Robot Manipulator Velocity Control with Low-DoF Teleoperation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing robotics technologies require highly customized and expensive hardware for controlling dexterous robot manipulators with many degrees-of-freedom, limiting usability and accessibility, and are inefficient in handling large action spaces during reinforcement learning-based tasks.

Innovation Solution

A method for controlling robot manipulators using a teleoperation device with fewer degrees-of-freedom, which involves formulating an optimization problem to determine joint velocities based on target velocities and constraints, allowing for effective control with a less complex input device and reducing computational resources.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If highly customized hardware is used to control dexterous robot manipulators with many degrees-of-freedom, then control precision and dexterity are improved, but hardware cost and complexity increase significantly

Engineering Contradiction:
Improvecontrol precisionVSAvoidhardware complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces an optimization controller as an intermediary computational layer between the simplified input device and the robot manipulator. This controller solves optimization problems to translate low-DoF control inputs into appropriate joint velocity commands, enabling precise control without requiring complex hardware at the input device

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces complex mechanical control hardware with computational algorithms. Instead of using specialized hardware with many degrees of freedom to control each joint, the system uses optimization-based software control to map simplified user inputs to multi-DoF robot movements, substituting mechanical complexity with computational intelligence

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Ease of manufacture

If teleoperation device with fewer degrees-of-freedom is used, then hardware cost and accessibility are improved, but control capability for high-DoF manipulators deteriorates

Engineering Contradiction:
Improvehardware accessibilityVSAvoidcontrol capability
Core Design Contradiction:
Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The patent changes the control parameter space by formulating optimization problems that map from a reduced-dimensional input space to the full robot state space. The optimization controller dynamically adjusts control parameters based on task requirements, enabling a simplified device to control complex manipulators across diverse tasks

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The optimization controller provides universal control capability that works across multiple tasks and robot configurations. The same simplified teleoperation device can control different high-DoF manipulators for various tasks by adjusting optimization parameters, achieving multi-functionality without hardware changes

Inventive Principle:
Principle #6Universality (Multi-functionality)

3Measurement precision

If conventional control methods are used for reinforcement learning tasks, then training accuracy is improved, but computational resource consumption and training time increase

Engineering Contradiction:
Improvetraining accuracyVSAvoidtraining efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent extracts and separates the computational burden of high-DoF control from the reinforcement learning agent. By pre-computing optimization mappings and using efficient solvers, the system reduces the action space complexity that the RL agent must handle, allowing faster training while maintaining policy accuracy

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The optimization controller performs preliminary computation to determine feasible joint velocities before execution. This pre-computation of control commands based on current state and desired task motion reduces real-time computational demands during RL training episodes, improving overall training efficiency

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4335598A1Action abstraction controller for fully actuated robotic manipulators
Publication Date: 2024.03.13 GDM HOLDING LLC
  • EP4335598A1 patent drawingFigure 1
  • EP4335598A1 patent drawingFigure 2
  • EP4335598A1 patent drawingFigure 3

AI summary

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for controlling a robot manipulator that has a plurality of joints. One of the methods includes obtaining a control input that comprises one or more velocity values that specify a target velocity of a reference point in a given coordinate frame; determining a respective joint velocity for each of the plurality of joints by generating a solution to an optimization problem formulated from the control input; and controlling the robot manipulator, including causing the plurality of joints of the robot manipulator to move in accordance with the respective joint velocities to approximate the control input.