Robot Effector Trajectory Planning With Relaxed Contact Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current multi-contact motion planning methods in robotics are computationally intensive and inefficient, particularly when dealing with complex tasks involving multiple alternate options, as they separate the problem into subtasks, losing interdependencies and requiring significant computational resources.
Innovation Solution
A method that acquires a graph of postures with contact points and kinematic poses, modifies constraint topologies, performs constraint relaxation, and generates effector trajectories using a trajectory generation algorithm, allowing for adaptive and optimized motion planning that considers overall contact sequences and optimality criteria.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If multi-contact motion planning problems are broken down into subtasks and solved separately, then the computational complexity of each individual step is reduced, but the interdependencies between computation steps are lost and overall computational requirements increase
Solution Approach 1:
The patent merges the previously separate computation steps (contact sequence search, contact pose estimation, and trajectory optimization) into a unified optimization framework. This allows the system to maintain computational tractability while preserving and exploiting interdependencies between steps, thereby improving overall computational efficiency without sacrificing the benefits of modular problem decomposition
Solution Approach 2:
The patent performs preliminary identification of contact sequences and contact poses before conducting the final trajectory optimization. This preliminary action provides a structured foundation that guides the subsequent optimization process, reducing the search space and computational burden while maintaining the ability to adapt to interdependencies among computation steps
2Reliability
If multiple alternate possibilities for addressing a specific task are examined individually, then the quality of the selected solution is improved, but the computational resources required increase significantly
Solution Approach 1:
The patent combines the evaluation of multiple alternate possibilities into a single unified optimization framework that simultaneously considers multiple contact sequences and their corresponding trajectories. This approach maintains solution quality by evaluating alternatives together rather than separately, while reducing computational resource consumption through efficient shared computations and a unified objective function
Solution Approach 2:
The patent implements a two-stage approach where a preliminary contact sequence search identifies candidate solutions, followed by a more comprehensive optimization that evaluates multiple alternatives. This partial action in the first stage reduces the overall computational burden while the second stage ensures high solution quality by thoroughly evaluating the reduced set of alternatives
Data Source
AI summary
A method and system for controlling at least one effector trajectory for at least one effector of a robot for solving a predefined task are proposed. A graph of postures is acquired, and at least one of a contact constraint topology and an object constraint topology are accordingly modified. A set of constraint equations based on at least one of the modified contact constraint topology and the modified object constraint topology are generated. Constraint relaxation is performed on the generated set of constraint equations to generate a task description including the relaxed set of constraint equations. The effector trajectory is generated by applying a trajectory generation algorithm on the generated task description. An inverse kinematics algorithm is performed on the generated effector trajectory for generating a control signal, and the effector is controlled to execute the effector trajectory based on the generated control signal.