In the wake of COVID-19, many autonomous vehicle (AV) companies halted physical road testing of their AV fleets due to safety issues. These interruptions presented a huge engineering challenge for furthering the research and development of next-generation AVs, predicted to generate $7 trillion.

To be certified as safe, AVs must drive billions of miles perfectly, which consumes a tremendous amount of development time and cost. How could automakers generate this critical road test data while fleets remained grounded?

The pandemic has accelerated the use of simulation, enabling engineers to create cutting-edge AV capabilities and validate AV safety, virtually driving billions of miles in mere months. In fact, simulation has increasingly replaced the need for hardware prototype-based testing. We call this approach software-based virtual prototyping.

The pandemic has also sparked widespread AV consumer interest. In 2019, only a small number of people were interested in riding in autonomous cars. Now, AVs excel in delivering critical goods to those quarantined and in need, and they are quickly turning into a must-have resource.

AVs: disrupting cultures, saving lives

Today’s cars are extraordinarily safe, incorporating advanced driver assistance systems (ADAS) that increase road safety by minimizing human error. ADAS is paving the way toward an enhanced level of safety that fully autonomous vehicles will eventually achieve. As ADAS continues to become more advanced, consumer adoption of AVs will increase.

AVs continue to prove their merit during the current pandemic. Around the world, consumers rely on AVs to deliver food, COVID-19 tests, and medical supplies. This helps reduce human interaction and exposure to the highly contagious virus, saving countless lives.

Global adoption of AVs could also help eliminate human driving errors—the primary cause of 94% of traffic accidents resulting in 1.25 million deaths per year. Engineered to exceed human driving capacity, AVs will soon reduce these tragedies.

AVs will soon transform the automotive industry into a robo-taxi mobility industry that’s projected to generate $10 trillion by 2030. When cars drive themselves, consumers will pivot from buying cars to hailing robo-taxis when needed. Ultimately, AVs will create ubiquitous mobility, liberating many people—such as children, the elderly, and individuals with disabilities—who must rely on others for transportation today.

Speeding AVs to market

For OEMs to capitalize on these emerging AV opportunities and win the race to market, they must rapidly prove that AVs and ADAS are safe and reliable. How can this highly complex task be achieved?

One way is to physically drive billions of miles in challenging driving conditions and difficult weather scenarios. This impractical alternative remains expensive, incredibly time consuming, and carries substantial risk as it would be nearly impossible to anticipate and adequately test for every safety-critical scenario.

Only simulation can hurdle that tremendous development obstacle. Performing virtual safety validation helps OEMs drive millions of miles in days, provides dynamic insights into AV and ADAS safety and performance, and radically expedites product development compared to real-world road tests.

Significantly advancing AV safety

When AVs are widely deployed they will face unpredictable driving scenarios each day.

What happens if the AV’s advanced perception software encounters hazardous driving scenarios—known as “edge cases”—that are too difficult to classify because the AV’s algorithmic training didn’t educate it on the highly diverse real-world driving situations it would likely experience? For example, a dog darting across the street may exhibit an abnormal size or physical shape that may prove too challenging for the AV to identify.

Or in another case, perhaps the AV’s perception software misinterprets optical phenomena like sun glare as a physical object such as a car headlight. Teaming simulation with powerful artificial intelligence-based perception stress testing and a risk analysis system rapidly and affordably assesses and certifies AVs across these challenging edge case scenarios, removing risks to drivers, pedestrians, and animals.

Glare and other issues could also impact an AV’s sensor array—including cameras, radar, LiDAR, and more—which must be tested and certified to ensure accurate capture of safety-critical driving data. Failure of one or more of these sensors could deliver disastrous consequences.

How can these issues be safely tested? Simulation duplicates how the entire vehicle system will perform across a wide range of real-life driving scenarios, including complex weather and lighting situations. That helps engineers confirm that the AV and its sensors provide their intended functionality when facing countless road and environmental challenges—substantially decreasing the need to build physical prototypes and conduct road testing.

COVID-19 has accelerated the use of simulation to significantly advance the safety of next-generation AVs. AVs have also seen increased adoption around the world and are increasingly relied on to deliver life-saving medical supplies and food to those in need during the pandemic. As simulation continues to help validate AV and ADAS safety, AVs will be must-have resources for years to come.