Differential evolution strategy for over-actuated control allocation problem with nonmonotonic nonlinearities

Over-actuated systems with redundant actuators often arise in aerospace, automotive vehicles, marine vessels, and robotics applications. Due to this redundancy, control allocation methods are typically utilized to distribute control effort amongst many actuators to achieve a desired effect. Traditional methods assume that a perfectly linear relationship exists between the control moments and the effector deflections. However, it is intrinsically nonlinear, especially in the event of failure of one or more actuators. In this paper we consider the constrained control allocation problem with nonmonotonic nonlinearities for over-actuated systems. To solve this problem, we propose a differential evolution (DE) strategy. The performance of proposed scheme is compared with that of other approaches for an aircraft model, including a cascaded generalized inverse (CGI) approach, a direct allocation (DA) method, and a fixed-point algorithm (FXP). The methods are first compared using open-loop measures such as the ability to attain desired moments for a prescribed maneuver. Then the methods are compared in closed-loop with a basic control law under failure conditions. The simulation results show the effectiveness of proposed approach.