Analytical Formalism for Data Representation and Object Detection with 2D LiDAR: Application in Mobile Robotics

Area of Science:

• Robotics
• Computer Vision
• Sensor Fusion

Background:

• Laser scanners (LiDAR) are crucial for mobile robot navigation in diverse environments due to their accuracy.
• Current research lacks a standardized data representation and modeling strategy for LiDAR sensor data.
• This heterogeneity hinders the development and integration of new applications.

Purpose of the Study:

• To develop a formal analytical approach for object identification and localization using 2D LiDAR data.
• To establish a common formalism for representing and processing LiDAR measurements.
• To facilitate the design and implementation of advanced robotics applications.

Main Methods:

• Formal definition of Laser Imaging, Detection, And Ranging (LIDAR) sensor measurements and their representation.
• Development of an analytical framework for object identification, property extraction, and localization.
• Experimental validation in semi-structured environments relevant to autonomous navigation.

Main Results:

• Successful demonstration of multiple object detection and identification based on the proposed analytical representation.
• Validation of the approach's feasibility in generic, semi-structured environments.
• Quantification of accuracy and precision in object localization.

Conclusions:

• The proposed analytical approach provides a unified formalism for 2D LiDAR data processing in mobile robotics.
• This standardization simplifies object identification, localization, and property analysis.
• The framework supports the development of diverse LiDAR-based robotic applications.