Nuport Robotics just landed a deal with Canada’s largest retailer, Canadian Tire, and the Canadian government to test its self-driving truck capability. The technology uses lidar, cameras, and radar to guide trucks. And the company has a vision to make every truck self-driving … not just brand-new ones coming right out of the factory.
That’s a big deal, because there’s about 80 million trucks on the planet right now that don’t have self-driving capability. And transport companies aren’t going to spend hundreds of thousands of dollars for new trucks just for some new tech.
Not, at least, if they can get it aftermarket.
In this episode of TechFirst with John Koetsier, I chat with the CEO of Nuport, Raghavender Sahdev, about how his solution works, and what their new project with Canadian Tire means.
TechFirst podcast: making every truck autonomous
Apple Podcasts
Spotify
Google Podcasts
Overcast
Other podcasting platforms
Watch: Nuport wants to make every truck self-driving
(Subscribe to my YouTube channel so you’ll get notified when I go live with future guests, or see the videos later.)
Read: 80 million autonomous trucks
(This transcript has been lightly edited for length and clarity.)
John Koetsier: How do you make 80 million trucks drive themselves?
Currently, there are about 80 million trucks on the planet. Are we going to replace them all with self-driving versions … or give the existing trucks self-driving capability? That will make them cheaper. It’ll make them more environmentally friendly, more efficient.
It’ll also make them safer.
To dive into the question we’re chatting with Raghavender Sahdev, the founder of Nuport robotics. They just signed a major deal with one of the largest retailers in Canada, Canadian Tire.
Welcome, Raghavender.
Raghavender Sahdev: Hi, John. It’s a pleasure to meet you.
John Koetsier: Hey, it’s a real pleasure to have you as well. You’re releasing some big news. Tell me about this new partnership with Canadian Tire.
Raghavender Sahdev: So our partnership with Canadian Tire is involving the deployment of Nuport autonomous driving technology. And what that does is it’s currently involving two trucks and we’re enabling a more safer, efficient, and a more eco-friendly technology to be used to — in a way, improve the current transportation systems, specifically tailored for the middle mile transportation.
And the project entails us to have a predetermined set of routes, so two trucks running on a predetermined set of routes to be automated to increase the safety, efficiency, and increase driver experience; and on the whole, create a better transportation system in Canada to begin with. And at the end of the day, our target is to — starting with Toronto, we’ll be going national throughout different cities in Canada.
John Koetsier: So your business is autonomous trucking, building trucks that essentially drive themselves, and you’re working on kind of the short-haul or the medium-haul autonomy — that would seem to be harder than long-haul, right?
I mean, long-haul, you’re working on highways, freeways, and just driving straight for hours on end. Short-haul, you’re in cities and you’ve got congestion and you’ve got lights; you’ve got pedestrians and cyclists and other things going on. Is that correct? Is that tougher?
Raghavender Sahdev: Yes, short-haul in general, depending on the region in which you deploy, it depends on the region in which you deploy. Like from our perspective, we are not deploying it in downtown city driving. That’s not what we are targeting, which is the beauty of our chosen choice because we are doing short distances, but it’s not exactly city driving. It’s in the industrial regions, meaning it’s a bit away from the city, so there’s not as much traffic that you would see as you would find in a downtown city road. So you don’t have people just like illegally crossing or jaywalking, stuff like that. So you’re kind of safe from that aspect. Yes, there are other vehicles in the scene, but the minimum speed limit on the roads that we are currently operating on is on the order of 60-70 kilometers per hour.
John Koetsier: Okay.
Raghavender Sahdev: So that’s the key difference.
John Koetsier: So what parts of the journey are you covering? Presumably a truck is leaving someplace where there’s some goods in it that need to be transported somewhere else and going to another place — whether that’s a factory, whether that’s a warehouse, whether that’s a store — and then it needs to pull up to a loading dock or something like that. What parts of the journey does your tech cover?
Raghavender Sahdev: So our technology specifically [is] addressing, like for retailers, it’s between a distribution center to a nearby rail terminal or a shipping port. So the truck loads up and the goal of the truck is to transport a set of goods from a given point in a distribution center to a nearby rail terminal, and the average distance is around seven miles — seven to ten miles.
Our peak, we can do any route which is less than 25 miles. So the terminal point is all freight needs to, at the end of the day, go from a city, like from a distribution center and gets put on the rail. And then from the rail it’s transported to different cities in the same country, or maybe it’s going to a different — it might go from Canada to the U.S., or it might go within Canada. So that’s the terminal end points that we are operating on.
John Koetsier: Does the self-driving tech … is it in control the whole time from loading dock to loading dock?
Raghavender Sahdev: So it’s basically, typically operating from a staging area to another staging area at the rail terminal.
So it’s going from a Canadian Tire distribution center to a nearby rail terminal staging area. We do have the ability to do dock-to-dock as well, but our current use case is not exactly dock-to-dock.
John Koetsier: Interesting. Interesting. And is there somebody on the truck right now as well? Like a lot of self-driving companies that are bringing their technology to market, they have a safety driver in place. Do you have that as well?
Raghavender Sahdev: Yeah, we do have drivers in the truck because we must understand that the job or the role of a driver is not just to drive the truck, but the role of the driver is to make sure that the truck is in good shape before the truck gets driven; to make sure that the loading and unloading the freight — having the truck stops at the starting point where the truck comes from, enters the rail terminal or enters the distribution center.
So there is human involvement as well.
Currently we will be having a human in the loop as enhancing their experience still by leading to building a more efficient system. And at the same, it’s basically having an enhanced driver experience and reducing the bottom line for the clients.
John Koetsier: Right. Can you tell us a little bit about the technology that is involved? Are you using cameras? Are you using lidar? What technology are you putting in a truck?
Raghavender Sahdev: Yeah. So we are using cameras, lidars, radars, GPS — so we are using the full stack of sensors because we want to have maximum safety.
And so what essentially we do is we want to cover the entire 360° view around the truck, and for that to happen, we do rely on cameras and lidars — and radars, primarily — because each of these sensors has a unique set of features and abilities. For example, a radar works really well in snow and rain time. However, a camera might not work as well in a snowstorm or a heavy rain. That’s when our system has the ability to — our system basically incorporates multiple sensor modalities so that we have redundant systems in place. So that if one system fails, that system has the ability to fall back on another sensor modalities, and we have different algorithms which take care of this.
John Koetsier: So, that’s really interesting because you have multiple data sets coming in. You’ve got the grossest of the gross, the GPS data — I’m somewhere here on this map within, I dunno, 10 meters, five meters, two meters, whatever the case might be. You’ve got the radar, which is telling you there’s a car 20 meters ahead, 60 meters ahead or something like that. You’ve got lidar, which is actively sending out pulses of light and turning exact distances and details and dimensions of objects. And you’ve got cameras as well.
So you need some pretty sophisticated software to integrate all that into sort of situational awareness, correct?
Raghavender Sahdev: Yeah, that’s 100% true, and that’s what is our core set of expertise. So we are not a sensor supplier, we’re not providing these sensors.
We are basically using off-the-shelf sensors, integrating cameras, lidars, radars together, fusing that data using a proprietary, state-of-the-art sensor fusion techniques and incorporating it into different perception algorithms, motion planning controls, and creating the right set of maps and the right amount of data to be used to train these algorithms — and in future, also test and validate the robustness of these algorithms.
So that is something that is core to our Nuport’s technology, is the different AI systems that get deployed on an onboard computer using all these set of algorithms specifically tailored for short distances.
John Koetsier: How many sensors do you have to put on a truck? So you differ from your competitors in that you’re using existing trucks, and we’ll get into that in a moment. But as you take a truck that somebody says, maybe Canadian Tire says, ‘Hey, make this a self-driving truck’ — how many cameras, sensors do you have to put on this thing?
Raghavender Sahdev: So you need sufficient amount of cameras and sensors. You need the extent of cameras and sensors that it has the ability to cover all around the truck. So, you need to have multiple cameras, like, oh, I wouldn’t be giving the exact details, but I’ll just run some numbers here.
So you can have anywhere from 5-10 cameras, depending on the truck’s model. We have a standardized, retrofit solution, which is agnostic to the truck model. So we make sure we use multiple cameras, multiple radars, and multiple lidars to make sure we have the right amount of coverage all around the truck so that we don’t have any blind spots all around the truck, which basically establishes a safety cocoon around the truck. So that there’s no point around the truck that is not being covered by each of the sensors.
Camera covers everything. Radar covers everything. Lidar covers everything.
So we ensure if there’s ever an obstacle, we are able to detect it and act accordingly by telling our algorithms to behave in the right manner, following the exact traffic laws and behaving with the right mindset, if you would call it, for the algorithms.
John Koetsier: What’s really interesting about that is that we see data from Tesla, which not everybody buys Autopilot, right, which is their self-driving tech. But we see data from Tesla that the number of accidents that Teslas get into is far fewer than the average car, because they have such a rich sensor package.
That would be interesting. Do you have data on the accident rate of drivers, or even with the self-driving system, trucks that are using your data set and your software? Because that would seem to be a much safer truck to be driven because you have that great situational awareness.
Raghavender Sahdev: Yeah. So we do have some early results. I wouldn’t be able to share the right exact number, but we have seen, we have validated by using our technology you can have a drop in accidents. But it’s too early to comment on the exact numbers from Nuport’s side. We’ll probably be sharing some numbers later in the year, but right now it’s a bit early to share the exact totals.
John Koetsier: That would be super interesting to hear because, of course, if you have an accident, the costs of that are enormous. We’re not just talking about repairing the truck. We’re talking about a truck out of service. We’re talking about a replacement vehicle. We’re talking about maybe an injured driver.
Raghavender Sahdev: Exactly.
John Koetsier: We’re talking about insurance claims and other things like that. And if you can reduce accidents by 50%, 25%, 75%, you can easily pay for your system, which has got to be multiple thousands because you’ve got all those sensors and the integration there as well.
Raghavender Sahdev: You’re absolutely right.
John Koetsier: You know, you’re taking a different tack than many other autonomous vehicle startups. You’ve got Tesla, which is building a Tesla Semi. We see some examples of that out there. You have some other companies that are building self-driving or autonomous vehicles for Amazon and for others as well.
But you’re not building your own vehicle. You’re saying, ‘Hey, there’s a fleet of millions of vehicles out there; we’re going to make them all smart and autonomous.’ Why did you choose that direction? What’s the benefit?
Raghavender Sahdev: The underlying benefit of going that route is because we have an inherent difference in our business model of going the retrofit route.
That gives us the ability to scale faster, to get to market faster, and we can — like right now we’re in the process of deploying and validating our technology. Once it’s all validated and it’s ready to go, it’s much easier to get it to market and scale faster and deploy it with multiple other clients, because the client now does not have to buy an autonomous truck. They can use their existing assets and leverage Nuport’s technology and start saving money from day one.
So at the end of the day, the client’s life is much easier. We are building the technology, thinking about what a client would look at when they’re purchasing automated, autonomous technologies.
John Koetsier: That is really interesting because if you look at replacing a truck, I don’t know, you might be looking at a $500,000 number or something like that, maybe $300,000, who knows what the case might be. Do you have any sense of where you’ll come in for the price of your self-driving kit?
Raghavender Sahdev: Yeah, we have an exact estimate of what would be our price, but— [crosstalk]
John Koetsier: But will you tell me? [laughter]
Raghavender Sahdev: I wouldn’t be able to tell you. Maybe let’s catch up in a few months or a year, then I can give you the exact numbers, but I wouldn’t be able to share the exact number currently.
John Koetsier: Okay.
Raghavender Sahdev: Just because of confidentiality with that.
John Koetsier: I totally understand, that’s not a problem. I’m going to guess that you’re in the $5,000-15,000 range, somewhere around there. You don’t have to comment, wink, or anything like that … but I’m going to guess it’s somewhere in that range with the sensor package that you’re talking about.
Now, let’s talk about what autonomous trucks gives us, how it changes our world. You’ve said that there’s still a safety driver in there. That may not be the case all the time. Maybe in five years or three years it won’t be. How does making trucking and transport autonomous help the world?
Raghavender Sahdev: So, I’m going to specifically speak about autonomous trucking. So when you make autonomous trucking or when you incorporate smart technology into trucks to basically train or to basically have the next generation of trucks, what that does is it makes the whole system more safe.
So you don’t have as many accidents as you would have with the traditional driving systems.
It makes it more eco-friendly, so your fuel consumption is reduced. Just because the way an automated technology would have the truck driven with a driver, it will basically reduce your carbon emissions by a significant factor, which is again good for the environment.
The third is it is also responsible for increasing your operational efficiency. So the client can directly, what the client could basically do with maybe 10 trucks, now the client can do it with eight trucks or seven trucks. So at the end of the day, the client is increasing its operational efficiency, increasing the throughput of the distribution center, and at the end of the day they are able to have a more robust and a more safer and efficient and a more eco-friendly system.
John Koetsier: Excellent. So, TechFirst is about innovators who are inventing the future. Where do you see your tech in about 10 years?
Raghavender Sahdev: So in 10 years, I imagine a world where Nuport has scaled to become a multi-billion dollar company, and we have Nuport trucks all over, not only North America, but in the entire globe at different places in Europe, Australia, Asia, depending on where there is maximum need for autonomous trucking.
At the end of the day, trucks drive the entire supply chains across throughout the world. So you’re either going through trucks, you’re going through an airplane, or you’re going through a rail terminal, or a ship. So trucks, being one of the major contributors towards driving the supply chains and the transportation industry, we imagine our trucks to be throughout North America, Europe, Asia, Australia, and different parts of the world.
John Koetsier: Wonderful. Is there an OEM product in the future such that somebody can buy their truck from whatever manufacturer that they prefer but your technology is installed right at the very beginning?
Raghavender Sahdev: Yeah. We are working on a model. I wouldn’t be able to give you the exact details on that, but we are, we have a plan for doing something similar.
John Koetsier: Excellent. Wonderful. Raghavender, thank you so much for taking the time. I really do appreciate it.
Raghavender Sahdev: Thanks a lot, John. It has been a pleasure speaking with you, and I’m looking forward to the next few months and the next few years and see what are the next exciting steps for Nuport.
Big reader, huh? Subscribe to the podcast already
Made it all the way down here? Who are you?!? 🙂
The TechFirst with John Koetsier podcast is about tech that is changing the world, and innovators who are shaping the future. Guests include former Apple CEO John Scully. The head of Facebook gaming. Amazon’s head of robotics. GitHub’s CTO. Twitter’s chief information security officer, and much more. Scientists inventing smart contact lenses. Startup entrepreneurs. Google executives. Former Microsoft CTO Nathan Myhrvold. And much, much more.
Subscribe on your podcast platform of choice: