Abstract:In this study, we present a novel method for enhancing the computational efficiency of whole-body control for humanoid robots, a challenge accentuated by their high degrees of freedom. The reduced-dimension rigid body dynamics of a floating base robot is constructed by segmenting its kinematic chain into constrained and unconstrained chains, simplifying the dynamics of the unconstrained chain through the centroidal dynamics. The proposed dynamics model is possible to be applied to whole-body control methods, allowing the problem to be divided into two parts for more efficient computation. The efficiency of the framework is demonstrated by comparative experiments in simulations. The calculation results demonstrate a significant reduction in processing time, highlighting an improvement over the times reported in current methodologies. Additionally, the results also shows the computational efficiency increases as the degrees of freedom of robot model increases.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **to improve the computational efficiency of humanoid robot whole - body control (Whole - Body Control, WBC), especially the challenges faced when dealing with high - degrees - of - freedom (Degrees of Freedom, DOF) robots**. Specifically, the paper proposes a novel method to enhance the computational efficiency of WBC by simplifying the dynamic model of the floating - base robot. This method divides the robot's kinematic chain into a constrained chain and an unconstrained chain, and further simplifies the dynamics of the unconstrained chain through centroidal dynamics, thereby reducing the computational burden. This method allows the problem to be decomposed into two parts for more efficient computation. ### Main contributions of the paper: 1. **Introduced a new method for constructing a simplified rigid - body dynamic model**, achieving high computational efficiency in whole - body control. 2. **Showed how to integrate the simplified dynamic model into various iWBC (instantaneous Whole - Body Control) formulation methods**, providing a unique approach different from traditional methods. ### Specific methods for solving the problem: - **Classifying and simplifying the kinematic chain**: Divide the kinematic chain of the floating - base robot into three categories: virtual chain, constraint chain, and unconstrained chain. Simplify the dynamics of the unconstrained chain to six degrees of freedom through the concept of centroidal dynamics. - **Constructing the simplified whole - body dynamic model**: Map the dynamics of the unconstrained chain to the six - degree - of - freedom centroid space through projection techniques, and reconstruct the joint space into a reduced - dimension vector. - **Task control**: According to the simplified dynamic model, divide task constraints into two categories: tasks related to the joint space of the simplified model and tasks within the unconstrained chain. ### Experimental verification: The effectiveness of this method was verified through simulation experiments. The results showed that in single - support and double - support scenarios, the computation time was significantly reduced, increasing by 299% and 222% respectively, proving the superiority of this method in improving computational efficiency. ### Summary: This paper successfully solves the problem of high computational complexity in humanoid robot whole - body control by introducing a new dynamic simplification method, providing new ideas and technical means for efficiently implementing multi - task control.

Efficient Computation of Whole-Body Control Utilizing Simplified Whole-Body Dynamics via Centroidal Dynamics