2D LiDAR Marker Tracking for Low-Cost Mobile Robot Pose Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mobile robot systems face challenges in cost-effectively recognizing and tracking objects, especially in crowded environments, due to the high cost and accuracy issues of deep learning-based methods and LIDAR sensors.
Innovation Solution
A method and device using a 2D LiDAR-based marker with high-brightness reflective materials, where point cloud data is acquired, processed, and clustered to estimate the pose of a target object, allowing for robust and accurate object tracking without requiring high-performance hardware.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep learning-based methods are used for object recognition and tracking, then recognition accuracy is improved, but hardware cost and system complexity increase
Solution Approach 1:
The patent introduces a marker as an intermediary element attached to the target object. This marker serves as a mediator between the LIDAR sensor and the object to be tracked, enabling accurate detection through simple geometric feature recognition rather than complex deep learning algorithms. The marker acts as a bridge that simplifies the recognition process while maintaining high accuracy.
Solution Approach 2:
The patent replaces expensive deep learning hardware with a simple, low-cost LIDAR sensor and a inexpensive marker. The marker is a simple geometric pattern that can be easily manufactured and attached to objects, serving as a disposable or temporary solution that eliminates the need for high-performance computing hardware.
2Measurement precision
If LIDAR sensors are used for distance measurement, then measurement precision is improved, but system cost increases
Solution Approach 1:
The patent uses a 2D LIDAR sensor to capture point cloud data of the marker, creating a digital copy or representation of the marker's geometric features. This 2D point cloud copy is then processed to extract pose information, eliminating the need for expensive 3D LIDAR sensors or cameras while maintaining measurement accuracy.
Solution Approach 2:
The patent replaces complex mechanical or optical 3D sensing systems with a simpler 2D LIDAR-based system. By substituting the sensing mechanism and using computational geometry to reconstruct 3D pose information from 2D point clouds, the system achieves cost reduction while maintaining measurement precision.
3Device complexity
If image sensors are used for object detection, then system cost is reduced, but measurement precision and reliability decrease due to lighting changes and occlusion
Solution Approach 1:
Instead of using passive image sensors that rely on ambient light, the patent actively illuminates the marker using LIDAR laser pulses. This inversion from passive to active sensing eliminates dependency on lighting conditions, ensuring consistent detection accuracy regardless of environmental light levels or occlusions.
Solution Approach 2:
The patent substitutes image sensors with LIDAR-based point cloud sensing. This replacement transitions from optical imaging susceptible to lighting and occlusion to active laser ranging that provides reliable distance and geometric information independent of ambient light conditions.
4Reliability
If high-performance hardware is used for deep learning-based tracking, then object tracking reliability is improved, but system cost increases
Solution Approach 1:
The patent extracts and isolates the essential geometric features of the marker from complex environmental data. By focusing only on the marker's specific geometric pattern and its point cloud representation, the system eliminates the need for high-performance hardware designed for general object recognition, achieving reliable tracking through simplified feature extraction.
Solution Approach 2:
The patent changes the detection parameters from complex visual features requiring deep learning to simple geometric parameters (point cloud coordinates, distances, angles) that can be processed by basic algorithms. This parameter transformation from visual to geometric domain enables reliable tracking with low-performance hardware.
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
This approach enables a cost-effective and accurate object recognition and tracking system for mobile robots, improving computational efficiency and reliability in various environments.
Implementation Method 1
A LIDAR (Light Detection And Ranging) sensor is a sensor that measures a distance to a reflection body by shooting a laser pulse and measuring a return time
Implementation Method 2
measures a distance to a reflection body by shooting a laser pulse and measuring a return time
Implementation Method 3
extracting second point cloud data corresponding to a high-brightness reflective material included in a marker attached to an object
Data Source
AI summary
Provided is a method for estimating a pose of a target object using a 2D LiDAR-based marker in a mobile robot, the method including acquiring first point cloud data including distance and reflection intensity information using a 2D LiDAR sensor of a mobile robot, extracting second point cloud data corresponding to a high-brightness reflective material included in a marker attached to an object from the first point cloud data, clustering the second point cloud data into at least one cluster, identifying a target cluster corresponding to the target object among the at least one cluster, and estimating a pose of the target object based on change in the target cluster.