Placement-Informed Robotic Picking for Tight Bin Packing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robotic pick and place systems often operate independently, leading to inefficient space utilization and suboptimal object placement due to the lack of integration between the pick and place processes, especially in tight bin packing scenarios.
Innovation Solution
A method that integrates placement information into the pick process by computing object masks and place region masks, calculating costs for object-region pairs based on place constraints, and selecting objects to be picked based on these costs to optimize both object selection and placement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the pick and place systems operate independently, then the system complexity is reduced and ease of operation is improved, but space utilization efficiency deteriorates and packing optimization is worsened
Solution Approach 1:
The patent merges the pick and place systems by integrating place information into the pick process. The place system determines placement poses and feeds this information back to the pick system, which then selects objects considering both pick and place constraints. This unified approach optimizes space utilization while maintaining system operability through integrated decision-making.
Solution Approach 2:
The patent implements feedback by having the place system provide placement information back to the pick system. The place constraints and placement poses computed by the place system are used by the pick system to make informed object selection decisions, creating a closed-loop system that improves packing efficiency.
2Productivity
If the pick system selects objects without considering place information, then the pick process is simplified and faster, but the overall packing efficiency and bin utilization deteriorate
Solution Approach 1:
The patent applies preliminary action by having the place system pre-compute placement poses and place constraints before the pick system selects objects. This advance preparation of placement information allows the pick system to make informed decisions that optimize packing efficiency without significantly delaying the pick process.
Solution Approach 2:
The patent introduces place information as an intermediary between the pick and place systems. This intermediary data, including placement poses and place constraints, enables the pick system to select objects that will optimize final packing efficiency while maintaining operational speed.
Data Source
AI summary
A method for performing placement informed robotic picking of objects includes acquiring a first image of a pick scene including a number of objects and acquiring a second image of a placement area that receives objects picked from the pick scene by a robot. Object masks are computed by performing instance segmentation based on the first image. Place region masks are computed by clustering locations in the second image based on a height level from a floor of the placement area. A cost is computed for respective object-region pairs, each object-region pair defining a pairing between an object mask and a place region mask. The cost is defined at least in part by a place constraint. An object is selected to be picked from the pick scene by the robot by selecting an object-region pair based on the computed cost.