Cristian F. Rubio A.

Cristian F. Rubio A.

Electronic Engineer | Ubidots Software Engineer III B | IEEE Volunteer

LiDAR SLAM Algorithms Comparison

Published:

Overview

Open-source implementation and comprehensive comparison of three state-of-the-art LiDAR SLAM algorithms using the complete KITTI dataset. This project addresses the computational challenges of real-time 3D mapping with high data-rate laser scanners.

Project Description

Benchmarking of LiDAR Odometry and Mapping (LOAM) derivatives for autonomous navigation:

  • A-LOAM: Advanced LiDAR Odometry and Mapping
  • FLOAM Fast LOAM
  • ISCLOAM: Intensity Scan Context LOAM
  • LeGO-LOAM: Lightweight and Ground-Optimized LOAM

Evaluated across 11 KITTI sequences (00-10) covering urban, highway, and country environments with Velodyne HDL-64E sensor data.

Key Features

  • Complete ROS implementation of three SLAM algorithms
  • Automated testing pipeline for KITTI dataset sequences
  • Comparative analysis on computational cost and positioning accuracy
  • Visualization tools for trajectory and map generation
  • Ready-to-use launch scripts for reproducibility

Evaluation Metrics

  • Computational Cost: Processing time and resource usage
  • Absolute Trajectory Error (ATE): Overall positioning accuracy
  • Relative Pose Error (RPE): Frame-to-frame consistency
  • Robustness: Performance across diverse environments (urban, highway, country)

Technologies Used

Impact

Provides robotics researchers and practitioners with empirical data to select the most appropriate SLAM algorithm for their specific applications, balancing accuracy and computational efficiency.

Publication

H. F. Murcia and C. F. Rubio, “A Comparison of LiDAR Odometry and Mapping Techniques,” 2021 IEEE 5th Colombian Conference on Automatic Control (CCAC), Ibagué, Colombia, 2021, pp. 192-197, doi: 10.1109/CCAC51819.2021.9633299

Acknowledgement

This work was supported by Universidad de Ibagué under research project 19-489-INT.