Mobile Robot Obstacle Attribute Recognition Using Multi-Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing moving robots lack the ability to accurately determine the attribute of obstacles, leading to inefficient obstacle recognition and avoidance, which affects their driving and cleaning performance, as they cannot differentiate between movable and immovable objects, resulting in potential collisions and disturbances to users.
Innovation Solution
A moving robot equipped with an image acquisition unit, sensor unit, and controller that recognizes obstacle attributes using machine learning, allowing it to adjust its driving pattern based on the identified attributes and output preset sounds for movable obstacles, thereby enhancing stability, user convenience, and cleaning efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the robot uses simple obstacle detection methods (infrared or ultrasonic sensors only), then the device complexity is low, but the measurement precision of obstacle attributes is insufficient
Solution Approach 1:
The patent combines multiple sensors (infrared sensor, ultrasonic sensor, and camera) into an integrated obstacle detection system. The infrared sensor detects thermal radiation from obstacles, the ultrasonic sensor measures distance through sound wave reflection, and the camera captures visual information. By merging these different sensing modalities, the system achieves comprehensive obstacle attribute recognition including type classification, distance measurement, and visual identification, thereby resolving the contradiction between measurement precision and device complexity.
Solution Approach 2:
The camera serves as an intermediary device that bridges the gap between simple distance detection and complex obstacle understanding. While infrared and ultrasonic sensors provide basic distance and presence information, the camera captures detailed visual characteristics of obstacles. The controller processes camera images to identify obstacle types (e.g., wires, steps, movable objects), transforming simple sensor data into rich attribute information without requiring direct complex sensing from the other sensors.
2Reliability
If the robot follows simple avoidance rules based on obstacle height, then the ease of operation is high, but the reliability of avoidance behavior is insufficient
Solution Approach 1:
The patent segments obstacle recognition into multiple classification categories: movable obstacles (pets, people, shoes), immovable obstacles (wires, steps, fan bases), and unknown obstacles. Each category has specific avoidance behaviors programmed in the controller. This segmentation allows the system to handle different obstacle types with specialized strategies, improving reliability by ensuring appropriate responses for each category while maintaining ease of operation through rule-based control logic.
Solution Approach 2:
The controller performs preliminary classification of obstacles using camera image recognition before executing avoidance maneuvers. By identifying obstacle type in advance (e.g., detecting a wire before collision), the system can select the appropriate avoidance strategy beforehand - such as rotating 90 degrees for wires, jumping for steps, or stopping for movable obstacles. This preliminary action ensures reliable avoidance behavior while keeping the control logic organized and manageable through predefined response rules.
3Productivity
If the robot performs aggressive avoidance maneuvers, then the productivity of obstacle avoidance is high, but the object-affected harmful factors increase (disturbing users)
Solution Approach 1:
The patent applies different avoidance behaviors based on the local characteristics of detected obstacles. For movable obstacles like pets or people, the robot performs gentle maneuvers such as stopping or slow rotation to minimize disturbance. For immovable obstacles like wires or steps, the robot executes more decisive actions like 90-degree rotations or jumps to ensure effective avoidance. This localized adaptation of avoidance intensity resolves the contradiction by matching response aggressiveness to obstacle type, maintaining high productivity while reducing harmful effects on users.
Solution Approach 2:
The avoidance behavior is made dynamic and adaptive based on real-time obstacle classification. The controller adjusts the avoidance strategy according to the identified obstacle type: gentle stopping for movable obstacles, moderate rotation for medium-risk obstacles, and aggressive maneuvers for hard-to-avoid obstacles. This dynamic adjustment allows the system to maintain high avoidance effectiveness while minimizing user disturbance, as the response intensity is continuously optimized based on the current obstacle situation.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The robot can accurately recognize obstacle attributes and adjust its movement accordingly, improving its ability to avoid obstacles and interact with users, resulting in enhanced stability and efficiency in navigation and cleaning tasks.
Implementation Method 1
The infrared sensor determines the presence of the obstacle and a distance thereto based on an amount of light reflected from the obstacle and a time taken to receive the reflected light
Implementation Method 2
The ultrasonic sensor emits ultrasonic waves, and, if there is an ultrasonic wave reflected by an obstacle, the ultrasonic sensor determines a distance to the obstacle based on a time difference between when the ultrasonic wave is emitted and when the reflected ultrasonic wave is received
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
Disclosed is a moving robot including: a travel unit configured to move a body; an image acquisition unit configured to acquire a surrounding image of the body; a sensor unit having one or more sensors configured to detect an obstacle while the body moves; a controller configured to: upon detection of an obstacle by the sensor unit, recognize an attribute of the obstacle based on an image acquired by the image acquisition unit, and control driving of the travel unit based on the attribute of the obstacle; and a sound output unit configured to: output preset sound when the recognized attribute of the obstacle indicates a movable obstacle. Accordingly, the moving robot improves stability, user convenience, driving efficiency, and cleaning efficiency.