Comparative Analysis of Sensor Fusion Techniques for Odometry Estimation in a Four-Wheeled Holonomic Drive Robot Using Wheel Encoders and IMU

Abstract

This thesis presents a comparative study of odometry and localization methods for a four-wheeled holonomic robot using wheel encoders and a BNO055 IMU. Three localization tiers were developed and evaluated: open-loop motion control, encoder-based odometry with PID speed control, and sensor fusion using a complementary filter. Experimental results showed that closed-loop encoder control significantly reduced navigation error, while IMU fusion improved heading stability during movement. The robot platform includes a 3D-printed mecanum-wheel chassis, ESP32-based control system, and a browser-based monitoring and logging framework. The study demonstrates an effective low-cost localization pipeline and provides an open-source testbed for future research on sensor fusion, UWB integration, and advanced localization algorithms for holonomic robots.