A Distributed ESP32-ROS Architecture for Real-Time EKF-SLAM on Low-Cost Mobile Robots

Soumyajeet Mahapatra, Asim Kr Naskar

Keywords: State Estimation, EKF, SLAM, Resource-Constrained Systems, ESP32, ROS, Mobile Robotics, Control Systems

TL;DR: The paper demonstrates a low-cost distributed ESP32–ROS system that enables real-time EKF-SLAM by offloading heavy computation to a host PC while the microcontroller handles high-rate sensing.

Abstract: This paper presents a distributed computational architecture enabling real-time EKF-based state estimation on low-cost, resource-constrained mobile robots. Our framework strategically partitions tasks: an ESP32 microcontroller manages high-frequency, real-time odometry and LiDAR data acquisition, while a host computer running the Robot Operating System (ROS) executes the computationally demanding EKF-SLAM algorithm. This hybrid, low-cost system, which leverages the ESP32's built-in Wi-Fi for communication, is experimentally validated to produce accurate and stable state vector estimates $[x, y, \theta]^T$. This architecture provides a practical and scalable blueprint for implementing advanced estimation and control algorithms on inexpensive robotic hardware.

Submission Number: 11