Image-Based Robot Control for Cluttered Bin Picking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robotic systems face challenges in efficiently picking objects from cluttered or randomly arranged environments, where access is blocked by other objects or clusters, leading to slowed operations in bin picking scenarios, especially in automated manufacturing and packaging settings.
Innovation Solution
An image-based system that acquires images of the object area, computationally analyzes the objects, determines preferred non-object picking actions to move objects without physical picking, such as tilting, shaking, or redistributing them to create a more uniform surface, and transmits signals to actuators or robotic controllers to execute these actions, improving object accessibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the robot directly attempts to pick objects from cluttered environments, then the picking operation can proceed, but access is blocked by other objects or clusters significantly slowing down the operation
Solution Approach 1:
The system performs preliminary actions by analyzing the cluttered environment with image sensors and computationally determining preferred non-object picking actions (such as shaking, tilting, or redistributing objects) before the robot attempts to pick objects. This preliminary analysis and preparation of the environment enables faster and more efficient object access without direct robot intervention for each obstacle removal.
2Measurement precision
If the robot uses image sensors and machine vision to locate objects, then the system can identify object positions, but clusters of parts or objects still prevent effective picking
Solution Approach 1:
The system introduces an intermediary computational layer between image sensing and robot picking. This intermediary processes image data to identify clusters and determines preferred non-object picking actions (such as shaking or redistributing) that will disperse clusters and improve object accessibility, thereby enhancing picking effectiveness beyond what simple location detection can achieve.
3Device complexity
If the system employs a robot motion controller as the central control structure, then all information passes through it, but the system lacks the capability to determine environmental modifications for improved picking
Solution Approach 1:
The robot motion controller is enhanced with multi-functionality by integrating image sensor processing and computational determination of preferred non-object picking actions. This allows the existing central control structure to not only control robot motion but also analyze the environment and determine environmental modifications (such as shaking or tilting) that will improve picking conditions, thereby increasing adaptability without adding separate complex control systems.
Data Source
AI summary
Vision based systems may select actions based on analysis of images to redistribute objects. Actions may include action type, action axis and/or action direction. Analysis may determine whether an object is accessible by a robot, whether an upper surface of a collection of objects meet a defined criteria and/or whether clusters of objects preclude access.