Effective identification of sliding mode control parameters employing linear quadratic regulator and evolutionary computation techniques. Issue 1 (28th November 2019)

• Record Type: Journal Article
• Authors: Banazadeh, Afshin; Khoshrooz Azad, Reza; Emami, Seyyed Ali
• Is Part Of: Advanced control for applications. Volume 2:Issue 1(2020)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2019-11-28
• Subjects: genetic algorithm -- linear quadratic regulator -- particle swarm optimization -- sliding mode control; Automatic control -- Periodicals; Automatic control; Periodicals; Electronic journals; 629.8
• DOI: 10.1002/adc2.21
• Languages: English
• ISSNs: 2578-0727

Abstract

Abstract: In this paper, a new technique to reduce the search space in evolutionary algorithms for operative identification of sliding mode control parameters is proposed, especially in the case of a rigid satellite. The reduced domain provides a solution closer to the actual optimum at a lower computational cost. The optimized sliding mode controller provides superior performance that eliminates the need for trial‐and‐error exploration. In this approach, a linear quadratic regulator is initially designed for the equivalent linear system. In order to reduce the convergence time and computational cost, the regulator history is then used as a reference to provide optimal space for the sliding mode control parameters. Consequently, boundaries of the optimal space are fed to the genetic algorithm to optimize the control parameters. On the other hand, the particle swarm optimization uses the midpoint of the optimal space as an initial guess. The validation process is performed by tuning the sliding mode control parameters for a rigid satellite, taking into account uncertainties associated with the inertia matrix. The simulation results suggest the convergence to the near‐optimal solution, and the stability and robustness of the controller are enhanced by this technique. Abstract : A new technique to reduce the search space in evolutionary algorithms for operative identification of sliding mode control (SMC) parameters is introduced. An LQR is first designed for the linearized model and subsequently, the regulator history is used as a reference for two evolutionary algorithms to optimize the control parameters. The reduced domain provides a solution closer to the actual optimum at a lower computational cost and the optimized SMC provides superior performance which eliminates the need for trialand‐error exploration.