Apparatus for localizing cleaning robot, cleaning robot, and controlling method of cleaning robot
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing localization methods for cleaning robots, such as the Monte Carlo localization method, are inefficient in determining the current position and traveling direction, especially when the number of particles is small or when the robot is moved to a different position, leading to increased calculation time and potential missed cleaning areas.
Innovation Solution
The method involves classifying straight lines in the local map by angles to determine the main segment angle, acquiring actual sensor data while rotating 360 degrees, and using a ray casting method to extract virtual sensor data for accurate position determination and direction adjustment, allowing the robot to quickly and accurately recognize its position and direction without unnecessary movement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If the Monte Carlo localization method is used with a small number of particles, then the calculation time is reduced, but the localization accuracy deteriorates
Solution Approach 1:
The system performs preliminary actions by extracting straight line information and angle distribution data from the map before localization. By pre-processing the environmental map to identify walls and obstacles as straight line segments with their angles, the system prepares reference data that can be quickly compared with sensor readings during localization, eliminating the need for time-consuming particle simulations while maintaining accuracy
Solution Approach 2:
The patent replaces the mechanical particle filter simulation process with an information processing approach based on straight line angle matching. Instead of using physical particles that randomly move and converge, the system uses sensor data to detect straight lines and compares their angles with pre-extracted map features, substituting a computational geometric approach for the stochastic mechanical simulation
2Ease of operation
If the cleaning robot moves to an arbitrary position by user, then the robot can be repositioned for efficient cleaning, but the localization time increases due to meaningless movement required by conventional methods
Solution Approach 1:
The system creates a copy of the environmental map and extracts straight line information from it. By having a digital representation of the environment with pre-extracted geometric features, the robot can perform localization through data comparison rather than physical movement, allowing instant position determination even after user repositioning without requiring the robot to move and scan the environment again
3Productivity
If the cleaning robot determines traveling direction without accurate localization, then the robot can start cleaning immediately, but cleaning coverage deteriorates with missed areas
Solution Approach 1:
The system uses feedback from sensor data by detecting straight lines in the environment and comparing them with the map data to determine the robot's position and orientation. This feedback mechanism provides continuous verification of the robot's state, enabling accurate traveling direction determination and ensuring complete cleaning coverage while maintaining high productivity
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A cleaning robot includes a data acquisition unit that acquires actual sensor data by measuring a distance from a current position to an object to be measured; a local map acquisition unit that acquires a local map by scanning the vicinity of the current position based on an environmental map stored in advance; and a processor that determines coordinates of the current position for the local map by performing matching between the local map and the actual sensor data, and determines a traveling direction based on the current position by calculating a main segment angle of a line segment existing in the local map.