Mission-based optimal control of Stewart manipulator

Purpose - The purpose of this paper is to show the merit of using mission information in tuning the controller gains for Stewart manipulator instead of the generic inputs previously developed in literature. Design/methodology/approach - The paper introduces two optimization techniques based on mission information. The first technique, a partial-information technique, uses gain scheduling that applies different controllers for different mission tracks. The second technique, a full-information technique uses a single robust controller by considering the full mission data. For demonstrating these techniques' feasibility; a nonlinear numerical simulation for a Stewart manipulator was built and tested using a generic mission. This mission consists of two piecewise trajectories (tracks). The proposed techniques were compared with one of the previous optimization techniques in literature, no-information technique, in which a step response is used to search for optimal controller gains without any information about the mission. Genetic algorithms were used to search for the optimal controller gain in each case with different cost functions. Findings - Based on the numerical simulations, the proposed mission-based optimization techniques have superior performances compared with no-information technique. Research limitations/implications - The proposed techniques were applied in a joint space or for a decentralized control. The work can be extended to be applied in a task space or for a centralized control. Originality/value - The paper proposes two novel optimization techniques: partial- and full-information techniques for tuning the controller gains of a Stewart manipulator, where mission information was imbedded into the cost function. These two techniques are generally applicable for other nonlinear systems such as aircraft stability and control augmentation systems.