Cancel
Showing results for 
Search instead for 
Did you mean: 

Virtual testing, validation of autonomous vehicle safety

Community Manager Community Manager
Community Manager

Automakers around the world are looking to mass produce autonomous vehicles, but they must first conquer some big challenges before that mass production can happen.


I recently spoke with Dave Lauzun, vice president of transportation and automotive at Siemens PLM Software, about these challenges and the steps companies must to take so production runs smoothly.

 

One of the biggest challenges is testing and validating autonomous vehicle safety. Dave said estimates from Toyota show that 8.8 billion miles (or 14.2 billion kilometers) of testing would be required to fully validate autonomous vehicles – which is impossible to do.

 

Instead, companies must look at combining virtual (totally computer simulated), semi-virtual (real vehicle with a simulated environment) and physical (normally driving) validation of these vehicles.

 

“There’s a lot of great work that has to happen in the validation area to make sure that autonomous vehicles are safer – and many feel many times safer – than conventional vehicles,” he said.

 

In this video, Dave talks more about these three types of validations, as well as the collaboration between Siemens and the American Center for Mobility to set up simulated environments for testing autonomous vehicle safety in crowded cityscapes, on highways and in intricate situations, such as going in and out of tunnels.

 

 

Watch our other videos to learn more about testing and perfecting driverless technology and the role software and sensors play in autonomous vehicles.

 

Tell us: What do you think is the biggest validation challenge for proving the safety of autonomous vehicles

 

About the author 

Toni Boger is the editor-in-chief of Digital Transformations, the Thought Leadership blog for Siemens PLM Software. As the content strategist for the Siemens PLM Thought Leadership program, she oversees the content creation, management, publication and promotion for all content in the program. She graduated from Saginaw Valley State University with a Bachelor of Arts degree in Communication and English. Prior to joining Siemens, Toni worked as an associate site editor for TechTarget, a technology media company.

Comments
Valued Contributor

Here are two challenges. First, making the miles spent testing really count by identifying and simulating as many fringe scenarios as possible. 8 billion miles of straightforward uneventful driving is meaningless. Second, creating highly capable simulation platforms with realistic graphics that focus on software-in-the-loop. Self-driving cars are fundamentally a software challenge, and depend on computer vision so near-real graphics and non-abstracted logic are critical for validation.