Conveyance Robot Classification Using Container Position Information
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing conveyance systems, such as those described in Japanese Unexamined Patent Application Publication No. 2018-103290, are unable to classify objects that cannot be detected by visual sensors, resulting in these objects being excluded from the conveyance process.
Innovation Solution
A conveyance system that includes a recognition device to identify objects of specific types and a work robot to classify objects based on recognition by the device or position information within a conveyance container, allowing unrecognized objects to be classified and included in the conveyance process without the need for physical markers, thereby reducing costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If visual sensors are used to detect objects, then objects of detectable types can be classified, but objects that cannot be detected by visual sensors are excluded from conveyance
Solution Approach 1:
The patent introduces position information as an intermediary to bridge the gap between unrecognized objects and classification. Instead of relying solely on visual recognition, the system uses the position where an object is placed in the conveyance container as a mediator to determine object type, enabling classification of objects that visual sensors cannot detect.
Solution Approach 2:
The patent replaces the optical detection mechanism (visual sensors) with a mechanical positioning mechanism. By tracking where objects are placed physically in the container rather than what objects look like visually, the system can classify objects regardless of their visual detectability.
2Measurement precision
If physical markers are attached to objects for recognition, then object identification accuracy improves, but system costs increase
Solution Approach 1:
The patent enables objects to be classified through their own placement behavior rather than requiring external markers. The position information is obtained automatically from where the object is placed in the container, eliminating the need for objects to carry or display additional identification elements.
Solution Approach 2:
The patent extracts the identification function from physical markers and transfers it to position information. By removing the need for markers and using only positional data, the system achieves identification without the added cost and complexity of physical attachment elements.
Data Source
AI summary
A conveyance system includes: a first recognition device configured to identify whether or not a target object is an object of a first type; and a work robot configured to classify the target object as the object of the first type when the target object is identified as the object of the first type by the first recognition device and to classify the target object as an object of a type corresponding to position information in a conveyance container when the target object is an object that is not identified as the object of the first type by the first recognition device.