Unified Task Space Control for Redundant Robot Motion Retargeting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in reconstructing, retargeting, tracking, and estimating human or animal motion from low-dimensional task descriptors, particularly due to the redundancy of degrees of freedom in robotic systems, which complicates the inversion of kinematics from task-description space to joint space.
Innovation Solution
A unified task space control framework is developed that decomposes control structures into tracking control of task descriptors in Cartesian space and controls internal motion in the null space, using methods like first and second-order closed-loop inverse kinematics and inverse dynamics, with regularization of the inverse Jacobian matrix, to generate commands for robotic systems based on observed human motion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If task oriented control strategies are used to control redundant robotic systems, then dexterity and versatility are enhanced, but the ill-posedness of inverting kinematics from task space to joint space remains unresolved
Solution Approach 1:
The control structure is segmented into two independent parts: task space control (handling the task objective) and joint space control (handling the redundant degrees of freedom). This segmentation resolves the ill-posed inversion problem by treating task and redundancy separately, while maintaining dexterity through task-oriented control.
Solution Approach 2:
The problem is solved by transitioning from a single-space control approach to a two-space control framework (task space and joint space). This dimensional separation allows independent handling of task objectives and redundancy resolution, converting the ill-posed single-space inversion problem into two well-posed control problems.
2Reliability
If artificial performance indices are introduced to resolve the ill-posed inverse kinematics problem, then a solution can be obtained, but the treatment becomes overly complex and computationally intensive
Solution Approach 1:
The complexity of artificial performance indices is extracted and replaced by the simpler null space projection method. The essential function of resolving redundancy is separated from the task control, achieving reliable solutions without computational overhead of complex optimization indices.
Solution Approach 2:
Instead of directly solving the complex inverse kinematics with performance indices, the solution copies the task space control results and supplements them with null space motions. This copying approach maintains solution reliability while avoiding the computational complexity of direct inverse solutions.
3Manufacturing precision
If the full kinematic chain is controlled to achieve accurate task execution, then task precision is improved, but the control of internal self-motion becomes difficult to manage
Solution Approach 1:
The control is segmented into task space control (ensuring task precision) and null space control (managing internal self-motion). This segmentation makes the system easier to operate by allowing independent adjustment of task execution and internal motions without interfering with each other.
Solution Approach 2:
The null space acts as an intermediary between task execution and internal self-motion. It mediates the control by providing a separate channel for internal motions that does not interfere with task precision, making the overall control easier to manage.
Data Source
AI summary
A control system and method generate joint variables for motion or posing of a target system in response to observations of a source system. Constraints and balance control may be provided for more accurate representation of the motion or posing as replicated by the target system.