Gait analysis and control of a quadruped robot on rugged surfaces

In this thesis, in order to improve step trajectory control of a quadruped robot (4AR) diversity the optimal controller designs and by designing the unique control structure to walking on uneven/rugged surfaces the modeling, simulation, and analysis studies were carried out. By considering the experimental studies, the design criteria and physical parameters of the 4AR have been determined. The forward/inverse kinematic analysis and the numerical calculations have been performed, and the solutions of closed-form mathematical dynamics based on the Newton-Euler method had been presented. To provide a basis for gait studies, the optimal PID, Hybrid LQR-PID, Fuzzy Logic Controller (FLC) designs have been realized using heuristic algorithms which not proposed before and unique approaches and evaluated comparatively for trajectory control of the step motion. A PID based Impedance Controller is designed to control the interaction between the end-point of the leg and the surface. In addition, the three degrees of freedom leg structure with physical and dynamic parameters belonging to a quadruped robot (HyQ2max) was modeled, and to increase diversity the PIλDµ controller was designed, its performance was compared to the classical PID controller. The foot trajectory control of the system had been improved by the studies on the step motion. the solid model of the system was designed in a CAD program and the dynamic model of 4AR was obtained by it is transferred to the ADAMS environment. A PD based torque controller is developed which could adapt its own gains according to the step phases along the walk. With the simultaneous running of ADAMS and MATLAB/Simulink programs, the planar walking simulations on the uneven ground were carried out at a total of four different types; the trot, pace, and canter dynamic gaits as well as the static gait. In this thesis, a unique control approach in which the body and step motion could be controlled together and their effects on the controller output could be adjusted, and also that included the impedance position-force control that regulates the interaction between the end-point and the ground was proposed for gait control on rugged surfaces. The parkour consisting of four different types of standard roughness was designed, the rugged gait control and simulation of 4AR were carried out. It has been shown that the proposed control structure could successfully control the reference gait planning on the rugged parkour while keeping the balance and stability of 4AR. Detailed numerical results of all the system variable and error values of the system were presented and evaluated comparatively with similar results in the literature.