Indoor Navigation with Model Vehicles

Abstract

This paper describes a vision-based approach to miniature robot vehicle position estimation and navigation, that integrates data from on-board camera, tachometers and steering servo. The goal is to navigate a miniature vehicle on an indoor track using visual data from lane markings and dead reckoning from odometry. Then the estimated position of the miniature vehicle will be visualized in real-time and overlayed on a photo of the actual track. This will be used to evaluate how well this estimation performs and what its limitations are.