September 23, 2020
Talk of fully autonomous vehicles is no longer new, even if progress might not be as fast as some early projections suggested. Despite this, investment in the development of more sophisticated, advanced driver-assistance systems (ADAS) on the journey to full autonomy continues at pace around the globe.
It's not just the technology that is evolving. Public sentiment is also changing. In an exclusive Ansys survey survey on autonomy, 80% of respondents said they would ride in a self-driving car, a 10% increase from the previous year. In addition, over 70% said they think self-driving cars are better drivers than humans, or will be in 10 years.
Given that the technology and consumer support are advancing in lock step, it’s no surprise that some analysts are suggesting the market for automotive electronics and software will double by 2030.
A modern car contains around 100 million lines of software code. To become fully autonomous, an estimated 10X increase is required — that's a staggering 1 billion lines of code. To put that in perspective, one of the world’s most advanced fighter aircraft, the F-35, contains around 10 to 20 million lines of code. As judged by lines of code, a fully autonomous vehicle could be up to 100X more complex than the most advanced aircraft. That represents a very significant engineering challenge.
For those in the race to market autonomous vehicle technology, demonstrating the safety of these highly complex vehicles in a practical timeframe and cost-effective manner is the overriding challenge. Trying to do this through physical testing would require hundreds of millions, if not billions, of driven road miles to cover the myriad potential scenarios a vehicle could encounter. In addition, addressing highly unusual situations, or edge cases, with physical testing is difficult, costly and potentially dangerous.
Simulation provides the answer to the challenge of autonomous vehicle development. In one example, a leading automated driving system maker reported driving more than 5 million miles physically, but more than 5 billion miles through simulation. That’s a 1,000X difference. This complementary approach of highly scalable, highly accurate simulation-based testing and targeted physical testing will make autonomous vehicles a reality.
Ansys Autonomy is an autonomous vehicle development solution that includes high-fidelity multiphysics simulation, automatic generation of safety-certified embedded software and functional safety analysis in an open environment. It is being used for optimization; simulation data and process management; workflow customization; and access to cloud and high-performance computing (HPC) resources. The solution is open for third-party tool integration and is backed by our simulation expertise.
Focused on safety by design and safety by validation, an overview of the solution components is shown below.
To explain how simulation and the Ansys Autonomy solution supports autonomous vehicle development, we have published a new e-book that details the engineering challenges underpinning the critical market needs, shows how some of the pioneers in the industry are addressing these challenges and articulates the impact of simulation on autonomous vehicle development.