Simulated Box Placement for Stable Mixed-Item Palletizing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for palletizing and depalletizing heterogeneous items are inefficient due to the variety of items, variations in size, weight, and mix, making it challenging for robotics to stabilize and arrange items effectively, leading to unstable stacks and increased resource intensity.
Innovation Solution
A system that uses a combination of geometric modeling, simulation, and machine learning to determine optimal placement of items on a pallet by simulating different scenarios and using a scoring function to evaluate stability and efficiency, allowing for simultaneous processing of item placement and robotic control to speed up the cycle time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If human workers manually select and stack items based on judgment and intuition, then the palletized stack stability is improved, but the productivity and resource intensity worsen
Solution Approach 1:
The patent replaces the mechanical system of manual human stacking with a robotic system that uses computer vision and machine learning algorithms to determine optimal placement. The robotic system captures images of items, processes them through neural networks to predict stable configurations, and automatically positions items on pallets, thereby maintaining stability while dramatically increasing productivity.
Solution Approach 2:
The system performs preliminary analysis of item attributes (weight, dimensions, center of gravity) and uses machine learning models to predict stable stacking configurations before actual placement. This pre-computation of optimal arrangements allows the robotic system to execute placements efficiently without real-time decision delays.
2Ease of operation
If items are selected and stacked in arbitrary order from conveyor or bins, then the ease of operation is improved, but the stability of the palletized stack worsens
Solution Approach 1:
The system implements feedback loops where the robotic system continuously monitors the pallet building process, captures images of placed items, and uses machine learning models to evaluate stack stability in real-time. Based on this feedback, the system adjusts subsequent item selection and placement decisions to maintain or improve stability, even when items arrive in arbitrary orders.
Solution Approach 2:
The system dynamically changes placement parameters (position, orientation, layer configuration) based on item attributes and current stack state. The machine learning model evaluates multiple possible placements and selects the optimal one that maximizes stability, allowing the system to adapt to arbitrary item sequences while maintaining stable configurations.
3Stability of the object's composition
If computational models and simulations are used to determine optimal item placement, then the stability and efficiency of palletization is improved, but the device complexity and computational burden increase
Solution Approach 1:
The patent divides the complex palletization problem into separate modular components: computer vision module for item recognition, attribute extraction module for measuring physical properties, machine learning module for stability prediction, and robotic control module for execution. Each module operates independently with defined interfaces, reducing overall system complexity while maintaining high stability through coordinated operation.
Solution Approach 2:
The system introduces an intermediary layer of machine learning models that act as mediators between the raw sensor data and the robotic control system. These models pre-process and interpret complex data, providing simplified decision recommendations to the robotic system, thereby reducing the computational burden on the control system while maintaining accurate stability predictions.
Data Source
AI summary
A robotic system is disclosed. The system includes a memory that stores for each of a plurality of items a set of attribute values. The system includes a processor(s) that uses the attribute values to simulate the placement of items, including by determining, iteratively, for each next item a placement location at which to place the item on a simulated stack of items on the pallet, using the attribute values and a geometric model of where items have been simulated to have been placed to estimate a state of the stack after each of a subset of simulated placements, and using the estimated state to inform a next placement decision. The steps of determining for each next item a placement location and estimating the state of the stack until all of at least a subset of the plurality of items have been simulated as having been placed on the stack.