Physically informed neural networks in the context of cloud-based vehicle functions.

Physically informed neural networks in the context of cloud-based vehicle functions.

CloudbasiertesFahren ITIV

Context

Physically informed neural networks (PINNs) can be used in the automotive industry to compute functions that have been moved to the cloud. Since there are few resource constraints in the cloud.

PINNs learn from a combination of data, such as sensor data and weather data, and physical laws. Predictions generated by PINNs can then be sent back to the vehicle, for example, to optimize vehicle efficiency.

Goals

  • Overview of the state of the art of physically informed neural networks and cloud functions.
  • Investigation of a cloud functionality, which will be implemented as a PINN
  • Implementation of the methods with subsequent optimization and validation

Requirements

  • High self-motivation, independent and solution-oriented way of working
  • Programming skills (Python, C++, ...)
  • Experience with machine learning is an advantage