dSpace aligns with Cognibit for AI-based traffic simulation models

dSpace aligns with Cognibit for AI-based traffic simulation models

Business news |
By Christoph Hammerschmidt

Traffic situations are shaped by the individual behaviour of road users. Emotions such as fear, surprise and joy or limitations such as limited vision of road users also play a role. Only if all these aspects are realistically included in the simulations for the development and testing of functions for automated driving will highly automated and autonomous driving become safe in practice. The AI-based driver model Drivebot from software company Cognibit (Munich, Germany) makes it possible to replicate human-like behaviour in the simulation and recreate realistic traffic scenarios.

In a first step, dSpace will couple its simulation environment Automotive Simulation Models (ASM) with Cognibit’s AI-based driver model Drivebot in order to include unforeseen behaviour of human drivers in simulations. In the future, behavioural models of other road users such as pedestrians or motorcyclists will be successively integrated into the simulations. 

Realistic driver models play an important role in AD/ADAS development. This is because the functions for autonomous driving are designed in the first step for the so-called “operational design domains” (ODD), i.e. for the specific operating conditions. The automated driving systems must function in these defined operational domains. If this range is left, the system prompts the driver to take over control. To make the assistance systems even more realistic in the next step, additional factors can be included with Cognibit’s AI-based driver models. 

Cognibit’s AI-based driver model can be used not only for the first-person vehicle, but also in a multi-agent simulation for the surrounding road users in the ASM simulation environment to be able to simulate realistic surrounding traffic. The realistic movement of the Fellows allows simulated traffic scenarios to be efficiently varied in order to identify corner cases and thus determine the limits of a driving function. 

“Driving functions for SAE Level 2 up to 5 need to be carefully tested in terms of safe interaction with human road users,” says Christopher Wiegand, strategic product manager of the Automated Driving & Software Solutions business unit at dSpace. “Incorporating neuroscience and sensorimotor insights into the simulation will produce valid and meaningful simulation results and make driving functions safer.”

Related articles:

KPIT, dSpace, Microsoft join forces for homologation of autonomous vehicles

dSpace acquires stake in AI start-up neurocat

dSpace, BTC join forces for ADAS simulation and validation

CarMaker 11.0 simulation software offers new functionalities for virtual test driving

AVL, TU Graz develop adaptive simulation software for automated driving

Co-simulation platform enables cross-company virtual prototyping


If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles