Inverse kinematics and adaptive gait controllers deployed on Raspberry Pi, enabling a quadruped to maintain balance and traverse uneven terrain under dynamic perturbations.
This project designs and deploys full-body gait planning and control for a custom quadruped robot. The focus was on enabling stable locomotion on uneven terrain and robustness to external perturbations — key challenges for field-deployable legged robots.
Inverse Kinematics: Per-leg IK solvers compute joint angles from desired foot-placement trajectories. The solvers handle the full 3-DOF leg kinematic chains and are efficient enough for real-time execution on Raspberry Pi.
Adaptive Gait Controller: A trot gait baseline was implemented with adaptive phase adjustments based on terrain height estimates from foot-contact sensing. When the controller detects loss of contact or unexpected ground contact, it redistributes body weight and adjusts step height.
Simulation & Validation: All gaits were first developed and tuned in PyBullet simulation with procedurally generated terrain before deployment to the physical platform.
The adaptive gait controller improved stability on uneven terrain by approximately 30% compared to a fixed-gait baseline. The robot maintained balance under lateral pushes and recovered within 2 steps in perturbation tests.