Incremental Control Allocation with Axis Prioritization on the Boundary of the Attainable Control Set

Pseudoinverse-based control allocation provides a computationally efficient way of allocating desired pseudo controls onto actuator commands with bounded run-time. In the presence of actuator saturation, the cascaded generalized inverse is often used to iteratively allocate the remaining pseudo controls and improve use of the control volume. Direction-preserving versions further avoid the problem of large angular deviations by scaling of the pseudo control vector. However, when used in an incremental control setting, the requirement of preserving direction can lead to fast deterioration of control authority close to the boundary of the attainable pseudo control set. An approach is proposed that solves this problem for incremental control allocation. The method is based on a direction-preserving pseudoinverse approach to retain the desirable properties of bounded run-time and to avoid potentially arbitrary local solutions due to actuator rate saturation. The approach is then extended by a step-wise reduction of dimension to move along the boundary of the attainable pseudo control set in a non-direction-preserving way, amounting to a left pseudoinverse, to retain control authority. Two versions are presented that either minimize the pseudo control error or prioritize control around specific axes.