Humanoid robots serve as a critical platform for advancing research in embodied intelligence, integrating dexterous manipulation, agile locomotion, and artificial intelligence. To be effectively deployed in daily human environments, robots must be capable of traversing complex, uncertain terrain—such as cobblestone surfaces and sloped ground—while simultaneously maintaining sufficient upper-body stability to execute tasks. 

In a paper published in IEEE Robotics and Automation Letters, researchers propose a whole-body impedance coordinative control framework for a wheel-legged humanoid robot to achieve adaptability on complex terrains while maintaining the robot’s upper body stability. 

As part of this framework, the authors define passive stable manipulation as maintaining the stability of a manipulated object without actively adjusting its position, relying instead on compliant coordination and whole-body dynamics. This capability is essential for tasks such as carrying water or balancing items during locomotion. 

Problem/Solution

The authors outline related research and strategies for legged robots, including prior studies addressing impedance adaptation, and note that the joint study of impedance adaptation and coupling, especially for passive stable manipulation during dynamic locomotion, has remained underexplored in humanoids.

Traversing complex terrains introduces significant challenges due to variable ground compliance, friction, and unforeseen perturbations, demanding high adaptability from locomotion controllers. Concurrently, maintaining upper-body stability is critical, especially when handling fragile or imbalanced objects, as minor oscillations can lead to task failures or safety risks.

The proposed whole-body coordinative control method, tested on a newly developed quadruped humanoid robot, effectively enables passive, stable manipulation for both indoor water-carrying and outdoor box-carrying tasks across varying, uneven terrain. The framework contains a bi-level control strategy:

  • The outer level is a variable-damping impedance controller that optimizes damping parameters to maintain upper-body stability while holding an object. 
  • The inner level employs whole-body control (WBC) optimization that integrates real-time terrain estimation based on wheel-foot position and force data. It generates motor torques while accounting for dynamic constraints, joint limits, friction cones, real-time terrain updates, and a model-free friction compensation strategy. 

Experiment Results 

According to the researchers, the quadruped humanoid robot is designed for human-centered environments, as it is expected to play an increasingly important role in daily settings such as homes and hospitals. The proposed 80 kg wheel-legged robot poses significant inertial challenges yet achieves passive upper-limb stability and balanced locomotion across varied, uncertain terrains. The robot adopts a wheel-leg hybrid structure that supports improved mobility while retaining key features of legged locomotion. 

Schematic of robot with control tasks for base position/orientation, wheel-centroid, centroid, and arms.

 

Throughout the experiments, the researchers note that the robot was not programmed with any prior accurate terrain information.

  • Experiment 1: Indoor Scenario

The indoor experiment involves the robot traversing a sequence of challenging terrains, including flat ground, uphill and downhill slopes, cobblestone surfaces, and wave-shaped terrain, requiring continuous adaptation to abrupt terrain changes. During the task, the robot holds a plate with both arms, carrying a bottle of water. The results demonstrated that the proposed algorithm can effectively suppress force oscillations caused by terrain variations, enhancing the robot’s upper-body manipulation stability during tasks such as transporting water over uneven surfaces.

  • Experiment 2: Outdoor Scenario

The outdoor experimental scenario is designed in a daily-life setting, where the robot carries a heavy object while traversing environments with unknown surface roughness and friction. The terrain includes brick sidewalks, concrete pavement, small concrete steps, and manhole covers, presenting realistic and challenging conditions. The results demonstrate that the proposed controller can effectively suppress external disturbances and maintain whole-body balance and passive stable manipulation, even when carrying heavy loads over uneven terrain, highlighting its robustness against real-world surface variations.

Experiments on both indoor and outdoor challenging terrains validate the framework’s effectiveness in enhancing balance, adaptability, and object transport stability. 

Conclusion

The article presented a whole-body impedance coordinative control framework for a newly developed wheel-legged humanoid robot. The proposed whole-body coordinative control method, tested on a newly developed quadruped humanoid robot, effectively enables passive, stable manipulation for both indoor and outdoor tasks across varying, uneven terrain.

Future work will extend the proposed framework to bipedal humanoid robots and incorporate active manipulation strategies to enable dynamic interaction with the environment.

We have thousands of articles related to these industries and more! Also learn more about our eBooks and eLearning collections on cutting edge technologies.