In context
Legged robots have long drawn inspiration from biology to achieve agile locomotion, but controlling them for precise trajectory tracking remains challenging. Traditional controllers often rely heavily on accurate models and complex parameter tuning, limiting adaptability. By 2023, researchers were increasingly exploring bio-inspired central pattern generators (CPGs) for smooth gait generation, yet integrating them with high-level planning for accurate path following was still an open problem.
What was reported
A team from Harbin Engineering University proposed a hybrid control scheme that fuses model predictive control (MPC) with a CPG network for a hexapod robot. The CPG, based on Hopf oscillators, generates smooth, coordinated joint signals for the six legs, while an upper-level MPC adjusts the CPG parameters in real time to achieve trajectory tracking. A transfer function maps the robot's desired velocity and angular velocity to CPG frequency and steering parameters, enabling seamless integration.
The controller was validated through simulations and experiments tracking circular and straight-line paths. Under initial position and heading errors, the robot quickly converged to the reference trajectory, maintaining position error within -0.1 to 0.1 m and heading error within -27° to 20°. The MPC-CPG approach demonstrated high tracking accuracy alongside smooth, coordinated motion, confirming its effectiveness.
Why it mattered
This work addresses a critical gap in legged robotics: combining bio-inspired locomotion with model-based optimization for precise path following. The hybrid architecture offers a template for improving autonomy in legged robots operating in cluttered or hazardous environments, such as inspection or search-and-rescue, where accurate trajectory tracking is essential. It also highlights the potential of merging neural-inspired control with classical optimization for robust industrial automation.
With the MPC-CPG controller, the robot not only has a high trajectory tracking accuracy, but also shows good motion smoothness and coordination.
Source: 《机器人》期刊 (robot.sia.cn) · Published 2023-03-06 · “基于模型预测-中枢模式发生器的六足机器人轨迹跟踪控制”