My research activities are in two research areas:
My main contributions in sensor/multi-sensor based control mainly concern the research and proposal of new robot-environment interaction models (patterns) in order to propose dynamic sensor baased control laws for two fields of application: manipulator robots and single and multi-drone aerial systems.
In aerial robotics, I mainly worked on the generic modeling of flying machines of any type (rotary (n rotors) or fixed wing) and on the control-guiding laws associated with these models. Moreover, since the creation of the ARMEN-LS2N team, I have directed my research in this field towards two axes:
I m in charge of the animation of "Theme 2: Interaction with the environment" of ARMEN team.
I thus took in hand this topic in accordance with my research project and in particular the multi-sensor aspect which presents a particular interest because it is still little exploited and developed. Indeed, there is much work on data fusion for localization, but still too little work on control fusion to perform a robotized task, involving both different interaction modes, different sensors and a large number of redundant and associated information. Sylvain Vandernotte's thesis work, co-supervised with P. Martinet, is fully in line with this axis by developing and proposing two new innovative results: the second order kinematic model of the most used primitives and a dynamic multi-sensor referenced control. Recently, in order to exploit my expertise in CRC mixed with the very innovative notion of "hidden robot" (an ARMEN exclusivity), I co-supervised with S. Briot Minglei Zhu's thesis, whose objective is to use recent work on the control of the hidden robot for the optimal design of the parallel robot based on visual servoing approaches.
ARMEN's main objective is to "Design and control robots and their interactions with the environment." In this perspective, I co-supervision with S. Briot and P. Martinet, Damien Six's thesis (2015--...) whose objective is to design a flying parallel robot with an original architecture made up of two quadrirotors, capable of moving an effector to perform tasks with 2 degrees of freedom. ARMEN is the first research team to position itself on this promising new theme and on which it is a precursor. The extension of D. Six's work is now the subject of a new thesis whose objective is the study and realization, here too, of a novel architecture in the form of a large flying gripper consisting of four quadrirotors for the capture and manipulation of large objects (thesis by Li Zhongmou, since November 2017, co-supervised with I. Fantoni and V. Begoc). Another facet of the ARMEN team is to design systems with drastically reduced energy consumption and robust control laws. This is the subject of R. Balderas' thesis which I co-supervision with S. Briot.