What's New in ASCMO-STATIC & ASCMO-DYNAMIC V5.15

MLP models can be exported to ETAS Embedded AI Coder to efficiently run these models on an ECU.

File menu > Export Model > Embedded AI Coder

Export to Embedded AI Coder
Model training, Automated Machine Learning, and various optimizations can now use Export Job to Docker to run the optimization in a Docker container, e.g., in the cloud.

Model Properties (ASCMO-STATIC)

Automated Machine Learning

Single Result Optimization

Multi Result Optimization

Global Optimization

Model training and Automated Machine Learning can now use Export Job to Docker to run the optimization in a Docker container, e.g., in the cloud.

Model Configurations (ASCMO-DYNAMIC)

Automated Machine Learning
LSTM models without peepholes can be exported to ETAS Embedded AI Coder for efficient deployment on an ECU.

Export to Embedded AI Coder

ODCM: The output sigma is now displayed as an untransformed value to improve comparability with real-world measurements. Due to this algorithmic change, ODCM projects are automatically converted. If those projects have active sigma limits or thresholds, they are reset to NaN. Users can redefine these values manually in the ODCM Manual Frontend. A log message informs users when such a conversion occurs during project loading.

Additionally, an automatic mode for the sigma threshold has been introduced. The predicted output for a new measurement is now also displayed.

Manual ODCM Front End