Abstract:
The monitoring and control of local or remote supervised systems dedicated to carry out missions and complex tasks require different levels and modes of interaction with the human operator when facing different situations. Examples of such systems are found in many application areas including critical areas such as the art of war, the dismantling of nuclear power plants, the space exploration, the télé-surgery, etc.
The subject is related to a very complex area of futurist technology. It aims to carry out the analysis, development, implementation and performance evaluation of a multimodal interface to ensure flexible and adaptable Human-Robot interaction. This multimodal interface will be the pivot of the whole Human-Robot system and is inspired from the natural Human-Human interaction which operates and combines high-level control modes (speech, gestures, locating objects by vision, manual manipulation).
The main objective of this thesis is to conduct an analysis concerning the human robot interaction in multimodal supervised systems. It intends to study the influence of the nature and the number of available modes, the human factors, the performance of the operational system, the complexity of the task or mission. Of course, because of its own complexity caused by the mode multiplicity and by the possibilities of their combinations; the degree of structural complexity of the system is expected to increase and mobilize the multiple sensory skills of the operator. But from the adopted design that is user centered and that uses a high abstraction level; the functional and operational complexity of the overall system will be masked enabling flexibility and adaptability of
use ensuring therefore better performance.
The major contributions of this thesis are:
The development of a multimodal interface combining different high-level control modes:
voice control, gesture control, control by pointing on an image (click-and-point) and joystick control.
The design and implementation of a new robotic system which is the parallel cable-based robot that open lines of research for original applications, particularly in the areas of rehabilitation.
The development of an original application that uses a 4 cables robot and the multimodality to built up a special class of a rehabilitation and learning center.
The proposal of a method for performance evaluation of multimodal supervised systems that takes into account some human factors by fitting an evaluation method called Analytical Hierarchy Process (AHP), which allows merging qualitative and quantitative criteria.
