Work

Beyond Odometry Accuracy: How Encoder Resolution and IMU Fusion Affect Particle Filter Localization in Differential-Drive Robots

Year: 2026

Type: article

Abstract: Encoder resolution and sensor fusion architecture directly influence odometric drift in autonomous mobile robots (AMRs), yet their interaction with particle filter localization remains insufficiently characterized. This study evaluates four encoder resolutions (90, 512, 1024, and 4096 PPR) in wheel-odometry-only (System A) and EKF-based IMU-fused (System B) configurations across multiple navigation scenarios using ROS AMCL localization. Increasing resolution significantly reduced drift, with 512 PPR achieving most of the improvement observed at 4096 PPR and diminishing returns beyond 1024 PPR. IMU fusion consistently improved odometry accuracy and reduced variance, although its benefit depended on trajectory complexity rather than encoder resolution alone. However, improved odometry did not uniformly translate into improved localization. In several cases, the IMU-fused system exhibited reduced AMCL stability despite lower odometric drift. These results demonstrate that encoder selection and sensor fusion must be considered alongside particle filter parameterization, highlighting the importance of estimator compatibility in ROS-based AMR localization architectures.

Source: Robotics

DOI: 10.3390/robotics15090164

Authors: Lim Yun Ler, Thangavel Bhuvaneswari, Min Thu Soe, Muhammad bin Hishamuddin

Institutions: Multimedia University, Telekom Malaysia Berhad (Malaysia)

Language: English

Topic: Robotics and Sensor-Based Localization

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: gold

Estimated APC paid: $2,172

Funder: Telekom Malaysia Berhad