Driving the industry forward: Q&A with Noah Lassar, Waymo

Sumit Dutta

Vehicles must not only be efficient, fast, comfortable – and in this industry, smarter than humans – they must also be reliable.

Noah Lassar, Manager of Reliability Engineering at Google / Waymo, is responsible for ensuring that the electrical components of the car keep working. One might say it’s his job to stress out the smart car in the lab – to anticipate faults in the field and nip them in the bud. We speak to him about the future of autonomous vehicles...

Noah Lassar

Do you favor fully autonomous vehicles or do you think the industry should draw the line at driver assistance?

The technology to make fully autonomous vehicles safer than conventional vehicles is already here. Why would we draw the line at driver's assistance? 

In your opinion, how reliable will autonomous cars be compared to conventional automobiles? What is the projected longevity of an autonomous car?

Autonomous vehicles can be just as reliable as conventional vehicles. Reliability isn't a function of autonomous vs. conventional. The added sensors and computing power can be made just as reliable as any electronic equipment.

What is the key concept in guaranteeing automobile reliability?

All vehicle components need to be designed to survive in an automotive environment and be tested to demonstrate that they achieve their reliability goals. System reliability must also be demonstrated by actually exposing the vehicle to the stresses it will encounter in the real world. In the case of autonomous systems, the stresses include not only environmental stresses like shock, vibe, temperature and humidity, but also situational stresses.

Is there a resource, strategy, or procedure that is under-utilized in the industry, but would go a long way to maximizing reliability?

The automotive industry hasn't fully taken advantage of reliability prognostics, or the ability to predict failure before it actually occurs.  There is a huge opportunity to improve the customer experience by fixing issues before the customer encounters them.  A tremendous amount of parametric information is being generated by vehicles all the time, and much of this information can be used to identify degradation so that corrective actions can be taken in advance of actual failure.

Presented in association with the Autonomous Vehicles Summit