Moving robot and control method thereof
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing moving robots face challenges in reliably recognizing and avoiding obstacles, especially in narrower spaces, which affects their traveling and cleaning performance, and lack effective escape algorithms when a clear avoidance path is not detected.
Innovation Solution
The implementation of a moving robot equipped with an image acquisition unit, including depth and RGB sensors, and a control method that utilizes machine learning for obstacle recognition and avoidance, allowing the robot to autonomously detect and adapt to its environment by learning obstacle attributes and determining appropriate navigation paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If infrared or ultrasonic sensors are used for obstacle detection, then the robot can detect obstacles and avoid them, but the recognition reliability is insufficient especially in narrower spaces
Solution Approach 1:
The patent replaces traditional infrared and ultrasonic sensors with a camera-based vision system combined with machine learning algorithms. The image acquisition unit captures visual information of obstacles, and the control unit processes this data through learned models to detect and classify obstacles, substituting mechanical sensing systems with an optical-information processing system that provides higher recognition reliability.
Solution Approach 2:
The patent changes the detection parameters from physical measurements (infrared radiation, ultrasonic wave reflection) to visual parameters (color, shape, texture, depth information). By capturing images and analyzing multiple visual parameters simultaneously, the system achieves more reliable obstacle recognition, particularly in confined spaces where traditional sensors struggle.
2Adaptability or versatility
If the robot uses traditional sensor-based obstacle avoidance, then it can navigate open spaces, but it fails to detect avoidance paths in narrower spaces
Solution Approach 1:
The patent adds depth information as an additional dimension to obstacle detection by implementing a depth sensor that works alongside the RGB camera. This creates a three-dimensional understanding of the environment, allowing the robot to perceive narrow spaces more accurately and detect avoidance paths that are invisible to traditional two-dimensional sensor systems.
Solution Approach 2:
The patent implements preliminary learning of obstacle attributes and environmental features before actual navigation. The machine learning models are trained in advance to recognize various obstacle types, patterns, and spatial configurations, enabling the robot to quickly adapt to and navigate narrow spaces without real-time trial and error.
3Adaptability or versatility
If the robot lacks an escape algorithm, then the system remains simple, but the robot cannot escape from confinement scenarios
Solution Approach 1:
The patent implements a feedback-based escape algorithm that continuously monitors the robot's navigation state and obstacle configurations. When the system detects that the robot is in a confinement scenario (surrounded by obstacles with no valid avoidance path), it triggers an escape sequence that uses learned patterns to identify potential exit paths, providing adaptive response to confinement situations.
Solution Approach 2:
The escape algorithm dynamically adjusts the robot's behavior based on real-time environmental assessment. Rather than following fixed escape patterns, the system uses machine learning to evaluate the current confinement situation and generate adaptive escape trajectories, allowing the robot to handle diverse confinement scenarios effectively.
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
Enhances the robot's ability to reliably recognize and avoid obstacles, improving its traveling and cleaning performance by enabling autonomous adaptation and escape from confinement scenarios without manual input.
Implementation Method 1
The infrared sensor is to determine the presence of the obstacle and the distance from the obstacle based on the quantity of light reflected from the obstacle
Implementation Method 2
The ultrasonic sensor is to determine the distance from the obstacle using the difference between a time point at which an ultrasonic wave is emitted and a time point at which the ultrasonic wave reflected from the obstacle is received
Data Source
AI summary
A moving robot includes a travelling unit to move a main body, an image acquisition unit to acquire an image around the main body, a sensor unit including at least one sensor to sense an obstacle during moving, a storage unit to store information on a location of the sensed obstacle and information on a location of the moving robot, to register, into a map, an obstacle area of the obstacle, and to store an image from the image acquisition unit in the obstacle area, an obstacle acquisition module to determine a final attribute of the obstacle using the attributes of the obstacle obtained based on machine learning, and a control unit to detect attributes of obstacles through the obstacle recognition module if detecting a confinement state by the obstacles and to control the traveling unit to move one of the obstacles.