In context
By late 2024, quadruped robots had shown promise in industrial inspection and field operations, but their deployment on unstructured terrain remained limited by reliance on vision and complex model-based control. Researchers at Fuzhou University addressed this gap with a vision-free control scheme that fuses proprioceptive sensors for terrain adaptation.
What was reported
The team proposed a terrain perception and stiffness adaptive control method for quadruped robots operating on uneven ground. An error-state Kalman filter (ErKF) fuses data from GPS, IMU, and joint encoders to estimate the robot's position, velocity, and attitude without visual input, enabling high-frequency state estimation.
A terrain-aware attitude controller (TAAC) uses the estimated states to compute an optimal body orientation that aligns with the local ground plane, suppressing high-frequency body jitter in fluctuating environments. A stiffness adaptive whole-body controller (SAWBC), based on impedance theory, adjusts controller stiffness by solving an optimal stiffness variation law, improving convergence of attitude errors and dynamic stability.
Simulations and prototype experiments demonstrated enhanced terrain adaptability and smoother motion compared with conventional methods. The stability of the system was verified using Lyapunov theory.
Why it mattered
This work advanced vision-free control for legged robots, reducing dependence on external terrain knowledge and visual processing. It pointed toward more robust, self-contained quadruped platforms for industrial inspection, search-and-rescue, and logistics in unstructured environments.
The simulation results and prototype experiments show that the proposed method significantly enhances the terrain adaptability and motion smoothness of the robot.
Source: 《机器人》期刊 (robot.sia.cn) · Published 2024-11-15 · “基于地形和刚度自适应的四足机器人姿态控制”