AI Mobile Robot Obstacle Learning for Faster Avoidance Response
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional mobile robots face limitations in recognizing and responding to various obstacles while cleaning, often resulting in collisions and errors due to delayed obstacle recognition and limited ability to adapt to new environments, leading to ineffective navigation and potential damage.
Innovation Solution
A mobile robot equipped with an image acquirer and obstacle detection unit that captures images while traveling, analyzes obstacle features through pre-captured images, and determines response motions based on obstacle type, allowing for immediate avoidance and adaptation to new obstacles by transmitting data to a server for learning and updating obstacle information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the mobile robot uses sensor-based obstacle detection (infrared rays or laser beams), then the robot can detect obstacles within a predetermined distance, but the robot responds too late and may collide with the obstacle causing damage
Solution Approach 1:
The robot performs preliminary actions by capturing images continuously before obstacle detection is triggered. When an obstacle is detected by sensors, the system already has pre-captured images ready for immediate analysis, eliminating the time delay between detection and response. This allows the robot to identify obstacle types and select appropriate response motions without waiting for image capture after detection.
Solution Approach 2:
The system skips the sequential process of first detecting an obstacle and then capturing images. Instead, it rushes through the process by having images already captured and stored before detection occurs, allowing immediate obstacle type identification and response motion selection upon sensor detection.
2Reliability
If the mobile robot immediately changes path when an obstacle is detected, then the robot avoids collision, but the robot cannot clean the corresponding region
Solution Approach 1:
The robot dynamically adjusts its response based on obstacle type identification. Instead of always taking the same avoidance action, the system selects from multiple response motions (avoidance, approach, stop) depending on the specific obstacle type recognized from image analysis, optimizing both safety and cleaning efficiency for different situations.
Solution Approach 2:
The system changes the parameter of response motion selection based on obstacle characteristics. By analyzing image data to identify obstacle types (e.g., furniture, pets, people), the robot adjusts its behavior parameters to choose appropriate actions that balance collision avoidance with cleaning coverage maintenance.
3Reliability
If the mobile robot uses conventional obstacle recognition methods, then the robot can recognize known obstacles, but the robot cannot adapt to new or unknown obstacles in restrictive test environments
Solution Approach 1:
The system implements feedback by continuously capturing images, analyzing obstacle types, and using this information to improve future obstacle recognition. The server stores recognized obstacle types and their characteristics, providing feedback that helps the robot adapt to new obstacles encountered in different environments beyond restrictive test conditions.
Solution Approach 2:
The robot performs self-service by autonomously capturing images, analyzing obstacle types using image processing algorithms, and selecting appropriate response motions without external intervention. This self-service capability enables the robot to adapt to new environments and obstacles independently, beyond what was possible in controlled test environments.
Data Source
Figure 1(a)~1(b)
Figure 2(a)~2(b)
Figure 3(a)~3(b)
AI summary
An artificial intelligence (AI) mobile robot and a method of controlling the same for learning an obstacle are configured to capture an image while traveling through an image acquirer, to store a plurality of captured image data, to determine an obstacle from image data, to set a response motion corresponding to the obstacle, and to operate the set response motion depending on the obstacle, and thus, the obstacle is recognized through the captured image data, the obstacle is easily determined by repeatedly learning an image, and the obstacle is determined before the obstacle is detected or from a time point of detecting the obstacle to perform an operation of a response motion, and even if the same detection signal is input when a plurality of different obstacles is detected, the obstacle is determined through the image and different operations are performed depending on the obstacle to respond to various obstacles, and accordingly, the obstacle is effectively avoided and an operation is performed depending on a type of the obstacle.