Nicolas Navet home page

Research in real-time dependable systems and risk assessment

My goal is to advance techniques, tools, and computing platforms that enable the development of provably safe systems in a time- and cost-efficient way, with applications in:

• Automotive and aerospace embedded systems: automated synthesis of E/E architecture (design-space exploration, machine learning) & timing and dependability verification  
• Risk assessment: quantify risks using probabilistic models derived from historical data 

Here is a list of my publications (available for download) and slides from recent talks. You can find here a short biography and a list of (academic) professional activities.

Selected works

• E/E architecture design automation:
  • A Model-Based Systems Engineering framework with DSE for E/E architectures in automated driving vehicles, with Robert Bosch.
  • Graph neural networks to speed-up the verification of Ethernet TSN networks in design-space exploration - study available here as well as a follow-up work.  
  • “QoS-Predictable SOA on TSN: Insights from a Case-Study”, with Renault. 
  • “Towards Computer-Aided, Iterative TSN-and Ethernet based EE Architecture Design”, with BMW.
    
  • “Early-stage Bottleneck Identification and Removal in TSN Networks“, with Volvo.
    
  •  “Practical Use-cases for Ethernet Redundancy”, with NXP.
    
  • "Multi-source software on multicore automotive ECUs - Combining runnable sequencing with task scheduling", with PSA.
• Communication networks for dependable systems:
  • Timing QoS protocols on top of Ethernet TSN (slides)
  • Scheduling frames with offsets provides a major performance boost on CAN (slides)
    
  • Configuration of FlexRay networks (slides)
  • Fine Tuning MAC Level Protocols for Optimized Real-Time QoS
  • Optimal configuration of TDMA / TTP/C networks - (slides)
    
  • Probabilistic analysis of CAN fault-confinement mechanisms
    
  • Patents on communication networks:  FR2976432 - FR2976434 - FR2976435.
    
• Model-Driven Engineering for embedded systems:
  • "Automated Fault Tolerance Augmentation in Model-Driven Engineering for CPS"
  • “Lean Model-Driven Development through Model-Interpretation: the CPAL design flow”.
• Low-power scheduling:
  • Optimal CPU frequencies for energy consumption, the case of FIFO tasks and the general case.
  • Stochastic model of battery discharge time
  • Multi-processor low-power scheduling
• Financial engineering
  • Regime-switching models for financial data
  • Heavy-tails behavior of financial data
    
  • Large deviations for risk assessment
    
  • Correlation between time series entropy and technical analysis returns