Robot Control Using Environmental Matching for Accurate Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robot control systems using machine learning struggle to accurately operate when environmental conditions differ from those during the learning process, leading to decreased inference accuracy and difficulty in adjusting operations accordingly.
Innovation Solution
A robot control system that includes an environment sensor and a control device capable of calculating a matching degree between current environmental conditions and those at the time of learning, adjusting operation commands and sensor settings to match the original conditions, ensuring accurate operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a learning unit is used to generate operation commands based on environmental sensor output, then the robot can execute work in various work scenes, but the inference accuracy decreases when the current environmental condition differs from the learning condition
Solution Approach 1:
The patent introduces an environmental condition matching degree calculation unit as an intermediary between the environment sensor and the learning unit. This unit calculates the matching degree between current environmental conditions and learning-time environmental conditions, then uses this information to adjust either the operation commands or sensor settings, thereby maintaining inference accuracy across varying environmental conditions without requiring extensive retraining
Solution Approach 2:
The patent dynamically changes parameters (either operation command parameters or environment sensor settings) based on the calculated environmental condition matching degree. When the matching degree is low, the system adjusts these parameters to compensate for environmental differences, allowing the learning unit to maintain high inference accuracy in conditions different from those during learning
2Adaptability or versatility
If the operation is adjusted according to user input of constraint conditions, then the robot can adapt to environmental changes, but it requires the user to correctly recognize and convert environmental changes into constraint conditions
Solution Approach 1:
The patent enables the system to automatically perform the function that previously required user intervention. The environmental condition matching degree calculation unit automatically detects environmental changes and calculates the matching degree, then the adjustment unit automatically adjusts operation commands or sensor settings based on this calculation, eliminating the need for users to manually recognize and convert environmental changes into constraint conditions
Data Source
AI summary
This robot control system is provided with a robot device, an environment sensor which acquires an environmental condition of the robot device, and a control device which controls the operation of the robot device by employing a trained model to generate an operation command for the robot device, corresponding to the current output of the environment sensor, wherein the control device is provided with: an environmental condition matching degree calculating unit which calculates a degree of matching between the environmental condition of the robot device during learning and the current environmental condition of the robot device, on the basis of the current output of the environment sensor; and an adjusting unit which adjusts the content of the robot device operation command on the basis of the calculation result of the degree of matching.