Robot and control method therefor
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cleaning robots are inefficient in cleaning wide areas due to limited sensor capabilities, leading to uncomfortable user experiences and potential damage to sensitive obstacles.
Innovation Solution
A cleaning robot equipped with a camera and AI models to capture and analyze images, differentiate between obstacles and structures, and adjust its movement speed and direction based on recognized data, generating a navigation map for efficient cleaning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a cleaning robot uses only local sensors to detect obstacles, then the device complexity is reduced, but the measurement precision and sensing range are limited
Solution Approach 1:
The patent combines multiple sensing modalities (camera for visual recognition, depth sensor for distance measurement, and local sensors for proximity detection) into an integrated sensing system. This merging allows the robot to overcome the limitations of individual sensors and achieve both wide sensing range and high measurement precision simultaneously.
Solution Approach 2:
The patent transitions from relying solely on local proximity sensors to incorporating a camera that provides two-dimensional visual information and depth sensors that add three-dimensional spatial awareness. This dimensional enhancement enables the robot to detect obstacles and structures at greater distances with higher precision.
2Device complexity
If the cleaning robot approaches obstacles until detected by local sensors, then the device complexity remains low, but the reliability of obstacle protection is insufficient
Solution Approach 1:
The patent implements preliminary obstacle detection using a camera and depth sensor before the robot approaches the obstacle. The system identifies obstacles and determines safe distances in advance, allowing the robot to adjust its trajectory and avoid potential collisions before they occur, thereby enhancing reliability without significantly increasing complexity.
Solution Approach 2:
The patent employs a feedback mechanism where the camera continuously captures images of the environment, the system processes this visual information to identify obstacles and structures, and the robot adjusts its movement based on this feedback. This closed-loop control ensures reliable obstacle protection while maintaining manageable system complexity.
3Productivity
If the cleaning robot uses a wide sensing range to clean large areas, then the productivity is improved, but the difficulty of detecting and measuring obstacles increases
Solution Approach 1:
The patent introduces an image processing algorithm as an intermediary between the camera input and obstacle detection. This intermediary component processes the visual information to extract meaningful features, segment the environment into navigable areas and obstacles, and provide structured data for navigation decisions, thereby simplifying the detection and measurement process while maintaining wide sensing range.
Solution Approach 2:
The patent segments the captured image into different regions (navigable floor areas, obstacles, structures, walls) using image processing techniques. This segmentation transforms the complex visual scene into discrete, easily identifiable elements, reducing the difficulty of obstacle detection and measurement while enabling efficient navigation across large cleaning areas.
Data Source
AI summary
The present invention relates to a robot and a control method therefor and, more particularly, to a robot and a control method therefor, the robot detecting peripheral obstacles and structures by using an image in front of same, thereby providing an efficient and safe moving path. A robot control method according to one embodiment of the present invention comprises the steps of: capturing an image; acquiring a first image of which the bottom surface is separated in the image; recognizing obstacles included in the image by using a learned first artificial intelligence model; allowing a robot to acquire and store a second image in order to perform a task on the basis of the information about the first image and the region including the recognized obstacles; and allowing the robot to move on the basis of the second image.