Autonomous Navigation of Humanoid Robots Using Kinect Sensor: A Vision-Based Approach for Real-Time Obstacle Detection and Path Planning

摘要

This research presents a novel perspective for an autonomously navigated humanoid robot system which unifies a Kinect Xbox 360 sensor, and L298N motor drivers with Arduino Uno for harnessing real-time depth sensing and vision-based obstacle detection. The robot can conclusively navigate, avoid obstacles and change the direction of its path if an obstacle is found. The Kinect sensor provides enumerated outer environmental data, allowing for decisive path planning and navigation. Arduino Uno processes the sensor data and sends directive commands to the motors via motor controllers through sequential communication. This paper proposes, and analyzes the navigation system’s performance, and proves its advantages compared to existing models.