Pitch Variation Filter for LiDAR-Only SLAM and Localization in Self-Balancing Mobile Robot
Doyeon Kim, Heoncheol Lee, Ka Hyung Choi
AI summary
Problem
Sudden pitch changes in two-wheeled self-balancing robots cause 2D LiDARs to capture ground or ceiling data, disrupting scan matching and destabilizing SLAM localization.
Approach
The Pitch Variation Filter identifies and removes pitch-affected point clusters using geometric variance, angular confidence, and temporal tracking, operating directly in the sensor frame without IMU data or scan alignment.
Key results
- Reduces SLAM errors by at least 24.31% across scan matching algorithms
- Filters pitch-affected clusters using geometric and temporal cues without alignment
- Operates independently of SLAM for seamless system integration
- Demonstrates robust performance on normal and disturbed movement sequences
Why it matters
Enables reliable, low-cost 2D LiDAR navigation for dynamic self-balancing robots, removing the need for expensive 3D sensors or IMUs in commercial applications.
Abstract
This paper addresses the Pitch Variation Problem in Two-Wheeled Self-Balancing (TWSB) robots that use 2D LiDAR for Simultaneous Localization And Mapping (SLAM). The issue arises from sudden accelerations or decelerations, leading to abrupt pitch variations that cause the 2D LiDAR to capture data from unintended surfaces, such as the ground or ceiling, destabilizing the robot’s position estimation. To mitigate this, we propose a novel preprocessing method that efficiently removes point clusters affected by pitch variation by leveraging their dis- tinct characteristics, without the need for an alignment process. Experimental results demonstrate that our method reduces errors by at least 24.31% across various scan matching algorithms. Furthermore, as the proposed method operates independently of SLAM, it can be seamlessly integrated into a wide range of systems and has been shown to substantially enhance SLAM performance when used alongside existing algorithms.