1 Introduction People with mobility impairments that use wheelchairs for travelling usually drive these vehicles either by hand or by using electric devices such as a joystick attached to one of the wheelchair’s arms. However, many of these people find great difficulties in controlling repetitive and/or accurate navigation operations due to poor motor control or sensory limitations. A field of research has evolved during the last years on providing mobility impaired people with assistant vehicles that complete their skills. In this sense, powered wheelchairs allow inserting automatic actions in the normal operation of the vehicle for assisting handicapped people and improving their mobility. Most problems found in this research have strong resemblance to those of mobile robotics. However, there is an important difference ([1]): in a wheelchair there is always a human that drives the overall operation of the vehicle (maybe only specifying general goals or maybe physically driving the wheelchair), which leads to an essential requirement: the wheelchair automatic system must include a suitable human-vehicle interface to be easily commanded. Such an interfacing comprehends two levels: human physical actuation on the wheelchair (voice recognition, eye- or face-movement commanding, etc.) and human accesses to the wheelchair’s internal status (goal communication, task planning interaction, etc.). There has been a lot of effort directed to solve the former level of interfacing ([2,3,4,5,6]), but fewer to the latter. The work presented in this paper aims to that less explored direction. Assistive Navigation using a Hierarchical Model of the Environment 1 Juan A. Fernández, Javier González, Cipriano Galindo, Antonio Reina, Antonio Muñoz System Engineering and Automation Department, University of Málaga Campus Teatinos - Complejo Tecnológico 29071 Málaga (Spain) e-mail: {jafma,jgonzalez}@ctima.uma.es Abstract: This paper presents an application of a hierarchical, symbolic model of large-scale space for representing the environment of a robotic wheelchair that has been automatized for assistance to mobility impaired people. Rehabilitation robotics has an issue that is not present in mobile robotics: the need for an interface with the human driver. Our hierarchical model demonstrates its suitability to represent the environment in a structure easily manageable by the human (it serves as a good interface with the human’s cognitive map) and at the same time its better performance than flat topological maps in route planning. The paper describes and illustrates an experiment in which the wheelchair, with the assistance of the person, builds its model of space and uses it for autonomous navigation. Keywords: Rehabilitation Robotics, World Modeling, Robot Path Planning and Navigation. Final Version appears in Proc. of Int. NAISO Symp. on Eng. of Intelligent Systems 2002