Multi-Sensor SLAM Framework for Humanoid Robots: Enhancing Environmental Perception and Localization in Webots using NAO with Virtual Sonar, IMU, and Camera Data

SLAM functions as a vital capability for autonomous robots which allows them to perform navigation and build maps of unexplored environments. This paper develops a multi-sensor SLAM framework which operates on the NAO humanoid robot within Webots simulation platform. The framework uses information from virtual sonar sensors alongside IMU (Inertial Measurement Unit) devices as well as cameras for improved perception of surroundings and better positioning accuracy. A robust approach based on Extended Kalman Filter (EKF) for state estimation control enables hazard detection using visual features which enhances mapping reliability. The proposed framework established in dynamic conditions shows positive experimental outcomes that enhance localization precision and mapping reliability. Empirical evidence demonstrates how the system operates effectively through simulations which provides proof for implementing this technology within humanoid robotic systems.