Analysis of Human Robot Interaction for Supervised Multimodal Systems
Proposed to: Mr Bouchemal Billel,
Department of Electronics, University Mentouri-Constantine Algeria
Promoter: Dr Abdelouahab Zaatri ,
Department of Mechanics, University Mentouri-Constantine
Background and Objectives of the thesis
Recently, some research in the field of man-machine interaction is shifting towards the study of supervised systems which are characterized by the integration of several modes of commands (multimodal systems).
These systems are designed to perform tasks and missions in complex environments that can be dynamic and sometimes hostile.
Examples of such modes of command are the voice command, the manual command by using joysticks, the image_based command, the Gestural command, the physiological command, etc.
The multimodal interaction is an interaction offers to the supervisor several modes of command in such a way that he/she can interact with a system (robot) in an easy manner for accomplishing more or less complex tasks and missions.
Multimodal interaction enables the operator to select the mode which is the most adapted to the current situation in order to perform the tasks while taking into account the performance of the supervised robotic system.
However, the complexity of multimodal system and the complexity of the tasks generate new problems that need to be studied. Indeed, Most supervised multimodal systems remain highly dependent on the robotic system, on the supervisor, and on the task or mission for which they were originally designed. This relative lack of flexibility compared to the highly sophisticated structural capacity of the multimodal system is an essential motivation which led to propose this subject of research.
Indeed, these systems should be analyzed and tested on generic prototypes in order to identify some design rules that will improve their adaptation to deal with a wide variety of tasks and missions.
130
These problems include the following elements:
_ The nature and number of modes available, the user interface design and its ability to more or less easy manage human-robot interaction for selecting and sequencing different modes.
_ Human Factors (influence of the degree of learning and experience of the human operator, preferences and motivations of the choice of modes and their sequencing by the operator, mental state of the operator in normal or critical circumstances...).
_ Performance of the operating system to drive (the degree of intelligence, autonomy and cooperation with the supervisor).
_ The degree of complexity of the task or the mission and the environment.
The objective of the Thesis is to carry out an analysis of supervised Multimodal systems in order to highlight:
The design of a generic multimodal architecture and its illustration by the realization of an experimental prototype including some modes of command; make a critical analysis of various modes implemented individually; then, study their combination to conduct typical tasks of increasing complexity.
Finally, to draw conclusions on the best correspondences and coupling between modes of command depending on the design of robotic system, its degree of autonomy, cooperation, human factors, the complexity of the task and the environment.
140