AGV Vision Supervision for Real-Time On-Site Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automated guided vehicles (AGVs) often perform poorly in real-world environments that differ significantly from the default datasets used for training, leading to low recognition rates due to variations in lighting, worker uniforms, and facility infrastructure, making it difficult for manufacturers to adapt the systems to each customer's facility.
Innovation Solution
An AGV system equipped with a mobile base, two cameras (one with a content filter for marker recognition), and a main control module that executes a recognition neural network program and a supervisor program, allowing for real-time supervised learning by updating the default dataset with user-input markers and real-world information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the AGV uses a default dataset for training the neural network, then the system can be manufactured and deployed quickly, but the recognition performance deteriorates when the actual environment differs from the default dataset
Solution Approach 1:
The AGV system performs self-training by automatically collecting images from its operating environment, generating pseudo-labels through its existing neural network, and updating its dataset without requiring manual intervention from manufacturers or customers. This enables the system to adapt to specific facility conditions while maintaining quick deployment.
Solution Approach 2:
The system dynamically updates the neural network's training dataset by incorporating new images from the actual operating environment, changing the data parameters to match real-world conditions. This allows the recognition model to adapt to variations in lighting, worker uniforms, and facility infrastructure.
2Reliability
If the manufacturer creates custom datasets for each customer facility, then the recognition performance improves, but the complexity and cost of deployment increases significantly
Solution Approach 1:
The system eliminates the need for manufacturer intervention in dataset customization by enabling the AGV to automatically collect, label, and integrate new images from its operating environment. The neural network generates pseudo-labels automatically, removing the need for manual annotation processes.
Solution Approach 2:
The system continuously collects images from the operating environment and uses them to update the training dataset, creating a feedback loop that progressively improves recognition performance. The supervisor program monitors recognition outcomes and triggers retraining when performance degradation is detected.
3Ease of operation
If the AGV operates in a self-navigation mode without real-time learning, then the system is simpler to operate, but the system cannot adapt to environmental changes
Solution Approach 1:
The AGV performs self-learning automatically during normal operations without requiring operators to switch between training and navigation modes. The system autonomously collects images, generates pseudo-labels, and updates its neural network in the background, maintaining operational simplicity while enabling continuous adaptation.
Solution Approach 2:
The system dynamically adjusts its learning process based on operational conditions, using the supervisor program to monitor recognition performance and trigger retraining only when necessary. This dynamic approach balances operational simplicity with adaptability by activating learning functions selectively.
Data Source
AI summary
An automatic guided vehicle (AGV) comprising a mobile base including a drive train and configured to drive the AGV in a self-navigation mode within a facility, a first camera configured to capture first image data of objects within the facility, a second camera configured to capture second image data of objects within the facility, the second camera including a content filter, and a main control module configured to receive the first and second image data from the first and second cameras. The main control module executes a recognition neural network program. The neural network program recognizes targets in the first image data. The main control module also executes a supervisor program under user control. The supervisor program is receives the second image data and recognizes markers attached to targets in the second image data. The supervisor program produces a supervised outcome in which targets to which markers are attached are associated with categories based on user commands. The supervised outcome adjusts weights of nodes in the recognition neural program.