Fast Exact Redistributed Pseudoinverse Method for Linear Actuation Systems

Abstract

Due to simplicity and real-time capability, the cascaded generalized inverse (CGI) method is a common control allocation scheme. However, depending on the specific problem, CGI may provide only an approximate solution and the available control volume cannot be fully used in most cases. In this brief, a CGI variant is established, which provides an exact solution for attainable control demands. The approach is based on the repeated usage of the pseudoinverse. A direction-preserving increment toward the control demand is achieved in each iteration step by saturating a single actuator. Control allocation examples for an F-18 airplane demonstrate that the proposed scheme delivers improved results compared with the original CGI method while requiring similar complexity.