A Smart Device on Wheels
John Culliton, a Ride and Handling Engineer with the electric vehicle company BYTON, will be one of the speakers at the upcoming Traction Summit Conference, (https://www.tractionsummit.com/) which will be held in San Jose, California, Oct. 16-18.
In this interview, Culliton touches on some of the steps being made by companies such as BYTON as autonomous vehicles move from the theoretical to the practical.
How does BYTON differentiate itself from more traditional/conventional vehicles now?
We are seeing a general trend toward more connected vehicles. I think BYTON is working on the next step evolution that is yet to be seen. We are one of the first OEM to partner with Amazon to have Alexa voice assistant in our car. We are working on technology for facial recognition. So when you get in any BYTON vehicle it can recognize you as yourself and configure all your settings, including seat positions, UI and UX settings, and lighting to what you would do in your personal vehicle. It is pushing the boundary of what next-generation technology can accomplish.
For one example, we have a tablet in middle of steering wheel that can be used to control our massive shared-experience display. The 48-inch screen can be controlled five different ways: voice, touch, hard buttons, gestures and facial recognition. It will be the first vehicle on the market to control through these five different methods.
What are the growth prospects for autonomous vehicles?
There is explosive growth for fully autonomous companies and investments. Both Silicon Valley VCs (venture capital firms) and OEMs are all over fully autonomous hardware and software companies. There also are fleets of autonomous vehicles driving around highly mapped parts of Silicon Valley, Phoenix, Pittsburgh, Las Vegas, and other locations in ideal conditions such as sunny clear days.
One encouraging fact is that the disengagement frequencies (how frequently a human driver takes over for the automated system) are decreasing rapidly, but we are still at a non-zero point. Furthermore, the methods for reporting disengagements are inconsistent and questioned by industry professionals.
There are companies promising fully autonomous robotaxis within the next 12 months. However, I believe these claims are overly optimistic and that the world of level 4 autonomy is father off than we think.
That said, basic level 1 and 2 features like automatic emergency braking, lane keep assist and lane centering are becoming commonplace, even in entry-level vehicles.
We are seeing a rapid influx of basic features as well as more advanced functions such as Super Cruise and Auto Pilots, which border on level 3 hands-free usage. And a few vehicles will have fully autonomous interstate driving, hopefully by 2020.
There are a few advancements that I believe need to be made before we can usher in the new, fully autonomous world of vehicles. One of them is non-linear dynamics control. Lots of this comes down to tire and its interface with the road. Most autonomous vehicles are using simple linear vehicle models, which is totally adequate for the controlled environment they are being tested in. However, no one is really addressing the limits of handling such as evasive maneuvers. Say the average driver in the Midwest has to deal with black ice in the middle of a turn a few times during the winter. Autonomous algorithms are not quite ready to do that yet.
What they need is real-time friction estimates, as well as prediction of what the friction state is going to be a few steps in the future.
That can be based on road and tire conditions as well as weather conditions. You can have a variety of sensors scanning the road, the environment, what the tire is doing. And you need to feed that data back into the algorithm to know what the friction is and going to be.
We are getting there, but we still have a long way to go. Another thing we need to catch up with are legislative standards. One difficulty we have is machine learning, the backbone of most autonomous driving algorithms, which is nondeterministic. You can give two identical machine learning algorithms the same problem. But, depending on their training data , they (the vehicles) can come up with different answers. As a society we need to come up with an agreed-upon set of scenarios that can be tested with these algorithms as well as passing outcomes.
As for progress in this area, there is hope. NHTSA is working with numerous third parties to develop industry-wide standards and testing protocols.
Their prediction is fully autonomous vehicles will hopefully see public roads in 2025, which is realistic. We can have lots of these problems sorted in 5-6 years.
How do tires fit into the autonomous vehicle world?
Tires are a critical component with any vehicle. And there are a variety of new challenges when trying to optimize for an electrified autonomous ready vehicle. Some of these involve connecting tires to the rest of the vehicle as well as the internet of things (IoT). We are working on predictive, cloud-connected algorithms that will give consumers and BYTON development teams insight as to how tires are used on our vehicles. This allows a start up like us to quickly build up our use cases and service trends across all markets as a newcomer to the industry.
We don’t have the legacy data that Ford or GM have. We need to learn quickly how vehicles are used and how tires fit into that and what the consumer is looking for to get the most out of them. I think we are in a good place off the bat with our tire portfolio. But we will be using data to quickly update our use cases and how we integrate the tire with the rest of the vehicle.
We also are investigating up and coming smart tire software and hardware. Just about every major tire manufacturer is working on some sort of smart tire device that can give you anything from an extension of the TPMS (tire pressure monitoring system) sensor to real time load treadwear and force generation characteristics. That last part is huge for autonomous driving. The tire is your only interface with the road. They have all the control authority over what the vehicle does. An autonomous driving algorithm has to know what sort of condition the tire is in to have full authority over what the vehicle does.
As for more innovations in the future?
We are leaning very heavily on our Tier 1 suppliers to help us to have any new technology as soon as it is ready. IN many ways, this includes the technology necessary for full and partial autonomy. We have access straight in the pipeline for the newest features in regards to autonomy and we will look to have it as soon as it is ready for the market.
Another benefit for connected vehicles like ours is an endless possibility future features via over-the-air updates. We will have that at launch and will really open up what we can do with our smart device on wheels. It will be constantly updated with newer features. We are not interested in planned obsolescence like many companies are.
To hear more about John Culliton and how BYTON is seeking to help revolutionize the automobile industry attend the annual Traction Summit Conference. To register, click here. https://www.tractionsummit.com/event-registrations