Multi-Sensor Pose Graph Mapping for Accurate Robot Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robots face challenges in generating accurate maps of spaces using multiple sensors, leading to errors in position identification and navigation due to sensor limitations and movement processes.
Innovation Solution
A robot system utilizing a combination of LiDAR and camera sensors, along with artificial intelligence, generates a map by creating a pose graph with LiDAR and visual branches, and sets correlations between nodes using wheel odometry information, LiDAR, and visual odometry to enhance the accuracy of Simultaneous Localization and Mapping (SLAM) and navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a robot uses multiple sensors to generate a map, then the map quality and position identification accuracy are improved, but errors are produced due to sensor features and movement processes
Solution Approach 1:
The patent combines data from multiple sensors (LiDAR, camera, wheel encoder) to generate a fused map. The controller integrates LiDAR frames, visual frames, and wheel odometry information to create a comprehensive pose graph that leverages the strengths of each sensor while compensating for individual weaknesses, thereby improving position identification accuracy while managing error propagation
Solution Approach 2:
The system implements feedback mechanisms by continuously comparing sensor measurements with the generated map and adjusting the pose graph accordingly. The controller uses the correlations between nodes to detect and correct errors in real-time, ensuring that position identification accuracy is maintained despite sensor limitations and movement-induced errors
2Measurement precision
If a robot uses various sensors to generate and correct maps, then map accuracy is improved, but device complexity increases
Solution Approach 1:
The controller serves multiple functions by simultaneously processing data from LiDAR, camera, and wheel encoders, generating the pose graph, calculating node correlations, and performing error correction. This multi-functional approach consolidates the complexity into a single processing unit rather than requiring separate dedicated systems for each function
Solution Approach 2:
The system creates a composite data structure (pose graph) that integrates information from heterogeneous sensors. By fusing LiDAR frames, visual frames, and wheel odometry into a unified representation with defined node correlations, the system manages the complexity of multiple sensors through a structured composite framework
3Measurement precision
If a robot stores detailed map information and node correlations, then navigation accuracy is improved, but information storage requirements increase
Solution Approach 1:
The map is segmented into discrete nodes and edges forming a pose graph structure. Instead of storing continuous detailed information, the system divides the environment into key positional nodes with correlated relationships, reducing storage requirements while maintaining navigation accuracy through the graph's topological structure
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 system allows for precise map generation and accurate robot positioning, improving navigation and reducing errors by fusing data from multiple sensors, thereby enhancing the accuracy of SLAM and map quality.
Implementation Method 1
a LiDAR sensor sensing a distance between an object outside the robot and the robot and generating a LiDAR frame
Implementation Method 2
a camera sensor photographing an object outside the robot and generating a visual frame
Data Source
AI summary
Disclosed herein are a robot that generates a map and configures a correlation of nodes, based on multi sensors and artificial intelligence, and that moves based on the map, and a method of generating a map, and the robot according to an embodiment generates a pose graph comprised of LiDAR branch, visual branch, and backbone, and the LiDAR branch includes one or more of the LiDAR frames, the visual branch includes one or more of the visual frames, and the backbone includes two or more frame nodes registered with any one or more of the LiDAR frames or the visual frames, and to generate a correlation between nodes in the pose graph.