New pole assignment algorithm with reduced norm feedback matrix

A new algorithm is described for the assignment of closed loop poles in linear time invariant multivariable systems. The approach is similar to the well known dyadic pole placement methods, but does not usually result in a unity rank controller. The algorithm can be put into iterative form in the sense that open loop poles can be relocated or preserved so that by repeating the assignment process, all of the open loop poles can be reassigned. There are only very mild constraints on the destinations of the poles. The paper also shows how any unused degrees of freedom can be exploited to reduce the control effort needed to achieve the pole assignment. Results are given for both the state and the output feedback cases.