High precision robotic manipulation with reinforcement learning and Sim2Real
High precision robotic assembly that handles high product variability is a key part of an agile and a flexible manufacturing automation system. To date however, most of the existing systems are difficult to scale with product variability since they need precise models of the environment dynamics in order to be efficient. This information is not always easy to get.
Reinforcement learning based methods can be of interest in this situation. They do not rely on the environment dynamics and only need sample data from the system to learn a new manipulation skill. The main caveat is the efficiency of the data generation process.
In this post-doc, we propose to investigate the use of reinforcement learning based algorithms to solve high precision robotic assembly tasks. To handle the problem of sample generation we leverage the use of simulators and adopt a sim2real approach. The goal is to build a system than can solve tasks such as those proposed in the World Robot Challenge and tasks that the CEA’s industrial partners will provide.
Formalization of the area of responsibility of the actors of the electricity market
The CEA is currently developing a simulation tool which models the energy exchanges between the actors of the electricity market but which models, in addition, the exchanges of information between those actors. The first results of this work show that, for some new energy exchange schemes, ’indirect’ interactions between actors may appear and may cause financial damage (for example, the failure of a source of production of one actor may impact the income of another). Thus, the borders which clearly delimited until now the areas of responsibility of each actors could be brought to blur and their areas of responsibility could "overla". The candidate will be responsible for:
- Formally define the area of responsibility of an actor in the electricity market,
- Model the interactions, including ’indirect’ ones, that may appear between these actors,
- Apply formal proof techniques (such as ’model-checking’) to detect overlaps in areas of responsibility,
- Define the conditions of exchange between the actors which would guarantee the non-recovery of the areas of responsibility.
Non-volatile asynchronous magnetic SRAM design
In the applicative context of sensor nodes as in Internet of things (IoT) and for Cyber Physical Systems (CPS), normally-off systems are mainly in a sleeping state while waiting events such as timer alarms, sensor threshold crossing, RF or also energetic environment variations to wake up. To reduce power consumption or due to missing energy, the system may power off most of its components while sleeping. To maintain coherent information in memory, we aim at developing an embedded non-volatile memory component. Magnetic technologies are promising candidates to reach both low power consumption and high speed. Moreover, due to transient behavior, switching from sleeping to running state back and forth, asynchronous logic is a natural candidate for digital logic implementation. The position is thus targeting the design of an asynchronous magnetic SRAM in a 28nm technology process. The memory component will be developed down to layout view in order to precisely characterize power and timing performances and allow integration with an asynchronous processor. Designing such a component beyond current state of the art will allow substantial breakthrough in the field of autonomous systems.
Measurement of active cell nematics by lensless microscopy
At CEA-Leti we have validated a video-lens-free microscopy platform by performing thousands of hours of real-time imaging observing varied cell types and culture conditions (e.g.: primary cells, human stem cells, fibroblasts, endothelial cells, epithelial cells, 2D/3D cell culture, etc.). And we have developed different algorithms to study major cell functions, i.e. cell adhesion and spreading, cell division, cell division orientation, and cell death.
The research project of the post-doc is to extend the analysis of the datasets produced by lens-free video microscopy. The post-doc will assist our partner in conducting the experimentations and will develop the necessary algorithms to reconstruct the images of the cell culture in different conditions. In particular, we will challenge the holographic reconstruction algorithms with the possibility to quantify the optical path difference (i.e. the refractive index multiplied by the thickness). Existing algorithms allow to quantify isolated cells. They will be further developed and assessed to quantify the formation of cell stacking in all three dimensions. These algorithms will have no Z-sectioning ability as e.g. confocal microscopy, only the optical path thickness will be measured.
We are looking people who have completed a PhD in image processing and/or deep learning with skills in the field of microscopy applied to biology.
Digital circuit design for In-Memory Computing in advanced Resistive-RAM NVM technology
For integrated circuits to be able to leverage the future “data deluge” coming from the cloud and cyber-physical systems, the historical scaling of Complementary-Metal-Oxide-Semiconductor (CMOS) devices is no longer the corner stone. At system-level, computing performance is now strongly power-limited and the main part of this power budget is consumed by data transfers between logic and memory circuit blocks in widespread Von-Neumann design architectures. An emerging computing paradigm solution overcoming this “memory wall” consists in processing the information in-situ, owing to In-Memory-Computing (IMC).
CEA-Leti launched a project on this topic, leveraging three key enabling technologies, under development at CEA-Leti: non-volatile resistive memory (RRAM), new energy-efficient nanowire transistors and 3D-monolithic integration [ArXiv 2012.00061]. A 3D In-Memory-Computing accelerator circuit will be designed, manufactured and measured, targeting a 20x reduction in (Energy x Delay) Product vs. Von-Neumann systems.
Application of a MDE approach to AI-based planning for robotic and autonomous systems
The complexity of robotics and autonomous systems (RAS) can only be managed with well-designed software architectures and integrated tool chains that support the entire development process. Model-driven engineering (MDE) is an approach that allows RAS developers to shift their focus from implementation to the domain knowledge space and to promote efficiency, flexibility and separation of concerns for different development stakeholders. One key goal of MDE approaches is to be integrated with available development infrastructures from the RAS community, such as ROS middleware, ROSPlan for task planning, BehaviorTree.CPP for execution and monitoring of robotics tasks and Gazebo for simulation.
The goal of this post-doc is to investigate and develop modular, compositional and predictable software architectures and interoperable design tools based on models, rather than code-centric approaches. The work must be performed in the context of European projects such as RobMoSys (www.robmosys.eu), and other initiatives on AI-based task planning and task execution for robotics and autonomous systems. The main industrial goal is to simplify the effort of RAS engineers and thus allowing the development of more advanced, more complex autonomous systems at an affordable cost. In order to do so, the postdoctoral fellow will contribute to set-up and consolidate a vibrant ecosystem, tool-chain and community that will provide and integrate model-based design, planning and simulation, safety assessment and formal validation and verification capabilities.
Simulation and electrical characterization of an innovative logic/memory CUBE for In-Memory-Computing
For integrated circuits to be able to leverage the future “data deluge” coming from the cloud and cyber-physical systems, the historical scaling of Complementary-Metal-Oxide-Semiconductor (CMOS) devices is no longer the corner stone. At system-level, computing performance is now strongly power-limited and the main part of this power budget is consumed by data transfers between logic and memory circuit blocks in widespread Von-Neumann design architectures. An emerging computing paradigm solution overcoming this “memory wall” consists in processing the information in-situ, owing to In-Memory-Computing (IMC).
However, today’s existing memory technologies are ineffective to In-Memory compute billions of data items. Things will change with the emergence of three key enabling technologies, under development at CEA-LETI: non-volatile resistive memory, new energy-efficient nanowire transistors and 3D-monolithic integration. At LETI, we will leverage the aforementioned emerging technologies towards a functionality-enhanced system with a tight entangling of logic and memory.
The post-doc will perform electrical characterizations of CMOS transistors and Resistive RAMs in order to calibrate models and run TCAD/spice simulations to drive the technology developments and enable the circuit designs.
Nanofabrication of spintronic spiking neurons
In the frame of the French national ANR project SpinSpike, Spintec laboratory is opening a postdoctoral researcher position. The candidate will work in collaboration with UMPhy CNRS-Thales and Thales TRT. The objective is the realization of proof-of-concept magnetic tunnel junction based artificial spiking neurons able to generate spikes and propagate them between coupled artificial neurons.
The candidate should have a strong background in nanofabrication and should be familiar with common techniques of optical and e-beam lithography as well as different etching techniques. The candidate can also be involved in the electrical characterization of the devices.
The position is expected to start on April 1, 2021 and go on for up to 2 years jointly between the RF team and MRAM teams of Spintec. The contract will be managed by CEA and funded by ANR Agency.
We offer an international and competitive environment, state-of-the-art equipment, and the possibility to perform research at the highest level. We promote teamwork in a diverse and inclusive environment and welcome all kinds of applicants. Further information about Spintec laboratory www.spintec.fr .
Simulation of PEMFC flooding phenomena
The proton exchange membrane fuel cell (PEMFC) is now considered as a relevant solution for carbon-free electrical energy production, for both transport and stationary applications. The management of the fluids inside these cells has a significant impact on their performance and their durability. Flooding phenomena due to the accumulation of liquid water are known to impact the operation of the cells, causing performance drops and also damages that can be irreversible. With the use of thinner channels in ever more compact stacks, these phenomena are becoming more and more frequent. The objective of this post-doc is to progress in the understanding of flooding in PEMFCs. The work will consist in analyzing the link between the operating conditions, the design of the channels and the materials used in the cell. It will be based on a two-phase flow modeling approach at different scales, from the local scale at the channel-rib level, up to, via an upscaling approach, the level of the complete cell. The study will also be based on numerous experimental results obtained at the CEA or in the literature.
High entropy alloys determination (predictive thermodynamics and Machine learning) and their fast elaboration by Spark Plasma Sintering
The proposed work aims to create an integrated system combining a computational thermodynamic algorithm (CALPHAD-type (calculation of phase diagrams)) with a multi-objective algorithm (genetic, Gaussian or other) together with data mining techniques in order to select and optimize compositions of High entropy alloys in a 6-element system: Fe-Ni-Co-Cr-Al-Mo.
Associated with computational methods, fast fabrication and characterization methods of samples (hardness, density, grain size) will support the selection process. Optimization and validation of the alloy’s composition will be oriented towards two industrial use cases: structural alloys (replacement of Ni-based alloys) and corrosion protection against melted salts (nuclear application)