Dual Sensor Robot Pickup for Random Pile Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automated pickup systems for robots face challenges in accurately identifying and picking up objects from a randomly piled container due to the need for precise three-dimensional measurements of multiple objects simultaneously, which is time-consuming and prone to errors.
Innovation Solution
A robot system equipped with a first visual sensor for obtaining height distribution information and a second visual sensor for measuring the position and posture of the target object once it is held, allowing for controlled pickup and placement in a predetermined posture and location.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If precise three-dimensional measurements are performed for multiple objects simultaneously before pickup, then the accuracy of object identification is improved, but the measurement time and processing complexity increase significantly
Solution Approach 1:
The measurement process is divided into two distinct phases: (1) a first visual sensor performs coarse height distribution measurement of multiple objects to identify the target object's approximate position, and (2) a second visual sensor performs precise three-dimensional measurement only after the robot holds the target object. This segmentation eliminates the need for simultaneous precise measurement of all objects, significantly reducing measurement time while maintaining accuracy for the selected target.
Solution Approach 2:
The system performs preliminary height distribution measurement using the first visual sensor to identify which object should be selected as the target. This preliminary action allows the robot to approach and hold the target object before initiating the time-consuming precise three-dimensional measurement, thereby optimizing the sequence of operations to minimize total measurement time.
2Measurement precision
If precise three-dimensional measurements are performed for multiple objects simultaneously, then the accuracy of target object identification is improved, but the computational complexity and data processing requirements increase
Solution Approach 1:
The computational task is segmented into two stages: (1) processing height distribution data from the first visual sensor to identify the target object's approximate location, and (2) processing three-dimensional measurement data from the second visual sensor only for the held target object. This segmentation dramatically reduces the volume of data requiring complex processing, as precise measurement data is generated for only one object rather than multiple objects simultaneously.
Solution Approach 2:
The system performs precise three-dimensional measurement only when necessary - specifically, only after the robot successfully holds the target object. This partial measurement approach avoids the excessive computational burden of measuring all objects in the container with high precision, while still ensuring accurate identification and positioning of the selected target object.
3Device complexity
If the robot attempts to identify the target object from randomly piled objects without preliminary height distribution information, then the measurement process is simplified, but the accuracy of selecting an appropriately positioned object decreases
Solution Approach 1:
The first visual sensor performs a preliminary measurement of the height distribution of objects in the container before the robot attempts pickup. This preliminary information allows the system to identify objects that are positioned favorably for pickup (e.g., objects that are higher and more accessible), thereby improving pickup accuracy without requiring complex real-time measurement during the pickup operation itself.
Solution Approach 2:
The measurement system is segmented into a first visual sensor for preliminary height distribution analysis and a second visual sensor for precise measurement during holding. This segmentation allows the system to use simple height information for target selection while reserving complex three-dimensional measurement for the critical phase when the object is being held, thus balancing complexity and accuracy.
Data Source
AI summary
A pickup device for picking up a target object from a plurality of objects randomly piled up in a container, and for placing the target object in a predetermined posture to a target location is provided. The device includes an approximate position obtaining part for obtaining information on an approximate position of the target object, based on information on a height distribution of the objects in the container, which is obtained by a first visual sensor. The device also includes a placement operation controlling part for controlling a robot so as to bring the target object into a predetermined position and posture relative to the target location, based on information on a position and posture of the target object relative to a robot, which is obtained by a second visual sensor.