LQR-Assisted Whole-Body Control of a Wheeled Bipedal Robot with Kinematic Loops (2005.11431v1)

Abstract: We present a hierarchical whole-body controller leveraging the full rigid body dynamics of the wheeled bipedal robot Ascento. We derive closed-form expressions for the dynamics of its kinematic loops in a way that readily generalizes to more complex systems. The rolling constraint is incorporated using a compact analytic solution based on rotation matrices. The non-minimum phase balancing dynamics are accounted for by including a linear-quadratic regulator as a motion task. Robustness when driving curves is increased by regulating the lean angle as a function of the zero-moment point. The proposed controller is computationally lightweight and significantly extends the rough-terrain capabilities and robustness of the system, as we demonstrate in several experiments.

• The paper presents a hierarchical LQR-assisted whole-body control framework for wheeled bipedal robots, incorporating detailed dynamic modeling and novel rolling constraint solutions.
• Experimental validation demonstrates enhanced robot stability and dynamic leg adjustment capabilities, enabling traversal of rough terrain and recovery from disturbances.
• This research significantly improves wheeled bipedal robot stability and environmental adaptability, offering a foundation for developing more sophisticated autonomous systems for complex environments.

Overview of LQR-Assisted Whole-Body Control of a Wheeled Bipedal Robot with Kinematic Loops

The paper "LQR-Assisted Whole-Body Control of a Wheeled Bipedal Robot with Kinematic Loops" authored by Klemm et al. presents a robust framework for controlling a wheeled bipedal robot known as Ascento. The key innovation of this paper lies in the hierarchical controller that integrates whole-body control (WBC) with a Linear-Quadratic Regulator (LQR). This methodology facilitates the balancing and navigation of Ascento across varied terrains by leveraging the full dynamics of the robot.

Core Contributions

1. Rigorous Dynamic Modeling: The authors derive comprehensive dynamics for Ascento, explicitly modeling the kinematic loops within its leg-like structures. This approach is essential for maintaining the robot's balance and agility.
2. Rolling Constraint Incorporation: The paper introduces a novel closed-form solution using rotation matrices to incorporate the rolling constraints of Ascento's wheels, ensuring precise control over its mobility.
3. LQR Integration: The balancing challenges posed by non-minimum phase dynamics of wheeled bipedal robots are addressed by integrating an LQR feedback loop into the WBC framework, enhancing the robot's stability during dynamic tasks.
4. Lean Angle Regulation for Robustness: The paper further refines control by regulating the robot's lean angle in relation to the zero-moment point (ZMP), crucial for maintaining balance when the robot navigates curves or uneven terrain.

Numerical and Experimental Results

The experimental validation of the proposed controller is robustly demonstrated through various tests, highlighting its effectiveness in real-world scenarios such as traversing rough terrain and recovering from disturbances. The LQR-assisted WBC markedly improved Ascento's ability to adjust leg extensions dynamically, a capability vital for maintaining an upright position and negotiating irregular surfaces.

Implications and Future Work

Practically, this research enhances the operational stability and environmental adaptability of wheeled bipedal robots, expanding their utility in fields like search and rescue. Theoretically, it paves the way for future work that could incorporate dynamic modeling of contact events, potentially enabling the design of robots capable of complex maneuvers, including jumping.

Further advancements could focus on improving state estimation through integration of additional sensors and employing cutting-edge estimation algorithms like Unscented Kalman Filters or visual-inertial odometry. Enhancements in system identification processes could also refine the model's accuracy, which would improve control precision and efficiency.

In conclusion, this paper offers a significant advancement in the control of wheeled bipedal robots, merging hierarchical control with dynamic adaptability, thereby substantially broadening the capabilities and applications of such systems. The methodologies and findings can serve as a foundation for developing more sophisticated robotic systems adept at navigating complex environments autonomously.