Robot Learning Extension Unit for Vibration-Resilient Position Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional robot learning systems face challenges in applying learning correction values effectively, particularly when the correction amount is large or when the work position changes randomly, leading to interference with vibration suppression and speed advantages, and are not suitable for systems using vision sensors with unfixed work positions.
Innovation Solution
A robot apparatus that calculates a learning correction value for new operations by obtaining the relationship between learning correction values and operation information, allowing for the calculation of another learning correction value without the need for sensor relearning, using a learning extension unit and transfer functions derived from spectrograms of operation data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a learning correction value of a different position is applied as-is, then the operation can be corrected, but it interferes with the effect of suppressing vibration
Solution Approach 1:
The patent applies local quality by creating position-specific learning correction values through spectrogram analysis. Instead of applying a generic correction value across different positions, the system analyzes vibration characteristics (spectrograms) at each specific position and generates tailored correction values. This ensures that each position receives a correction value optimized for its local vibration characteristics, thereby maintaining vibration suppression effectiveness while achieving operation correction.
2Reliability
If the learning correction value is made zero by deceleration for large correction amounts, then safety is improved, but the speed advantage is interfered with
Solution Approach 1:
The patent applies parameter changes by using spectrogram analysis to transform the approach from binary correction (zero or full correction) to continuous parameter optimization. The system analyzes vibration frequency and amplitude characteristics through spectrograms and adjusts correction values as continuous parameters rather than binary decisions. This allows the robot to operate at high speeds while applying precisely calibrated correction values that maintain safety without requiring deceleration.
3Measurement precision
If conventional learning methods are used with fixed position sensors, then learning accuracy is improved, but the system cannot be applied to situations where the position of work changes at random
Solution Approach 1:
The patent applies universality by creating a learning correction system that works across multiple positions and varying work conditions. The spectrogram-based approach captures vibration characteristics that are position-independent, allowing the same learning methodology to be applied whether the work position is fixed or changes randomly. The system generates correction values based on vibration patterns rather than absolute position coordinates, making it universally applicable to different positioning scenarios including vision sensor systems.
Data Source
AI summary
A robot apparatus includes a robot mechanism; a sensor provided in a portion whose position is to be controlled, of the robot mechanism, for detecting a physical quantity to obtain positional information of the portion; and a robot controller having an operation control unit for controlling the operation of the robot mechanism. The robot controller includes a learning control unit for calculating a learning correction value to improve a specific operation of the robot mechanism based on the physical quantity detected, while the operation control unit makes the robot mechanism perform the specific operation, with the sensor; and a learning extension unit for obtaining the relationship between the learning correction value calculated by the learning control unit and information about the learned specific operation, and calculates another learning correction value to improve a new operation by applying the obtained relationship to information about the new operation without sensor.