Dynamic modeling and Designing a dynamic based control algorithm for legged quadruped robot locomotion

In this study, a dynamic based control algorithm for a six-link quadruped locomotion is proposed. Up to now, a lot of robotic scientists have researched in quadruped locomotion but most of their researches are based on modeling of robot and its surrounding. Such methods are not able to generate a stable locomotion when the surrounding changes a little. So this is important to propose a dynamic based control algorithm. The algorithms that can guarantee the stability are classified to two categories of dynamic based and trajectory based methods. The trajectory based algorithms need detailed information of gait and surrounding which is not necessarily available. But the dynamic based algorithms use some geometric constraints to reach a stable controller. These geometric constraints generate the proper gaits. So in this study by employing the dynamic based control algorithm, we proposed a controller for generating the Trot and Pace gait on a straight and flat path for quadruped robot locomotion. Given that the quadruped robot has four degrees of freedom so three geometric constraints are needed to provide a rhythmic locomotion. In this study we showed that for step generating by quadruped robot, both the appropriate initial conditions for angular velocities and presence of a point mass on the neck of the robot are needed. Also in this study the stability of quadruped locomotion has been proved using Poincaré return map.