Redundant Robot Control With Passive Task Prioritization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for controlling kinematically redundant robots struggle with maintaining strict task hierarchy while handling multiple tasks, especially in the presence of task singularities and model errors, leading to issues like vibrations and loss of stability.
Innovation Solution
A method using passivity-based controller modules combined through optimization with weighted quadratic programs, allowing for soft prioritization of tasks and maintaining passivity even in overdetermined control scenarios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If hierarchical controllers with nullspace projections are used to maintain strict task hierarchy, then task priority control is improved, but task singularities cause loss of strict hierarchy and generate unwanted vibrations
Solution Approach 1:
The patent changes the control parameter from strict hierarchical weighting to a continuous potential field formulation. By defining a potential function that naturally encodes task priorities through spatial relationships rather than explicit weighting parameters, the system maintains hierarchy stability without the singularity problems of traditional weighted approaches. The potential field parameters (attractors, repellers, weights) are optimized to achieve both stability and singularity robustness simultaneously.
Solution Approach 2:
The patent replaces the mechanical hierarchical control structure (with its rigid task prioritization and nullspace projections) with a field-based potential formulation. This substitution transforms the control problem from a mechanical hierarchy maintenance issue into a field optimization problem, where task priorities emerge naturally from the potential field configuration rather than being imposed through hierarchical constraints.
2Adaptability or versatility
If inverse dynamics-based tracking controllers are used for flexible task handling, then adaptability is improved, but robustness against model errors and contact uncertainties deteriorates
Solution Approach 1:
The patent introduces dynamic optimization of the potential field parameters in real-time, allowing the controller to adapt to changing task requirements and environmental conditions. The dynamic reconfiguration of attractors, repellers, and weights enables flexible task handling while the underlying potential field formulation maintains robustness by avoiding the explicit model inversion operations that plague inverse dynamics approaches.
3Reliability
If multiple passive tracking controller modules are combined through optimization, then robustness is improved, but device complexity increases
Solution Approach 1:
The patent merges multiple passive tracking controller modules into a unified potential field-based controller. Instead of combining controllers through complex optimization frameworks, the invention integrates task-level controllers by formulating them as a joint potential field optimization problem, where all task objectives are unified into a single potential function that naturally coordinates all controller modules.
Solution Approach 2:
The patent creates a universal controller framework that can handle multiple task types and priorities through a single potential field formulation. The unified controller structure serves multiple functions (task execution, priority management, singularity avoidance, vibration suppression) simultaneously, eliminating the need for separate specialized controllers for each function.
Data Source
AI summary
A method for controlling a kinematically redundant robot (100) in order to fulfill multiple tasks. At least one passivity-based first controller module (102) is used, at least one task target description and at least one associated task mapping are computed for the at least one first controller module (102), at least one weighting is computed for the tasks, and the at least one first controller module (102) is integrated into an overall controller (104), using the at least one weighting. Moreover, the invention relates to a computer program product that includes commands which, when the program is executed with the aid of at least one processor, prompt the processor to carry out such a method.