Overview
In this webinar you will learn about how accurate, predictive, transient deep learning models of key engine and battery plant responses can be generated from measured data, using an automated process based on evolved Design of Experiments techniques and efficient test methods used over the past several decades in the field of Engine Mapping.
Highlights
- Automated end-to-end workflow from data measurement to validated models
- Deep learning plant models suitable for HIL testing and control design activities
- Design of Experiments (DoE) methodology used to ensure model coverage
- Test-point sequencing and steady-state detection methods to minimize test time
- Ordinary Differential Equation Neural Network model form for large number of inputs
- Model validation metrics to ensure objective continuous improvement of models
- Deep learning model training times less than 8 hours on a standard PC
About the Presenter
Peter Maloney is a MathWorks Senior Team Leader in the Simulink Automotive group. His main areas of focus at MathWorks are powertrain control and plant model development, powertrain calibration process development, and vehicle dynamics plant modeling. Before joining MathWorks in 2000, he designed, developed, and delivered spark-ignition engine control algorithms to production for engine airflow estimation, fuel delivery, air/fuel ratio control, and fuel system on-board diagnostics for Delphi Automotive Systems. Mr. Maloney has a B.S.M.E. from Texas Tech University, and a S.M.M.E. from the Massachusetts Institute of Technology.
Hari Rangarajan is part of the MathWorks Engineering Development group, where he works with Simulink Automotive Teams on electrification and autonomous driving projects to support latest shifts and trends in the automotive industry. He graduated from Ohio State University with a Master’s in Mechanical Engineering in 2021 and was the Control Systems team lead for the EcoCAR Mobility Challenge Competition in year 2020-21.
