PhD thesis, University of Evry Val d'Essonnes

Defended dy
Philippe Hoppenot

The disabled people aid has been developed for several years. Mobile robotics can help for it. Defined with the French Association against Myopathies (AFM), the aim of the project Intelligent Mobile Robot (RMI) is to embark a manipulator arm on an autonomous mobile base to give to disabled people the possibility to realise themselves everyday life tasks.
This work is divided in two parts. The first one, the more developed one in this thesis, consists in giving a certain autonomy to the mobile base (planification and navigation capacity). Solved with expensive systems, researches must be done on this issue to achieve it operational with low cost robots which are essential for medical robotics. To accomplish those two tasks, the robot must know its position in an partially known environment. During the mission, on line localisation takes into account the odometry corrected by the ultrasonic sensors measures. The main point is to match the measures with the environment. It is based on the approximate knowledge of the robot position given by the odometer. If, in spite of this on line localisation, the robot is lost (an automatic detection is performed) an off line localisation system is used. Only based on the ultrasonic measures, it gives the robot position in its environment.
The second part studies the man machine co-operation. In the field of disabled people assistance, automatism is not looked for to do instead of the person but must be an extension of the disabled operator capacity. This thesis feels an interest in the kind of information given back to the operator to know if the mission is performing well : blocking detection and/or localisation error detection.

Key words : Mobile robotics, assistance to disabled people, localisation, man machine co-operation, trajectory planification, navigation, odometry, ultrasonic sensors.