Why Simulation is a Driving Force for Autonomous Vehicles

Each year, roughly 1.25 million people around the world die in traffic accidents, the vast majority of which are the result of human error. In addition to the loss of life, traffic accidents result in an annual economic costs of $277 billion.

Traffic congestion and delays also have high costs. Drivers waste more than 3 billion gallons of fuel because of congestion and spend nearly 7 billion extra hours in their cars. This costs North America roughly $160 billion in 2015.

By eliminating human error, autonomous vehicle technology can solve these safety, economic and congestion problems.
 

Simulations Cut the Development Time and Cost of Autonomous Vehicles

That level of testing and training has extended development cycles. Remember, fully autonomous vehicles will require 8.8 billion miles of road testing. In comparison, Waymo, formerly Google’s self-driving car project, has completed only 20 million miles of physical road tests.

How can this process be accelerated?

Simulation will play a key role. Vehicle designers can use advanced simulation tools to virtually test these scenarios quickly and cost effectively. This not only accelerates development but also reduces the need for physical road tests.

For example, model-based development tools help designers create software for autonomous vehicles that have proven to be far more robust, prone to less errors and safer on the road. These tools can also cut development time in half and shave about 40% off the costs of code validation and verification.
 

The Benefits of Simulated Road Testing

With simulation, companies can quickly and economically complete the design explorations needed to develop a wide array of technologies that are required for autonomous vehicles (such as lithium-ion batteries and LIDAR systems).

Simulation also solves a number of other challenges associated with autonomous vehicle development. It enables developers to virtually test millions of driving situations in order to validate the vehicle’s high-fidelity sensors, actuators, LIDARs and cameras. Simulation can even drive the development and testing of complex automated driving control software.

Autonomous vehicle development continues to be a challenge of scale — both in terms of large fleets of vehicles and the compute resources required to crunch the data. Test fleets across the globe will generate petabytes of data on a daily basis. This data must be processed, sampled and utilized by multiple teams that work in parallel. Every update or change discovered during this development cycle can trigger additional design iterations. Simulation remains the best method for increasing the scale of development to meet these challenges.
 

Meeting the Autonomous Vehicle Challenge

At the recent Big Compute conference, Walt Hearn, regional vice president at Ansys, and Nidhi Chappell, head of product, Azure HPC/AI at Microsoft explained how advanced simulation tools are helping automotive companies overcome their development barriers.