Robot Abnormality Detection Using Disturbance Torque Distributions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing abnormality determination devices for robots may incorrectly identify normal operating conditions as abnormal due to significant variations in disturbance torque following maintenance or teaching procedures, leading to inaccurate assessments.
Innovation Solution
An abnormality determination device that calculates a measurement probability distribution of disturbance torque over a predetermined period and compares it to an evaluation normal model, ensuring reference points align, to accurately determine apparatus abnormalities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a constant threshold is used for abnormality determination, then the determination process is simple, but the accuracy of abnormality detection deteriorates due to great variations in disturbance torque after maintenance or teaching
Solution Approach 1:
The patent applies dynamics by transitioning from a static constant threshold to a dynamic threshold that adapts to different operational states. The determination device now uses probability distributions calculated from actual measurement data during specific periods (normal operation periods) to create state-specific thresholds. This allows the threshold to dynamically adjust based on the robot's operational context, improving detection accuracy while managing complexity through automated statistical calculations.
Solution Approach 2:
The patent implements parameter changes by modifying the threshold parameter from a fixed constant value to a variable derived from statistical analysis of measurement data. The threshold is now expressed as a probability distribution with parameters (mean, standard deviation) that change based on the operational period and state. This enables the system to accommodate variations in disturbance torque after maintenance or teaching while maintaining reliable abnormality detection.
2Measurement precision
If probability distribution comparison is used for abnormality determination, then the accuracy of abnormality detection is improved, but the computational complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the operational period into distinct phases: normal operation periods for building probability distributions and current periods for comparison. This temporal segmentation allows the system to accumulate statistical data during normal operations without continuously processing all data, reducing computational burden while maintaining high detection accuracy through targeted comparisons between baseline and current states.
Solution Approach 2:
The patent implements preliminary action by pre-calculating and storing probability distribution parameters during normal operation periods before abnormality detection is needed. This advance preparation creates a baseline reference that can be quickly compared against current measurements, reducing real-time computational complexity while preserving the accuracy benefits of probability distribution analysis.
Data Source
AI summary
An abnormality determination device includes a control unit for determining an abnormality of a robot, the control unit being configured to calculate a measurement probability distribution which is a probability distribution using disturbance torque acquired during a predetermined period as a random variable. The control unit causes an average of the calculated measurement probability distribution to conform to an average of an evaluation normal model which is a predetermined probability distribution, compares the measurement probability distribution with the evaluation normal model of which the respective averages conform to each other, and determines an abnormality of the robot in accordance with a result of the comparison.