Robot Localization Using LiDAR-Camera Fusion Under Low Sensor Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current robot localization methods require increased sensor accuracy and map information, which can be compromised by low accuracy from individual sensors or multiple detected positions, leading to reduced localization precision.
Innovation Solution
A robot utilizing a combination of LiDAR and camera sensors with artificial intelligence to generate and compare frames, calculate positions, and enhance accuracy by selecting the most accurate sensor data for localization, even when individual sensors have low accuracy or detect multiple positions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of sensors is increased to improve localization accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent combines data from multiple sensors (LiDAR, camera, odometry) into a unified localization framework. The sensor fusion algorithm integrates measurements from different sensor types to produce a single, more accurate position estimate, thereby improving localization precision without proportionally increasing system complexity through redundant sensor installations.
Solution Approach 2:
The localization system is designed to accept and process data from multiple sensor types universally. The framework can accommodate LiDAR, camera, odometry, and other sensors through a common interface and fusion algorithm, allowing the system to leverage diverse sensor capabilities without requiring separate processing pipelines for each sensor type.
2Measurement precision
If sensor accuracy is improved to enhance localization precision, then measurement precision is improved, but cost increases
Solution Approach 1:
The patent changes the parameter of sensor integration rather than relying on a single high-accuracy sensor. By transforming the localization approach from single-sensor high-precision to multi-sensor fused data, the system achieves improved accuracy through parameter combination rather than investing in expensive high-precision individual sensors.
3Measurement precision
If map information is increased to improve localization accuracy, then measurement precision is improved, but loss of information increases due to data management complexity
Solution Approach 1:
The patent extracts and compares specific features from sensor data against corresponding features in the stored map. Rather than managing and processing all raw sensor data and complete map information, the system extracts relevant localization features (such as landmark positions, edge configurations) for comparison, reducing data management overhead while maintaining localization accuracy.
4Measurement precision
If the number of sensors is increased to improve localization accuracy, then measurement precision is improved, but ease of operation decreases
Solution Approach 1:
The localization system operates autonomously by automatically fusing data from multiple sensors and comparing it with map information. The sensor fusion algorithm self-manages the integration of LiDAR, camera, and odometry data without requiring manual intervention or complex operational procedures, thereby maintaining ease of operation despite using multiple sensors.
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 method enables robust and accurate robot localization by integrating LiDAR and camera sensor data, improving position estimation and map accuracy, and compensating for low sensor performance, thereby enhancing overall localization precision.
Implementation Method 1
a LiDAR sensor sensing a distance between an object placed outside of the robot and the robot and generating a LiDAR frame
Implementation Method 2
a camera sensor capturing an image of an object placed outside of the robot and generating a visual frame
Data Source
AI summary
Disclosed herein are a method of localization using multi sensors and a robot implementing the same, the method including sensing a distance between an object placed outside of a robot and the robot and generating a first LiDAR frame by a LiDAR sensor of the robot while a moving unit moves the robot, capturing an image of an object placed outside of the robot and generating a first visual frame by a camera sensor of the robot, and comparing a LiDAR frame stored in a map storage of the robot with the first LiDAR frame, comparing a visual frame registered in a frame node of a pose graph with the first visual frame, determining accuracy of comparison's results of the first LiDAR frame, and calculating a current position of the robot by a controller.