Uri Lavi: Two Wheels, Neural Nets, and Zig-Zags
Uri Lavi on How Ride Vision Predicts Dangers for Motorcycles and What Needs to be Fixed in the Smart Mobility Industry
Uri Lavi is the CEO and co-founder of Ride Vision, an Israeli startup developing Collision Aversion Technology (CAT) for two-wheeled vehicles.
Using neural nets and computer vision, Ride Vision’s solution analyses a road situation and predicts potential dangers for the riders, there are no other products in the market that can do the same for two-wheelers. The beauty of the system is that it utilizes only standard cameras as visual sensors, making it cheap and easy to deploy.
Uri and his co-founder and CTO Lior Cohen are both motorcyclists and both come from PicScout, an image recognition technology company.
What excites you the most about mobility industry right now?
For me, the most exciting and intriguing thing is that today in our sector new technologies are being introduced every minute, but most of them are aimed at cars — four-wheelers as we call them. If you look at motorcycles or scooters (two or three-wheelers), surprisingly, there is practically nothing that is being developed for them. This is quite amazing since it’s a huge smart mobility segment and it keeps on growing as the cities become more congested.
Being motorcyclists and technology entrepreneurs, me and my co-founder Lior could not ignore that, so we developed an Advance Driver Assistance System (ADAS) for motorbikes. In a way, it is similar to four-wheelers’ ADAS systems, but if in the car world there are many solutions like that, in our world we almost do not have any competition. The reason is that although we work on the same problem — minimizing collisions — the required solution is very different because bike’s behavior on the road is very distinct. No matter the size, bikes share lanes (cars do not do that), they maneuver, they move in zig-zags. This behavior calls for a completely different solution with a different user interface.
It’s very easy to overstimulate a motorcyclist with the alerts so that he or she would stop paying attention. How do you solve that?
This is the cornerstone of our technology. When you build algorithms for cars, the rules are more straightforward: due to their sizes, cars usually go one after another, mostly adhering to lanes. Our algorithm is not like that: it is able to analyze how vehicles surrounding the rider behave and understand the maneuvers the bike is making. Combining these inputs, it interprets the situation as potentially dangerous or not. So, our system makes intelligence predictions and minimizes false alerts.
When you studied the road data, what were the most surprising facts you have learned?
We found that data on accidents involving two-wheelers is very similar for very different countries. For example, you would think that many accidents happen due to a two-wheeler being an unstable vehicle and slipping in bad weather or due to poor road condition. But the actual stats say that 98% of fatal accidents happen due to vehicles collision, not due to weather or road conditions. In addition, a good percentage of bike accidents resulting from being hit from behind. For the four-wheelers, it’s almost always a forward collision, but motorcycles and scooters are smaller and less noticeable, so car drivers often just do not pay enough attention to them. For motorcyclists themselves, it is hard to deal with a threat that comes from behind, so our system does it for them.
Can such accidents be prevented? - Ride Vision
The Challenges of Riding a Bike on a Public Road As most motorcycle riders know, riding a motorcycle on public roads…
So, by simply alerting riders about that, you can already reduce the number of road accidents big time.
Exactly, and it is especially relevant for the markets where two-wheelers are extremely popular, such as Asia, Latin America, and Europe. But our system does much more than that — we constantly monitor the situation from all the viewpoints. The thing about two-wheelers that this is the most cost-effective vehicles and in most of the cases they are used on congested roads to get somewhere or to deliver something. This means they maneuver through the traffic a lot and we want to help them be safer when they do it.
What is the hardest challenge Ride Vision is tackling now?
Product-wise we are ready, so the challenge now is to continue Ride Vision’s growth without losing the speed. We are the first ones with ADAS for two-wheelers and we want to maintain that position.
If you look at the mobility industry, what do you think needs to be fixed?
There is a lot of hype about autonomous technology. But fully autonomous future will not be realized in the next 50 years and before that, we must solve many problems. For example, infrastructure and regulations. How do you prepare cities to live autonomous? I think there needs to be more emphasis made on how the solutions will be deployed in the urban environment. The trick is that we are not starting cities from scratch, there are already man operated infrastructure and vehicles on the roads, what do you do with them?
Another thing is — we really do not need to wait for long to make the roads better and safer. For example, right lanes can be opened not only for buses but also for shared vehicles, like the ones your company, GoTo operates, there can also be special lines for two-wheelers, you can put more of them onto the same line. So there is a lot of work ahead and it is not only autonomous vehicles.
What industry book would you recommend to those interested in transportation and mobility?
I’m reading now The One Device: The Secret History of the iPhone, which is a product and not an industry book, but it’s interesting to extrapolate the insights into what we are doing.
You can meet Uri and the Ride Vision team at the Ecomotion Main Event on June 11th in Tel Aviv Expo, at the Autonomous & Connected Sector, booth #25
Originally published at https://goto.global.