Robot Gravity Direction Estimation Using Joint Torque Curves
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the direction of gravity relative to a robot are inaccurate due to deviations in estimated gravity, leading to errors in motion control, particularly when there are inaccuracies in linear gain and offset in joint torque measurements, and rely on dynamic models that may not be reliable.
Innovation Solution
A method that estimates the direction of gravity by recording torque information from two non-parallel revolute joints in a robotic system, using sinusoidal portions of torque curves to abstract gravitational effects, thereby eliminating influences from friction and decentration, and calculating reference angles to determine the direction of gravity without relying on the robot's dynamic model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If direct measurements from external sensors (e.g., inclinometer) are used to determine gravity direction, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The robot uses its own joint torque sensors to measure gravity direction by rotating its joints and recording torque information. The existing torque sensors serve dual purposes: controlling robot motion and measuring gravity direction, eliminating the need for separate external sensors like inclinometers.
2Device complexity
If dynamic model-based estimation is used to determine gravity direction, then device complexity is reduced, but measurement precision deteriorates due to inaccuracies in linear gain and offset
Solution Approach 1:
The patent replaces the dynamic model-based calculation method with a direct torque measurement method. Instead of using complex dynamic models that require accurate parameters for gravity direction estimation, the system directly measures torque at different joint positions and uses the sinusoidal characteristics of torque curves to determine gravity direction, eliminating model-related errors.
3Ease of operation
If dynamic model-based estimation is used, then ease of operation is improved, but reliability deteriorates due to dependence on accurate dynamic models
Solution Approach 1:
The robot performs self-calibration by rotating its own joints and using its built-in torque sensors to collect data. This self-service approach eliminates dependence on external calibration equipment and complex dynamic models, improving reliability while maintaining operational simplicity through automated procedures.
4Measurement precision
If torque information from multiple joints is recorded and processed, then measurement precision is improved, but loss of time increases due to multiple rotation and recording steps
Solution Approach 1:
The patent uses torque information from at least two joints to determine gravity direction, which provides redundant measurements that improve precision. By using multiple joints rather than relying on a single joint or complex dynamic models, the system achieves more accurate results through multiple independent measurements.
Data Source
AI summary
A method, system, and non-transitory, computer-readable medium are provided for estimating a direction of gravity with respect to a robot. The method includes rotating a first joint. A first torque information of the first joint is recorded during rotation of the first joint. A second joint is then rotated. A second torque information of the second joint is recorded during rotation of the second joint. The direction of gravity is then estimated with respect to the robot based on the first torque information and the second torque information. The method provides an efficient way for determining the direction of gravity with respect to the robot.