Skills and experiences

Intership on spatial and behavioral contexts in bio-inspired robotics

Neurocybernetics team, ETIS
Under the direction of N. Cuperlier and M. Belkaid

The autonomous navigation of a mobile robot in an unknown environment is a complex task that raises many problems related to perception, categorization, planning and motor control. Solving all these problems in an integrated way remains a challenge for roboticians. The approach developed in the laboratory since its creation is original in that it relies largely on the most recent discoveries in neurology. We will demonstrate in this thesis that models from hypocampal and parahympocampal system studies allow the robot to navigate and situate itself in unknown environments. For this purpose, we will rely on an internal laboratory tool: Promethe. The latter is a development environment based on middleware that manages both low-level robotic aspects such as speed and higher-level aspects such as neural network learning.

[...]

Intership on detection and segmentation by CNN
Convolutional Neural Network
Click for open the Wikipedia article.

ComSee team, Institut Pascal
Under the direction of T. Chateau

Applied deep learning to elements of the highway landscape (traffic signs, pedestrians and vehicles) is a crucial need for autonomous vehicle. During this intership, I use the Caffe library (and its Python overlay) to perform image learning and recognition. The early stages of the intership consisted in a discovery of computer vision theory, neural networks and the library. I then conducted several experiments in order to better understand how a convolutional network works and ultimately develop scripts to detect elements in an image. The application phase then began with the learning of several models, more and more precise (by discovering the effects of learning parameters on precision). Thus, a reliable panel model was created (95% accuracy on classification) as well as a car model (94% accuracy). The classification on vehicles failed completely but allowed me to understand the limits of the neural network.

[...]

Intership on aromatic hydrocarbons detectors by
ttb-MPc
Tetra-tert-butyl metallophthalocyanines
Click for open the article "Conducting Stacked Metallophthalocyanines and Related Compounds" - M. Hanack & M. Lang, 1994.
based QCM sensor
Quartz Crystal Microbalance
Click for open the Wikipedia article.

Minamat team, Institut Pascal
Under the direction of J. Brunet

The design and characterization of chemical microsensor dedicated to the detection of gaseous species is an important part of air quality evaluation. Indeed, closed environments accumulate quantities of puluants and detect them, even more when there are higly toxic at very low dose, is a major public health issue.

2014

Programming


C/C++

90% Complete

Python (with Matplotlib)

85% Complete

Visual Basic

70% Complete

Lab tools



Mecatronics


education

Machine Learning certification

Stanford University

  • Supervised learning (parametric/non-parametric algorithms, support vector machines, kernels, neural networks).
  • Unsupervised learning (clustering, dimensionality reduction, recommender systems, deep learning).
  • Best practices in machine learning (bias/variance theory; innovation process in machine learning and AI).
View full syllabus


Neuronal Dynamics certification

EPFL - École polytechnique fédérale de Lausanne

  • Hodgkin-Huxley models and biophysical modeling
  • Two-dimensional models and phase plane analysis
  • Variability of spike trains and the neural code
  • Noise models, noisy neurons and coding
  • Estimating neuron models for coding and decoding
View full syllabus


Master
Two-year postgraduate degree
Other designations: MS/MSc/M2
in Robotics Research

University of Clermont Auvergne

  • Mastering the scientific fondations of disciplines that make up robotics (mecanics, automatics, instrumentation, data processing)
  • Modelization and data optimization methods
  • Mecanisms, machines and robots' modelization principales
  • Image processing technics
  • Parameters estimation and multi-sensorial data fusion
  • Robotics systems command
View full syllabus


Bachelor
Three-year undergraduate degree
Other designations: BS/BSc/L3
in Mecatronics

University of Clermont Auvergne

  • Measurement technics in electronics, electrotechnics and automatics fiels: synthesis and analysis of electrical schematics, modeling of automatic systems and power management of a machine
  • Instrumentation technics: choice and use of measurement sensors; signal processing
  • Current technics in mecanics engineering: CAO tools, parts dimensioning, functional analysis, conception and sizing mechanisms
  • Implement essential cross-technology solutions in a context of design, integrated manufacturing and control of automated and robotic systems
View full syllabus

2014

Projects

Neural network viewer

Spatial and behavioral contexts in bio-inspired mobile robotics framework

Autonomous navigation of a mobile robot in an unkown environment is a hard task that raises many problems linked to perception, categorization, planning and engine control. To solve all that problems in an integrated way stay of challenge for roboticists. The developed approach of the laboratory, since his creation, was original because is supported, for a major parts, on most recents neurobiology discoveries. In this memory, we will demonstrate that models derivated from hypocampic and parahypocampic allow robot to navigates and to located himself in unknow environment. For this purpose, we will rely on an internal tool of the lab called Promethe. The latter is a development environment axed on a middleware managing as much low levels robotics aspects like speed as higher level aspects like learning via neural networks.
Download internship thesis (french)

Detection and segmentation by convolutional neuron network

Autonomous navigation of a mobile robot in an unkown environment is a hard task that raises many problems linked to perception, categorization, planning and engine control. To solve all that problems in an integrated way stay of challenge for roboticists. The developed approach of the laboratory, since his creation, was original because is supported, for a major parts, on most recents neurobiology discoveries. In this memory, we will demonstrate that models derivated from hypocampic and parahypocampic allow robot to navigates and to located himself in unknow environment. For this purpose, we will rely on an internal tool of the lab called Promethe. The latter is a development environment axed on a middleware managing as much low levels robotics aspects like speed as higher level aspects like learning via neural networks.

Download internship thesis (french)

Nine – a game engine

Nine is a personnal project done on many years, during holidays. It's the begining of a game engine that I've started to improve my C++ skills and understood some aspects of video games.
Used libraries :

  • SFML : Allow to manage display, sound, events and threads
  • Newton Dynamics : A physics engine
  • LUA : Nearly all the game works thanks to LUA scripts, allowing to modify a level, a tool, an object or a material, directly in a script file
It's from this basics libraries I start ths engine. Currently, it allow to :
  • display a level with lights, events, spatialized sounds and shaders
  • move a caracter at first person view with the ability to interact with the world (by raycasting)
Particular emphasis has been placed on resources management. All loaded items are follow in memory and a report of the game run is provided in an HTML page, update on the fly.
GitHub page

Mapper robot - Bachelor project

Done in 40 hours, during my bachelor’s last year, this project has for goal mapping an unknown environment, thanks to a Lego Minstorm robot. To do this, my partner and I used C++ language and bluetooth communication. The goal was to show, early in studies, the benefits of futur learned tools (signal processing, statistical maths and numerical methods, …).