This paper deals with solving robot cooperative task conflict using robot assistive behavior processing. These conflicts could occur in reactive multi mobile robot task. To manage such conflicts, each robot must assist the other units from their own capacities. A role method based on capacity adjustment is proposed. A role is a specific robot assistive behavior. Conflicts are then compensated by role switching. The same processing can be used in man-robot interaction. In this case, the robot has to define the efficient role in order to assist the man. A specific multi mobile robot conflict problem is studied: a constraint case of a reactive handling task with two mobile robots. This is a critical task since the object transported is only held by robot compression. Any degraded case, hardware failure or tactical conflict occurrences, would lead to a task deadlock. To define the global assistive role, on-line fuzzy reinforcement is used to adjust basic behaviors. In the studied case, no explicit communication is used to define the relationship between the robots in order to be generalized to man-machine cooperation. Our goal is to develop a strategy to manage interaction with handicapped persons. We studied a robotic rollator. In this case, the assistive role of rollator has to be adjusted in order to compensate the man's task

Evaluation systems for handicapped people through adapted interface are used to measure of people motor disability parameters and to provide assistance. In this work, a friendly reconfigurable evaluation and assistance platform for handicapped people is presented. A set of interfaces (mouse, joysticks, etc.) is used for controlling the pointing system. Different systems can be controlled (PCs cursor, robot Khepera, electric wheelchair). The main characteristics of this platform are the availability of several interface devices, exercises interaction with 2D models or 3D models, the saving of experimental data, the easy configuration of different scenarios, and the remote access capabilities. The system is illustrated through an implementation example and preliminary results are shown on a child with cerebral palsy

Due to specific pathologies, a part of the handicapped population does not have a correct adaptation of his wheelchair command and sometimes this adaptation is never found. This article presents a study about an automatic autoadaptation of the technical aid behavior with regard to the user characteristics. Different experiments are realized with two handicapped person on a robotic wheelchair. The obtained results present the possibilities of this adaptation on an electric wheelchair in order to take into account heavy difficulties of handicapped person commands (like cerebral motor infirmity, Sclerosis)

Due to technical progress and to the growth of the dependent population in our society (elderly and/or handicapped people), needs in advanced technical aids represent a great challenge for robotics. This article presents a study about technical systems for locomotion aid integrating robotic process and adaptive neural controller. The system specificities are an automatic auto-adaptation of the technical aid behavior with regard to the user characteristics. Different experiments have been realized with a robotic walker to evaluate our system. We also integrated it on a robotic wheelchair to show his portability on other technical aids. The obtained results present the possibilities of this adaptation on a walker system and on an electric wheelchair in order to take into account heavy difficulties of handicapped person commands (like Cerebral Motor Infirmity).

This article presents a study about a mobility aid system, which integrates adaptive control interface for a robotic walker. The objective is to give to the patient more flexibility in the choice of a technical aid. The specificity of our system is based on an auto-adaptive interface, which improves the configuration choice of the patient's driving commands. The results obtained from experimentations show the capabilities of our method for a technical aid. The learning process is applied on line to a neural network controller. The experiment is focused on sensors configuration and command of a powered walker interface.