Tendon-Driven Manipulator Self-Calibration via Zero-Torque Regression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Tendon-driven manipulators require frequent calibration to account for sensor drift and external impacts, but existing methods necessitate disassembly and external force references, which are impractical.
Innovation Solution
A method for calibrating tendon tensions in a tendon-driven manipulator without disassembly or external force references, using zero-tension and fully-tensioned data points to determine sensor calibration parameters through a regression process that satisfies the zero-torque constraint and minimizes error relative to nominal calibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional calibration methods are used with external force references, then measurement precision is improved, but device complexity and ease of operation deteriorate due to required disassembly and external equipment
Solution Approach 1:
The calibration method enables the manipulator to calibrate its own tension sensors using internal tendon tensions and zero-torque constraints, without requiring external force references or disassembly. The system uses its existing structure and control capabilities to perform self-calibration through regression analysis of tendon tension data collected at different manipulator configurations.
Solution Approach 2:
The patent introduces an intermediary calibration process that uses the manipulator's own tendon tensions and kinematic model as mediators between the sensors and the calibration outcome. Instead of directly applying external forces, the system uses the relationship between tendon tensions and joint torques (through the tendon map matrix) as an intermediary to achieve calibration.
2Measurement precision
If traditional calibration methods are used with external force references, then measurement precision is improved, but device complexity worsens due to additional equipment and disassembly requirements
Solution Approach 1:
The manipulator performs self-calibration using its existing components (tendons, sensors, actuators) without requiring external calibration equipment. The system leverages its redundant tendon network and control capabilities to generate the necessary calibration data internally, eliminating the need for additional devices.
Solution Approach 2:
The calibration method uses the manipulator's existing actuators and tendon system to serve multiple functions: both actuation and calibration. The same tendons and sensors used for normal operation are utilized for calibration, eliminating the need for separate calibration equipment and reducing overall system complexity.
3Ease of operation
If in-vivo calibration without disassembly is implemented, then ease of operation is improved, but measurement precision may worsen without external force references
Solution Approach 1:
The calibration process uses feedback from the manipulator's own sensor readings and kinematic model to iteratively determine calibration parameters. The system collects tendon tension data at different configurations, compares it with expected values based on the kinematic model and zero-torque constraints, and uses regression analysis to optimize calibration parameters that minimize the error.
Solution Approach 2:
The method changes the operational parameters of the manipulator during calibration by moving joints to different positions and measuring tendon tensions at each configuration. This generates multiple data points that are used in the regression process to determine calibration parameters, leveraging parameter variation to achieve calibration without external references.
4Measurement precision
If frequent calibration is performed to account for sensor drift, then measurement precision is maintained, but loss of time increases due to disassembly requirements
Solution Approach 1:
The manipulator can perform rapid self-calibration during normal operation or between tasks without requiring disassembly or external equipment. This significantly reduces calibration time compared to traditional methods, enabling frequent calibration to compensate for sensor drift while maintaining operational efficiency.
Solution Approach 2:
The calibration process can be performed preliminarily during manufacturing or initial setup, establishing baseline calibration parameters. Subsequent recalibrations can then be performed quickly using the same self-calibration method when drift is detected, reducing the frequency and time of full calibration cycles.
Data Source
AI summary
A method for calibrating tension sensors on tendons in a tendon-driven manipulator without disassembling the manipulator and without external force references. The method calibrates the tensions against each other to produce results that are kinematically consistent. The results might not be absolutely accurate, however, they are optimized with respect to an initial or nominal calibration. The method includes causing the tendons to be slack and recording the sensor values from sensors that measure the tension on the tendons. The method further includes tensioning the tendons with the manipulator positioned so that it is not in contact with any obstacle or joint limit and again recording the sensor values. The method then performs a regression process to determine the sensor parameters that both satisfy a zero-torque constraint on the manipulator and minimize the error with respect to nominal calibration values.