ESA GNC Conference Papers Repository

The paper focuses on actuator commanding, in particular on the distribution of a command (e.g. torque, angular momentum) from three-dimensional space onto a number of individual actuators, here reaction wheels or thrusters. The conventional pseudoinverse (L2) approach is compared with the minimax (L-infinity) approach considering three realistic application cases: (1) maximum wheel array capability, (2) wheel offloading cycle, and (3) propellant consumption. It is shown that, compared to L2, with the L-infinity approach a higher agility (larger torque or momentum capability of an actuator array), a larger wheel offloading cycle and less propellant consumption is achivable without changing actuators. In addition, the presented algorithms have predictable outputs, i.e. they are to be used for on-board application.