Keynote 1

  Toward advanced powertrain control technologies based on models
  Prof. Yudai Yamasaki, The University of Tokyo
  Powertrain control technologies of vehicles are important for carbon neutrality. The powertrain system is getting more complicated for its hybridization, and an internal combustion engine is also a complicated component itself in nature. The control system is also becoming complicated and the conventional control framework using look-up tables based on a huge number of experiments is difficult to continue. Furthermore, the framework of control is changing, as it is now possible to control the powertrain using and combining a variety of information, including connected data and driver’s data. Model-based control system is more useful and essential to combine different things, disciplines and researchers for advanced powertrain control. In this presentation, I will introduce our research activities on advanced control systems of automobiles based on models.


Keynote 2

  Model Predictive Control: An attempt to tell an unfinished story!
  Dr. Mazen Alamir, CNRS, Grenoble
  While it is true that any control design methodology is based on some modelling of the underlying dynamics, the most advanced of these techniques, namely Model Predictive Control (MPC) cannot be conceived in the absence of such a dynamic model. The predominant place of MPC in the control engineering landscape since the last two decades is justified by its ability to handle nonlinearities, constraints and optimality concerns on one hand and by the availability of dedicated efficient optimization solvers on the other hand. In this talk, the basics of MPC are introduced and the main related research topics are shortly and progressively discussed including among others: Stability issues, Real-time implementation, Handling uncertainties via stochastic MPC, GPU-based parallel implementation as well as appropriate use of Machine Learning based blocs. Industrial and real-life examples are used throughout the talk in order to illustrate the underlying tools and methods.


Powered By Whova