Development of an Intelligent Senior-Car in a Pedestrian Walkway

摘要

Abstract This research aims at development of a stable mobile robot (personal mobility robot) that can navigate automatically on sidewalks where ordinary citizens are coming and going. Many studies have investigated automated vehicles that move on public roads. In the case of a road environment, many infrastructures and traffic regulations exist. In contrast, in a sidewalk environment, human and robots cross the same area without clear rules. For these reasons, some paths or landmarks are often blocked by unknown obstacles such as bicycles and passing pedestrians. Therefore, the most important requirements for a mobile robot are robust localization and safe navigation without collisions. We proposed simple and novel algorithms for localization using a grid map and navigation for four-wheeled robots. We also developed an intelligent robotic cart (senior car) that moves autonomously along a given course to help people transfer luggage or themselves. The sensors used in this system are three small laser scanners and an optical fiber gyroscope. The robot estimates its global position by using the integrated sensor information and generates local paths for avoiding nearby obstacles. This article describes the structure of the developed senior car and details of the proposed algorithms. We also describe a field experiment of our system in the Tsukuba Challenge 2010. Our robot successfully and autonomously ran the 1.1 km course in a realistic environment. The experiment showed the robot could estimate its position and follow the specified path effectively. Keywords: Tsukuba ChallengeMobile robots for public spaceLocalization unaffected by unknown obstaclesNavigation for four-wheeled robots