Robot Object Learning via Remote Verification and Motion Capture
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robotic systems face challenges in accurately identifying objects due to failures in edge detection and object recognition, particularly when objects overlap in color, are translucent, or new, leading to inefficient operation and potential collisions or grasping errors.
Innovation Solution
An object identification training method for robotic devices involving a movable appendage with a visual sensor that captures images while moving through a predetermined learning motion path, with human operator verification and data association to improve machine learning model accuracy, allowing for additional data collection and confidence score-based decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the robotic device uses automated object recognition without human verification, then productivity is improved, but measurement precision deteriorates leading to misidentification errors
Solution Approach 1:
The system implements a feedback loop where the robotic device captures images, sends them to a remote computing device for human operator verification, and uses the verified images to train the machine learning model. This feedback mechanism allows the system to improve its object recognition accuracy over time while maintaining efficient automated operation.
Solution Approach 2:
The system performs preliminary data collection by capturing multiple images of objects during normal operation and sends these images ahead of time for human verification. This allows the machine learning model to be trained in advance with verified data, improving recognition accuracy before the robotic device needs to make critical identification decisions.
2Measurement precision
If the robotic device collects additional training data through human verification, then measurement precision is improved, but loss of time increases due to manual verification process
Solution Approach 1:
The system continuously captures images of objects during normal robotic operation and sends them for verification without interrupting the primary task. This continuous data collection approach allows the system to accumulate training data over time without adding separate data collection steps, thereby improving accuracy while minimizing time loss.
3Device complexity
If the robotic device uses a simple visual sensor system, then device complexity is reduced, but difficulty of detecting and measuring increases for translucent or overlapping objects
Solution Approach 1:
The system uses a remote computing device as an intermediary between the simple visual sensor on the robotic device and the machine learning model. This intermediary processes captured images, allows for human verification, and trains the model to handle difficult cases like translucent or overlapping objects, enabling simple sensors to achieve high detection accuracy through intelligent processing.
Data Source
AI summary
A method includes receiving, by a control system of a robotic device, data about an object in an environment from a remote computing device, where the data comprises at least location data and identifier data. The method further includes, based on the location data, causing at least one appendage of the robotic device to move through a predetermined learning motion path. The method additionally includes, while the at least one appendage moves through the predetermined learning motion path, causing one or more visual sensors to capture a plurality of images for potential association with the identifier data. The method further includes sending, to the remote computing device, the plurality of captured images to be displayed on a display interface of the remote computing device.