Obstacle recognition method for autonomous robots
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous robots face challenges in effectively navigating and performing tasks in complex environments due to limitations in object recognition, mapping, and path planning, particularly in identifying and avoiding obstacles like cords and pet waste while maintaining efficient navigation paths.
Innovation Solution
The implementation of an image sensor and processing unit on the robot to capture and analyze workspace images, generate feature vectors, compare objects to an object dictionary, and execute actions based on identified object types, combined with a method for creating and modifying navigation paths to avoid obstacles and maintain efficient cleaning patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the robot uses basic obstacle detection methods, then the navigation is simple and fast, but the robot cannot accurately identify specific objects like cords and pet waste
Solution Approach 1:
The patent segments the object recognition process into multiple stages: initial obstacle detection using simple sensors, followed by detailed classification using image sensors and machine learning models. This segmentation allows the system to achieve high recognition accuracy for specific objects like cords and pet waste while maintaining simple and fast navigation for areas without such objects.
2Reliability
If the robot avoids all detected obstacles, then the robot safely navigates the environment, but the navigation path becomes inefficient and time-consuming
Solution Approach 1:
The patent applies different navigation strategies based on the local characteristics of detected objects. For hazardous objects like pet waste and cords, the robot generates avoidance paths. For safe objects or open areas, the robot maintains efficient cleaning patterns. This local quality approach ensures safe navigation around dangerous obstacles while preserving cleaning efficiency in safe zones.
3Adaptability or versatility
If the robot uses comprehensive object classification, then the robot can distinguish between different object types, but the processing time and computational resources increase
Solution Approach 1:
The patent implements preliminary action by pre-training machine learning models with extensive object data before deployment. During operation, the pre-trained models can quickly classify objects like cords, pet waste, and furniture without requiring extensive real-time processing. This preliminary preparation enables comprehensive object classification while minimizing processing time during actual cleaning tasks.
Data Source
AI summary
A method for identifying objects for autonomous robots, including: capturing, with an image sensor disposed on an autonomous robot, images of a workspace, wherein a field of view of the image sensor captures at least an area in front of the autonomous robot; obtaining, with a processing unit disposed on the autonomous robot, the images; generating, with the processing unit, a feature vector from the images; comparing, with the processing unit, at least one object captured in the images to objects in an object dictionary; identifying, with the processing unit, a class to which the at least one object belongs; and executing, with the autonomous robot, instructions based on the class of the at least one object identified.