Localization Method for a Rehabilitation Mobile Robot using Visual and Ultrasonic Information

This paper presents a new robot localization method in an indoor environment and enters in the development of an assistance robot for disabled people called the ARPH project. The robot developed is composed of two principal parts: a mobile base for displacement and an embedded MANUS arm to handle objects. Each part is equipped with two kinds of sensors: a camera to have the environment structure and the object parameters and an ultrasonic system to recognize environment and object. Our approach to robot localization is based on combination of these two sensors information and uses a 3D/2D matching between environment model and image attributes. The aim is to compute the robot position and orientation in the navigation workspace in a fast way and with a good accuracy, to give to the mobile robot ability to realise different tasks like navigation, tracking and grabbing objects.