Joint Control Of Quadruped Robot Driven By Hydraulic Cylinder

Moving flexible and having strong adaptive ability to the complicated terrain, quadruped robot has wide application in the military, civil, resource exploration, geography adventure, etc. Joint rapid control is one of the core technologies of the high-performance quadruped robot driven by hydraulic system, in which the displacement, velocity and force control are researched, and appropriate control algorithms are developed and simulated in this thesis. The main content is as follows:(1) The necessity and feasibility of the hydraulic system used as driver on the quadruped robot are described, and valve-controlled hydraulic cylinder module as the final actuator of the walking control is determined, and a mathematical model of the module is established. According to the characteristics of valve-controlled hydraulic cylinder system, the possible control strategies are discussed.(2) The composition of the hydraulic position system and the transfer function module are analyzed, and the simplified mathematical model of system is obtained, based on this model, a sliding mode controller and a fuzzy sliding mode controller are designed, and control performances are compared between two controllers-facing to the parameter perturbation and external disturbances of hydraulic position system, sliding mode controller has good robustness and rapid response, but there are some chattering; and fuzzy sliding mode controller not only inherits the advantages of a single sliding controller, but also effectively reduces chattering.(3) Through calibration of the hydraulic velocity system, I-velocity system which is equivalent to position system is obtained. Using control methods of hydraulic position system for reference, a sliding mode controller and a fuzzy sliding mode controller are designed for joint hydraulic velocity system, and achieve effective velocity control.(4) According to the nonlinear characteristics of hydraulic driven system, the nonlinear mathematical model of hydraulic force system is derived, and a fuzzy sliding mode controller is presented with a fuzzy boundary adjustment function. This controller is robust to a large variation of parameters and nonlinearity, and ensures the high precision and optimum dynamic and static quality of hydraulic force control.(5) Based on the nonlinear mathematical model of hydraulic force system, a fuzzy neural network sliding mode control method is proposed. The method has good effect to nonlinearity and uncertainty of hydraulic force system, and can realize high speed and high precision dynamic force control of the joint hydraulic driven system under the high frequency signal.The simulation results indicate that the designed controllers realize the joint real-time control of a quadruped robot which is driven by hydraulic system, and satisfy the project designing requirement in this thesis. This platform has certain extensible ability, it can work as basic controller for robot control system, and it provides theoretical basis and key technical support for the development of quadruped robot.