Wireless Technologies

Ex Googlers: Humans will see that machines driving trucks is safer

While Google is busy perfecting its driverless car, former Googlers Anthony Levandowski and Lior Ron have joined forces to create Otto, a start-up aimed at equipping existing trucks with self-driving technology. According to the founder’s post, current trucks on the highways are causing fatalities every year and truck drivers experience a low quality of life. Otto want to change all of this by making trucks drive themselves thereby letting truck drivers rest and attend to other duties on the road.  

I catch up with co-founder Lior Ron by phone to find out more.


Why did you decide to form a separate company rather than work along Google’s current AI Car project?

Both me and Anthony spent nine years at Google and we have a lot of respect and appreciation for Google. I don’t think we would be here and talking about autonomous vehicles if it wasn’t for their vision and investment in the last decade. That being said, Google is pursuing a different market to what we are doing. Google is focusing on driving in cities. We are focusing on solving a very different problem, which is driving on trucks and highways. Its two different businesses and we do hope to even collaborate with Google in the future but the focus is just very different.

How are you funding this?

We are all self-funded to date. The group of us are a great team of self-driving veterans and experts that have been in that field for a decade in companies such as Google, Apple and Tesla. Because everyone is so excited about bringing the solution to market sooner rather than later we were able to fully self-fund the company to date. It’s also allowed us to make super-fast progress in assembling the technology and developing the team.

Your team at Otto consists of an interesting mesh of experiences with people from Google, Apple and Tesla. What’s it been like working together?

It’s been so exciting and a very special experience. When you have a group of such high calibre engineers coming together they hold each other to a high standard. We’re all familiar with self-driving technology so this is really about - this time making it happen. The culture is very unique which allows us to progress with urgency, speed and with a high level of execution.

What is a typical truck driver’s day like at the moment?

Truck driving is very much a noble and much needed occupation. At the same time it’s quite challenging. Truck drivers are being asked to drive more and more because demand for trucking is getting higher and higher. But the daily lives of truck drivers is challenging. Having to drive a truck to begin with - you need a lot of expertise. But then having to ship a cargo as soon as you can from one point to another, which can be thousands of miles away, means that you need to drive in as much time as you can. During that time it’s relatively mundane - like keeping your truck between the lanes on very long stretches of highway.

On one end there is this thing to push the drivers as much as possible because time is money. On the other hand, if you look at the fatality rate and accidents it’s very hard for humans to focus on driving for more than a few hours a day. Which is why right now, in the US and Europe, there is something called ‘hours of service’ that basically mandates truck driving to only happen nine to 10 hours a day which roughly amounts to 60 hours a week. The rest of the time, the truck driver has to rest and get a good night’s sleep. But the truck driver is pushed more and more by the pressures and necessities of the job  - so there is a conflict between getting the job done and driving the truck safely.

Almost 70% of cargo transport in the US is on trucks. It’s basically the backbone of the economy. So demand is more but supply is not catching up because it’s a hard job to attract new people into the profession. As a result the average age of the truck driver is 51, the churn rate in the industry is 91% - meaning you always have to hire two drivers to do a job of one every year.

So presumably autonomous trucks will free up more time for the truck drivers…

We can help the driving become safer and make the roads safer and enable the truck drivers to become more productive. So we are taking existing trucks and enabling them to self-drive. The driver is still in the cabin and instead of actively driving for hours a day, he can basically supervise the truck. On long stretches of the highway, the driver can take a nap in the cabin and he can attend to other duties without having to drive throughout the day. So the productivity is doubled. This means more livelihood for the drivers. They can finish the job faster and get back to families.  

Are there plans to get rid of the human truck driver completely at some point?

Definitely not in the near future as this is going to be a process. For the foreseeable future there will be a truck driver in the cabin. You still need someone to take the truck to the exit, someone to address any issues with the truck in unsafe weather or if the human needs to take over the truck. The main difference of our approach compared to the rest of the market is that we do aim for full self-driving capability on the highway, meaning that the truck driver is still in the cabin but can rest and take a nap.

There will be some natural job revolution. It’s happened to other professions when technology came in over the years and we have seen many examples of that in the economy. For us the question is: what will make a safer future and how can we actually eliminate those unnecessary thousands of deaths on the highways? How can we scale the economy to make sure we have enough supply for cargo shipments to do the job needed?

Have you faced any resistance from truck drivers? 

I think most of them are excited. They want to drive more so they can get back faster to their families and make more money. Right now they only see their families maybe a hundred days a year. So far this is the response we’ve gotten. I’m sure as time goes on there will be more debates as we extend the technology but so far it’s been positive.

People are generally having a hard time adjusting to driverless cars let alone driverless trucks. How do you think they will respond to a self-driving truck?

I think as people when we face new technology we tend to adjust our thinking based on what we know and don’t know. I think right now a lot of people are not fully informed about the benefits of the technology so I think this will change over time. The onus is on us to prove the safety benefits and once we show that the machine driving the truck is far safer than the human driving the truck than that will speak volumes. I also think new stuff takes time to get used to.

When you read something you can’t visualise - it’s a big jump. Plus this is a step by step process. Initially we will have the truck on the highway with a safety driver in the backseat just logging data. Then slowly but surely the truck will start driving itself initially for five miles then 10 miles [and so on]. The public will have time to get used to it but this is also a dialogue – about what it will take to bring self-driving trucks to the market.

What sorts of things will the technology on the truck be tracking?

We are tracking the overall compositions of the roads from all angles. We are looking at conditions of the highways, lanes, cars, debris, how safe it is to be on either side. The nice thing about a truck from a technology point of view is that you can mount the sensors on a high vantage point on top of the truck which is almost double the height of a car and allows you to see much further down the highway. Why do people like driving SUVs verses regular cars? It’s because they can see more and feel safer.

The other benefit in developing technology specifically for trucks is we only care about a smaller subset [which is the highway]. If you wanted to drive in the city, you would need to map every street in every city in every state you are going to. In our case there’s 220,000 miles of interstate in the US. We only need to map those and make sure we are doing a fabulous job in delivering on safety and performance of those routes. So it’s easier for us to test on highways and intimately know that to show safety stats.

Any specific challenges while testing this technology?

Nothing too entertaining to share right now, it’s been relatively smooth-sailing. I think the main challenges will come later. The bar for deploying these technologies on the roads commercially is high. We need to show the safety stats so we are embracing the challenge.

What components would you like to add in the future?

Think about all the options that open up once you have this specific technology. In the US, 15% of trucks are running empty because it is very hard to coordinate cargo and trucks. Once you have everything connected you can do a much better job. 28% of the pollution on the roads is because of the trucks. So once you have self-driving trucks you begin thinking about managing fuel efficiency better and what this means for the engine of the truck.

The UK wants to test driverless “HGV platoons” but there are concerns that they won’t work in the UK. Are autonomous trucks only suitable in certain countries?

We are focusing on the US but we do whole heartedly believe that this is something that can be used globally. Platooning is an interesting technology and has some benefits like fuel efficiency. It also has some challenges like how to navigate the trucks when coming off the exit.  The trucks have to be at the same time and same place, maybe even from the same fleet. So it’s very hard to coordinate and really scale that solution. That’s why we are just focusing on the single truck rather than platooning.

When will these trucks officially start driving on the highway?

I don’t have a specific timeline but we are moving with urgency. We started Otto in January and since then we are now 43 people and have a fleet of trucks equipped with the basic technology. We just completed our first driverless test and drove many miles with that. We are progressing as fast as possible. We still need to perfect the technology and show the safety benefits. Then we will be ready to bring it to the commercial market.


Also read:

London’s driverless pods are coming but how will the public react?

Google’s ambitious fully driverless goal still needs work

Driverless cars: Is semi-autonomous the future?

Driverless cars in the UK by 2030?


« Rugby player data offers plenty for hackers and companies


C-Suite Talk Fav Tech: Jeff Loeb, Ipswitch »
Ayesha Salim

Ayesha Salim is Staff Writer at IDG Connect

  • twt
  • Mail


Do you think your smartphone is making you a workaholic?