Vision-Guided Robotic Picking With Virtual Stack Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current product picking systems in warehouses lack efficiency and accuracy in selecting and arranging products based on specific business rules and conditions, leading to potential errors and inefficiencies in fulfilling orders.
Innovation Solution
A system comprising a machine-readable medium with business rules for arranging products into a virtual stack, a vision module with sensors to identify product conditions, and a robotic arm to select and arrange products physically based on the virtual stack, ensuring compliance with business rules and optimal stacking configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual product picking is used, then flexibility in handling diverse products is maintained, but picking accuracy and efficiency deteriorate due to human error and time consumption
Solution Approach 1:
The system enables self-service through autonomous robotic arms that independently inspect, select, and stack products based on visual detection and business rules, eliminating the need for human operators to manually handle each product while maintaining high accuracy
Solution Approach 2:
The patent replaces manual mechanical picking with an automated vision-guided robotic system that uses optical sensors to detect products and robotic arms to execute picking actions, substituting human mechanical operations with automated mechanical systems for improved precision
2Productivity
If automated robotic picking is implemented, then picking efficiency and accuracy are improved, but system complexity and initial investment increase
Solution Approach 1:
The robotic arm system is designed with multi-functionality to perform diverse operations including product inspection, selection, and stacking in a single integrated platform, allowing the system to handle various product types and picking scenarios without requiring separate specialized equipment for each function
Solution Approach 2:
The vision module acts as an intermediary between the control system and the robotic arm, translating visual detection data into actionable picking decisions through business rule evaluation, thereby simplifying the control architecture and reducing direct system complexity
3Loss of time
If products are manually arranged in stacks, then adaptability to different stacking configurations is maintained, but time consumption and labor costs increase
Solution Approach 1:
The system performs preliminary action by pre-planning the stacking configuration through virtual stack generation based on business rules and product characteristics before actual picking begins, allowing the robotic arm to execute pre-determined stacking sequences without real-time decision-making delays
Solution Approach 2:
The stacking system incorporates dynamics by allowing the robotic arm to adapt stacking configurations in real-time based on detected product variations and space constraints, enabling flexible rearrangement of products within the stack to optimize space utilization and meet different ordering requirements
4Measurement precision
If visual inspection of each product is performed, then product condition accuracy is improved, but inspection time and processing speed decrease
Solution Approach 1:
The vision system applies partial action by selectively inspecting only the critical features and conditions of products that are relevant to picking decisions, rather than performing exhaustive inspection of all product attributes, thereby maintaining detection accuracy while reducing inspection time
Solution Approach 2:
The inspection process achieves continuity of useful action by integrating visual detection with the picking workflow, where the vision module continuously captures and processes product images in real-time during the picking operation, eliminating separate inspection steps and maintaining continuous productive action
Data Source
AI summary
The disclosed embodiments include methods and apparatus for picking products. In one embodiment, the apparatus includes a machine-readable medium containing business rules for arranging at least one product of a physical stack having physical dimensions into a virtual stack, the virtual stack being a 3D representation of the at least one product, and also containing business rules for selecting the at least one product. The apparatus also includes a vision module including a sensor configured to identify indications of a condition of each product. The apparatus also includes a processor configured to perform an inspection of each product based on the condition of the product and the set of business rules for selecting the product. The apparatus further includes a robotic arm for selecting each product of that has passed the inspection and for arranging the product to form a physical stack based on the virtual stack.