PhD : Identification of multivariable models for human-robot co-manipulation with passivity certificates
This PhD is part of a joint collaboration between ICube (Edouard Laroche), LIAS (Guillaume Mercère) and CEA (Neil Abroug). The location is at CEA-Saclay.
Co-manipulation systems, also called cobots, are designed to interact with human operators physically in order to assist them in their work tasks. For instance, they provide force enhancement for handling or cutting tasks. Like any usual robotic systems, cobots are equipped with actuators and position sensors. Usually, they are also equipped with force sensors.
These robotic systems interact with uncertain and time-varying environments: on the one hand, with the operator who grips the handle more or less rigidly and, on the other hand, on the effector side, depending on the task it has to perform, either lifting a load of unknown mass or cutting an inhomogeneous material, to mention only simple cases. These environments are generally assumed to be passive: they only dissipate the energy provided to them. This property of passivity of the environment makes it possible to ensure the stability of the cobot in interaction with its environment, provided that it is itself passive.
Different approaches are available to identify a model of a robot. The so-called gray box methods rely on a model structure provided by the laws of physics, the parameters of which must be estimated. They result in passive models that are relatively compact and parsimonious, but have the disadvantage of not being able to easily encompass dynamics that are difficult to model physically but that can be crucial for the performance of the control. Black-box techniques (in the frequency domain with continuous-time models) are more suitable for obtaining powerful models for controller design in particular in the presence of flexible modes. However, the available methods are not able to ensure that the provided model is passive with satisfying accuracy.
In this thesis, methods will be developed for the identification of multivariable models from frequency data, that suit to the context of co-manipulation and with the goal of enabling the synthesis of robust control laws. The aim is to nicely combine the advantages of black-box and gray-box models to lead to passive representations capable of representing complex dynamics such as high frequency flexible modes. Experimental validation will be carried out on the demonstrators available at CEA-LIST.
- Profile of applicant
This PhD proposal mainly requires skills in applied mathematics, automatic control and system identification more specifically. Fluency in French or in English is required. Thus, we are looking for candidates graduated from a Master of Science or an Engineering school with a solid scientific background with a specialization in control system theory and, of possible, in robotics. Your curiosity, your ability to analyze, your communication skills, as well as your commitment, will be a guarantee of success for the project.
- Curriculum Vitae and cover letter,
- Master marks,
- Certificate of proficiency in English,
A letter of recommendation will be appreciated. Any other document deemed necessary by the candidate which can enrich the application. Please send these documents to Neil.ABROUG@cea.fr, firstname.lastname@example.org and email@example.com before April 27, 2018. In your cover letter, you must show your understanding of the subject and highlight your skills related to the topic. All documents will be provided in the form of a single pdf document whose title will be: NAME-Firstname-thesis-identifier-cobot.pdf.
Open Internship Positions within Project CAMMA
We are looking for motivated and talented students with knowledge in computer vision, machine learning and/or augmented reality who can contribute to the development of our computer vision system for the operating room.
Please feel free to contact Nicolas Padoy if you are interested to do your master's thesis or an internship with us (funding of ~500Euros/month will be provided during 3 to 6 months). The successful candidates will be part of a dynamic research group hosted within the IRCAD institute at the University Hospital of Strasbourg. They will thereby have direct contact with clinicians, industrial partners and also have access to an exceptional research environment. The CAMMA project is supported by the laboratory of excellence CAMI, the IdEx Unistra and the MixSurg Institute.
- Deep Learning for the Analysis of Large Surgical Video Databases
- Multi-view Human Body Tracking for the Operating Room
- 3D Simulation and Visualization of X-ray Radiations for Radiation Safety Analysis
- IHU Strasbourg / MIX-Surg Institute: http://www.ihu-strasbourg.eu/ihu/en/institut/presentation/
- IRCAD: http://www.ircad.fr/?lng=en
- AVR group: http://icube-avr.unistra.fr/en/index.php/Main_Page
- LabEx CAMI: http://cami-labex.fr/