Scientific direction Development of key enabling technologies
Transfer of knowledge to industry

PhD : selection by topics

See all positions [+]

Improvement of CdZnTe based gamma imager CdZnTe using machine learning

Département d'Optronique (LETI)

Laboratoire Architecture Systèmes Photoniques



Photonics, Imaging and displays (.pdf)

Gamma imaging is a technique widely used in medical imaging (molecular imaging, nuclear medicine) and security (transportation, industry). CdZnTe semiconducting detectors usage is currently emerging for SPECT (Single Photon Emission Computed Tomography, using gamma-cameras) and portable gamma imaging. Indeed, they enable performance improvements in speed, sensitivity and image quality. These detectors operate at room temperature and are sensitive to five physical parameters of the interaction: deposited energy E, interaction time T and the 3-dimensional position XYZ. These parameters are estimated by real-time analysis of anode electronics signals. However, the link between electrical signals and physical parameters is not fully known, as material physical properties are not uniform inside detector. The goal of this Ph.D. internship is to overcome these limits by using machine learning techniques to model actual detector response. Recent multi-layered deep learning technique enable to build and train complex and flexible system models, and to overcome our lack of knowledge about detector physics. The identification of internal physical parameters of the detector would allow to optimize estimation of interaction location, time and energy. This will lead to a better image quality and then capability to detect small and weak objects, enabling better diagnoses and lower false alarm rate. The student may have a background in applied mathematics (machine learning) and/or instrumentation physics. He/she need to have taste for multi-disciplinary research, mixing experimental physics and data science.

Download the offer (.zip)

Adaptative frequency tuning electronic systems for broadband vibration energy harvesting

Département Systèmes (LETI)

Laboratoire Autonomie et Intégration des Capteurs



Cyber physical systems - sensors and actuators (.pdf)

Energy harvesting is a theme whose aim is to supply power to wireless sensor nodes by replacing the source of electrical energy (battery, cables) with ambient energy. Vibration energy harvesting, in particular, makes it possible to exploit the mechanical energy of an environment and convert it into electricity in order to supply a wireless sensor node. The thesis will focus on the exploitation of piezoelectric materials on resonant structures to convert vibration energy into electricity. The use of mechanical resonators amplifies ambient vibrations, but the harvested power drops sharply when the spectrum of ambient vibrations no longer coincides with the harvester's resonant frequency. For the adoption of this type of system by industry, one of the major obstacles is therefore this frequency selectivity. The CEA and the University of Savoie Mont-Blanc (SYMME Laboratory) have recently proposed high-performance techniques to solve this problem by using harvesters that can be dynamically tuned by an electronic system. Indeed, coupled with intelligent electronics, a "strongly coupled" harvester has its mechanical behavior modified (its resonance frequency in particular), making it possible to follow the evolution of the input frequency. The objective of the thesis is to propose, dimension, simulate, fabricate and test innovative electronic architectures (based on discrete components and/or microcontrollers) allowing the automatic tuning and the search for the maximum power point of piezoelectric vibration energy harvesters. Particular attention will be paid to the low power and small size of the electronic architectures since the ultimate goal is to propose an autonomous circuit consuming a negligible part of the harvested electrical energy. At the end of the thesis, the selected architecture(s) will then be proposed to the CEA-Leti's integrated circuit department for miniaturization. A complete demonstrator (harvester, micro-converter and adjustment circuit) is targeted for the end of the thesis.

Download the offer (.zip)

Ecodesign methodology for new generations of batteries

Département des Technologies des NanoMatériaux (LITEN)

Laboratoire des Eco-procédés et EnVironnement



Electrochemical energy storage incl. batteries for energy transition (.pdf)

The development of the electrification of vehicles requires the design of cheaper and more efficient battery technologies. In response to this demand, many development paths are under study, such as new generations of Li-ion with reduced cobalt content or high energy density, all solid state lithium batteries or Li-Sulphur batteries, among other. Apart from the performance aspect, there is a real need to assess the environmental impact of these technologies over their entire life cycle (LCA), and to look at eco-design options for the development of the batteries of the future. The proposed thesis will aim at addressing these issues, using a multidisciplinary approach combining the skills of at least three laboratories from CEA LITEN. At the end of the thesis, the expected results will be: an environmental evaluation of the 3 new generation of battery technologies (advanced Li-Ion, Li-S and All-Solid), compared to reference battery technologies as well as an eco-design methodology to guide decision support in the development of low TRL battery technologies.

Download the offer (.zip)

Applied formal semantics in hardware compiler frameworks

Département Architectures Conception et Logiciels Embarqués (LIST-LETI)

Laboratoire composants logiciels pour la Sûreté et la Sécurité des Systèmes



Artificial intelligence & Data intelligence (.pdf)

The development of RISC-V instruction set architecture (ISA) is supported by new methodologies and tools which are dedicated to increase the productivity of hardware designs (i.e., high-level design languages and specialized compilation chains). At the language level, Chisel and FIRRTL Hardware Description Languages (HDLs) aim to raise the level of abstraction of hardware design. It thus becomes appealing to formally reason on functional and temporal properties of these high-level designs and rely on appropriate compilation extensions to transfer these high-level properties down to the level of generated Verilog, for example. In this PhD proposal, we target a formal verification framework for computer architectures to support the specification and verification of timing-related safety and security properties. The following two contributions are expected of this PhD thesis: 1) the design and implementation of a verification infrastructure based on formal executable semantics of Chisel and FIRRTL HDLs and 2) the design and implementation of an assertion language to express timing safety and security properties, which are to be verified with the aforementioned formal infrastructure. The scientific contributions of this thesis are expected to evaluated on a selection of the rich-set of architecture designs provided by the RISC-V ecosystem.

Download the offer (.zip)

Innovative chalcogenide materials for photonics applications: impact of integration processes and interfaces on their optical properties

Département des Plateformes Technologiques (LETI)




Photonics, Imaging and displays (.pdf)

Chalcogenide materials are now considered as most promising materials for many emerging applications in microelectronics or for optical sensors: photonics in the MIR, nonlinear photonics and applications, photonic neuromorphics, MIR sensors but also for OTS selectors for the new 3D resistive memories. The aim of this thesis is to study and control the impact of integration processes, encapsulation materials & interfaces on the optical properties of thin films of such chalcogenide materials to enable the future realization of highly efficient photonic devices. In this context, the phD student will realize photonic objects and structures based on chalcogenide materials using the classic microelectronics integration tools available on the 200/300 mm LETI platform such as sputtering deposition, optical and electronic lithography , plasma etching ... The obtained photonic structures will first be characterized using thin-film characterization tools (AFM, XPS, FTIR, Raman, XRD, XRR, ellipsometry / reflectivity as a function of temperature ...) available at the nano-characterization platform of CEA Grenoble (PFNC). Optical properties (propagation losses, quality factor Q of cavities, optical nonlinearities, optical phase shift, ...) of photonic objects (waveguides, interferometers, phase shifters, resonant rings, nonlinear structures ...) will be characterized on the photonics measurement benches of LETI as well as at the University of Bourgogne in Dijon. This work should allow the outcome of the thesis to develop efficient photonic devices and to go beyond the state of the art by making the best use of the unique optical properties of these new materials. This will go through a deep mastery of the development of these materials and their integration at the nanometer nanoscale by means of lithography/etching technique with a particular emphasis on the control of their interfaces (impact of etching and encapsulation materials/processes, passivation of electronic surface states , study of the potential of heterostructures ...).

Download the offer (.zip)

Side-Channel Analysis against the confidentiality of embedded neural networks: attack, protection, evaluation

Département Systèmes (LETI)

Laboratoire Sécurité des Objets et des Systèmes Physiques



Cyber security : hardware and sofware (.pdf)

One of the major trends of Artificial Intelligence is the large-scale deployment of Machine Learning systems to a large variety of embedded platforms. A lot of semi-conductor practioners propose "A.I. suitable" products, majoritarely with neural networks for inference purpose. The security of the embedded models is a major issue for the deployment of these systems. Several works raised threats such as the adversarial examples or the membership inference attacks with disastrous impact. These works consider the ML aglorithms through a pure algorithmic point of view without aking into consideration the specificities of their physical implementation. Moreover, advanced works are compulsory for physical attacks (i.e., side-channel and fault injection analysis). By considering a overall attack surface gathering the theoretical (i.e. algorithmic) and physical facets, this subject propose to analyze side-channel analysis threats (SCA) targeting the confidentiality of the data as well as the model (reverse engineering) of embedded machine learning systems and the development of appropriate protections. Several works have studied physical attacks for embedded neural networks but with usually naive model architecture on 'simple' 8-bit microcontrolers, or FPGA or at a pure simulation level. These works do not try to link the fault models or the leakages with well-known algorithmic threats. Being based on the experience on other critical systems (e.g., cryptographic primitive), the main idea of this PhD subject will be to jointly analysis the algorithmic and physical world in order to better understand the complexity of the threats and develop efficient defense schemes. The works will answer the following scientific challenges: (1) Caracterization and exploitation of side-channel leakages: how to exploit side-channel leakages (power or EM) to guess sensible information focused on the training data or information on the model architecture. (2) Evaluation of the relevance of classical countermeasures such as hiding or masking techniques for this kind of systems and threats. (3) Develop new protections suitable to embedded neural networks.

Download the offer (.zip)

173 Results found (Page 5 of 29)
first   previous  3 - 4 - 5 - 6 - 7  next   last

See all positions