Ltd.