Assessing Feasible Gait Dynamics of a Highly Coupled Quadruped Robot with Six Actuated Degrees of Freedom

Abstract

This thesis explores the feasibility of stable gait dynamics in a quadruped robot with six actuated degrees of freedom (DoF). Unlike conventional quadrupeds with twelve actuated DoF, this novel design employs a highly coupled structure, with each actuator influencing multiple legs. The reduction in DoF presents significant challenges in terms of locomotion control, workspace and stability during motion. The primary objective is to develop and validate quasi-static crawling gaits that maintain stability despite the robot’s constrained body workspace. The work begins by analyzing the coupled kinematic design, which was optimized in previous research (GAISSER 2024), to understand its constraints and motion capabilities. Using a parameterized approach, feasible gaits are identified through a combination of sampling in a configuration space and simulation-based evaluations. The sampling framework assesses stability at critical transition points in the gait cycle, using metrics such as stability margins, joint limits, and workspace constraints. Additionally, inverse kinematics are solved through simulation to generate joint trajectories, addressing the unique challenges posed by the coupled kinematic chains and singularities in the design. A prototype robot was developed to physically test the feasibility of the proposed gaits. The hardware was carefully designed accounting for practical considerations such as motor torque, link stiffness, and collision avoidance. Simulation results demonstrate the feasibility of straight and spinning crawling gaits, and these were further validated experimentally with the physical robot on flat surfaces. This research demonstrates the viability of stable crawling in quadrupeds with six actuated Degrees of Freedom (DoF), offering insights into locomotion strategies for underactuated robotic systems with highly coupled kinematics. The findings suggest potential pathways for extending this work to dynamic gaits, adaptive control, and more complex terrains, paving the way for cost-effective and energy-efficient quadruped robots in diverse applications.