Supporting Autonomy with Engineering Simulation

The focus of this ASSESS Initiative theme is to investigate opportunities, requirements, issues, and approaches related to Engineering Simulation to support the rapidly growing desire for autonomy of future products and processes.

There is a significant push across multiple industries (especially Automotive & Transportation) for Electrification, Autonomy and Connectivity. The validation of an autonomous system needs to be performed via simulation since the physical testing requires extensive modeling of the application domain of the product or process (e.g., millions of miles for automotive autonomy). The application domain for any autonomous product or process requires the creation and simulation of 1,000s of virtual scenarios to capture the product behaviour, control systems, perception systems, and their connectivity.

Engineering Simulation will play a major role in in the modeling of these 1,000s of scenarios and the prediction and validation of autonomous behaviour across the scenarios. This will require new approaches to enable a significantly broader range of simulations and real-time feedback driven simulations. The number of simulations required to support autonomy is much larger than the number of simulations currently used in engineering design and these simulations need to understand or interface with sensor data, software, and control systems.

The magnitude of the number of simulations required along with a need to ensure “full” coverage of the application domain will likely require leveraging Machine Learning in multiple ways to better leverage simulations completed and to determine simulations required.

The ASSESS Initiative will explore the different factors related to Engineering Simulation for Autonomy to provide understanding and guidance moving forward.