Mobile Robot Position Correction Using Image and Geomagnetic Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mobile robot devices face challenges in effectively using graph SLAM techniques due to high costs associated with 2D or 3D LiDAR, difficulty in position recognition in environments with few feature points, and limitations in robustness and accuracy, especially in environments with varying light conditions and magnetic field distortions.
Innovation Solution
A mobile robot device that combines an image sensor and multiple geomagnetic sensors to extract feature points, generate key nodes, and create a graph structure for position estimation, using algorithms like ORB and RANSAC for feature extraction and node matching, and Gaussian processes for distance difference analysis to correct the graph structure during position recognition failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2D or 3D LiDAR is used to realize graph SLAM, then position recognition accuracy is improved, but device cost increases
Solution Approach 1:
The patent combines image sensors and geomagnetic sensors into an integrated sensing system. The image sensor captures visual feature points while geomagnetic sensors detect magnetic field characteristics, and their data is fused through graph SLAM algorithms to achieve accurate position recognition without requiring expensive LiDAR hardware.
Solution Approach 2:
The patent uses image sensors to capture visual representations of the environment, creating a visual map that serves as a substitute for LiDAR's direct distance measurements. The geomagnetic sensor data provides additional spatial information, together replicating the positioning functionality of LiDAR at lower cost.
2Device complexity
If a camera is used for graph SLAM, then device cost is reduced, but position recognition becomes difficult in environments with few feature points
Solution Approach 1:
The patent merges camera-based visual feature extraction with geomagnetic sensor data. When visual feature points are scarce, the geomagnetic sensor provides alternative spatial information through magnetic field characteristic matching, ensuring continuous and accurate position recognition regardless of environmental visual features.
3Device complexity
If a camera is used for graph SLAM, then device cost is reduced, but robustness decreases due to sensitivity to light conditions
Solution Approach 1:
The patent combines image sensor data with geomagnetic sensor data to create a robust positioning system. The geomagnetic sensor provides light-independent spatial information that compensates for the image sensor's sensitivity to lighting conditions, maintaining reliable position recognition across varying environmental conditions.
4Device complexity
If geomagnetic sensors are used for position recognition, then device cost is reduced, but matching accuracy decreases in environments with insufficient magnetic field distortion
Solution Approach 1:
The patent fuses image sensor data and geomagnetic sensor data through graph SLAM algorithms. When magnetic field distortion is insufficient for accurate matching, the image sensor provides visual feature point information to supplement the geomagnetic data, maintaining matching accuracy across diverse environments.
5Measurement precision
If beacon or Wi-Fi is used for indoor position recognition, then position can be identified, but installation complexity increases and communication requirements are stricter
Solution Approach 1:
The patent enables the mobile robot to perform self-positioning using its own onboard sensors (image sensor and geomagnetic sensor) without requiring external infrastructure like beacons or Wi-Fi networks. The robot autonomously builds maps and determines its position through sensor data fusion and graph SLAM algorithms.
Data Source
AI summary
Provided are a mobile robot device and a control method thereof. The mobile robot device comprises: a driving unit; an image sensor; a plurality of geomagnetic sensors; a memory for storing at least one instruction; and a processor for executing at least one instruction, wherein the processor may obtain, while the mobile robot device moves by means of the driving unit, a plurality of image data through the image sensor and obtain sensing data through the plurality of geomagnetic sensors, extract a feature point from the plurality of image data and obtain key nodes on the basis of the feature point, obtain a node sequence on the basis of the sensing data, generate a graph structure that estimates a position of the mobile robot device on the basis of the key nodes and the node sequence, and correct the graph structure based on the mobile failing in position recognition.