Self-Learning Robot Object Recognition With Server Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Household robots often malfunction when encountering new objects due to their inability to accurately recognize and respond to them, leading to potential damage and the need for user intervention.
Innovation Solution
A self-learning robot system that includes a data reception unit, data recognition unit, recognition result verification unit, and server communication unit, allowing the robot to autonomously recognize new objects, verify recognition accuracy, and update its database for appropriate behavior without user input.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the robot uses a fixed database for object recognition, then the device complexity is reduced, but the adaptability to new objects deteriorates
Solution Approach 1:
The robot performs preliminary recognition attempts using its fixed database before initiating server communication. This preliminary action allows the robot to handle common objects independently, reducing server dependency, while still maintaining the capability to adapt to new objects through server-based learning when preliminary recognition fails.
Solution Approach 2:
The server acts as an intermediary between the robot's fixed database and new objects. When the robot encounters an unrecognized object, the server mediates by providing recognition results and updating the robot's database, thus bridging the gap between the fixed local database and the need for adaptability to new objects.
2Ease of operation
If the robot autonomously learns new objects without user input, then the ease of operation is improved, but the reliability of accurate recognition may deteriorate
Solution Approach 1:
The robot implements a feedback mechanism where recognition results are verified against a predetermined accuracy threshold. When recognition accuracy is insufficient, the robot automatically requests server assistance and uses the server's correction feedback to update its database, thus maintaining reliability while achieving autonomous learning.
Solution Approach 2:
The robot performs self-service by autonomously initiating server communication when recognition fails and automatically updating its own database with learned information. This self-service capability enables autonomous learning while maintaining reliability through systematic verification and correction processes.
3Adaptability or versatility
If the robot communicates with the server for every unrecognized object, then the adaptability to new objects is improved, but the loss of time increases
Solution Approach 1:
The robot applies partial action by communicating with the server only when necessary (i.e., when local recognition fails), rather than for every object. This selective communication approach maintains adaptability for new objects while minimizing time loss by avoiding unnecessary server interactions for recognized objects.
Solution Approach 2:
The robot performs preliminary recognition using its local database before initiating server communication. This preliminary action filters out commonly recognized objects, ensuring that server communication occurs only for truly unrecognized objects, thus reducing overall communication time while maintaining adaptability.
4Reliability
If the robot performs verification of recognition results, then the reliability is improved, but the device complexity increases
Solution Approach 1:
The verification mechanism uses feedback from server correction results to update the robot's database. This feedback-based verification improves reliability by ensuring accurate recognition while managing complexity through automated database updates rather than complex real-time verification algorithms.
Solution Approach 2:
The robot performs self-service verification by automatically comparing recognition results against the predetermined accuracy threshold and autonomously initiating server communication when needed. This self-service approach improves reliability without requiring complex external verification systems.
Data Source
AI summary
A self-learning robot, according to one embodiment of the present invention, comprises: a data receiving unit for sensing video data or audio data relating to an object located within a predetermined range; a data recognition unit for matching data received from the data receiving unit and data included in a database in the self-learning robot; a result output unit for outputting a matching result from the data recognition unit; a recognition result verifying unit for determining the accuracy of the matching result; a server communication unit for transmitting data received from the data receiving unit to a server, when the accuracy of the matching result determined by the recognition result verifying unit is lower than a predetermined level; and an action command unit for causing the self-learning robot to perform a pre-set object response action, when the accuracy of the matching result determined by the recognition result verifying unit is at least the predetermined level.