Robotic Pallet Stacking Using Physics-Based Collision Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for stacking dissimilar items on pallets, such as in shipping and distribution centers, are inefficient and unstable due to the variety of items and human reliance on intuition, making it challenging for robotics to handle diverse and unpredictable item arrangements.
Innovation Solution
A robotic system using vision data and geometric data, combined with a physics engine, determines optimal item placements and trajectories to ensure stability and efficiency by simulating and evaluating stack interactions, employing machine learning for decision-making and collision avoidance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If human workers manually stack items using judgment and intuition, then item placement stability is improved, but productivity is reduced
Solution Approach 1:
The patent replaces the mechanical human decision-making process with a computational system that uses vision data, geometric data, and physics engine simulations to determine optimal item placements. The control computer processes sensor data and runs physics-based simulations to predict stable configurations, substituting human intuition with automated computational analysis while maintaining stability requirements.
Solution Approach 2:
The system performs preliminary physics engine simulations and stability evaluations before actual palletization occurs. By pre-calculating optimal placements using vision and geometric data, the system determines the best configuration in advance, allowing robotic systems to execute pre-planned stable arrangements without real-time human intervention, thus improving both speed and stability.
2Productivity
If robotic systems are used to automate palletization, then productivity is improved, but reliability is reduced due to difficulty handling diverse item arrangements
Solution Approach 1:
The patent introduces vision systems and physics engine simulations as intermediary components between the robotic system and physical items. The vision system captures geometric data about diverse items, and the physics engine simulates various placement scenarios to predict stability outcomes. This intermediary computational layer enables the robotic system to reliably handle diverse item arrangements by translating visual information into actionable placement decisions.
Solution Approach 2:
The system dynamically adjusts placement parameters based on real-time vision data and item characteristics. By changing parameters such as placement position, orientation, and stacking sequence based on detected item geometry and physics simulations, the robotic system adapts to diverse item variations, maintaining high reliability across different item types and arrangements.
3Productivity
If items are densely packed to maximize space utilization, then productivity is improved, but stability is reduced
Solution Approach 1:
The patent applies partial physics engine simulations to evaluate multiple potential placements, selecting configurations that achieve sufficient density without compromising stability. Rather than maximizing density to the extreme, the system finds optimal balance points where space utilization is high enough to be productive while maintaining adequate stability margins, avoiding excessive packing that would cause collapse.
Solution Approach 2:
The system uses physics engine simulation results as feedback to adjust placement decisions. By evaluating predicted stability outcomes from simulated placements and using this feedback to select optimal configurations, the system iteratively determines placements that balance density and stability, ensuring densely packed arrangements remain stable under gravitational and handling forces.
Data Source
AI summary
A robotic system is disclosed. The system includes a communication interface configured to receive, from one or more sensors deployed in a workspace, sensor data indicative of a current state of the workspace, the workspace comprising a pallet or other receptacle and a plurality of items stacked on or in the receptacle. The system includes one or more processors that use a geometric model based at least in part on past item placements in combination with the sensor data to estimate a state of the pallet or other receptacle and one or more items stacked on or in the pallet or other receptacle, and use the estimated state to generate or update a plan to control a robotic arm to place a next item on or in, or remove a next item from, the pallet or other receptacle in a manner that avoids having the next item collide with any other item stacked on or in the pallet or other receptacle.