There is a significance of the invention of consumer airplanes where everybody can buy a ticket and fly. A self-driving car will be more significant than that. – Mr Sanjay Sood
Autonomous driving is the new buzzword that all the (Original OEMs are talking about. It will not only completely transform the way people travel but also improve the safety, comfort and efficiency of vehicles to a level beyond imagination. Innovations in-vehicle sensors, processing power and digital maps are driving the autonomous vehicles revolution. Companies like the HERE Technologies are leveraging their vast experience in digital mapping and location-based services, developing innovative technologies to power autonomous vehicles and smart cities of the future. Mr Sanjay Sood, VP- Highly Automated Driving, HERE Technologies talks to Siddharth Shankar of BW Businessworld about how digital mapping technologies are the key to a smarter, connected and autonomous future.
What is the role of HERE Technologies in the Automated Driving at a global level?
We are a leading provider of HD mapping technology; we have signed HD mapping deals for autonomous driving cars with BMW and Daimler. We have taken a leading position in terms of forming the eco-system strategic partnerships then we are investing 100s and millions of Euros in self-driving technologies, specifically HD maps and we will be a main part of the eco-system moving forward in terms of self-driving vehicles.
What is the One Map Alliance?
We have established the One Map alliance, which is essentially a global alliance between technology mapping providers that cover the global footprints. In China, we have paired up with a company called Navinfo which is owned by Tencent. In Korea, we have partnered with SKTE (South Korea Telecom), the largest telecom in South Korea. This allows us to offer a truly global offering for HD maps because, in certain countries like South Korea and China, you cannot be a foreign company mapping, you must have local partnerships. We have these partnerships in place and we are creating global specifications. So, if you are an automotive company and you want to build a self-driving car and you want it to work in China as well as have the same platform work in South Korea, Europe and Japan, you can work with the One Map Alliance to get a global offering that is compatible across the globe.
What do you think are the current trends in automated and connected vehicles?
Connectivity and intelligence are the biggest trends that we are seeing. One interesting thing that we are doing with many of our customers is we are taking the data from the vehicles. We are pooling it on our cloud and then using it to create services that make the vehicle smarter and safer. So, a couple of years ago we announced a road hazard warning which provides information from vehicles, like the real-time information about anti-lock braking system, or deploying of the airbag or whether the windshield wipers are on, etc. After receiving this data, we pool this it from many vehicles on the road and then serve dynamic warnings to drivers in that area. So, for example, you are driving on a highway and because the cars which are kilometres ahead of you have their windshield wipers on, your windshield wipers automatically get turn on.
Can you share some light on HD Live Map?
We are working on some sort of next generation of mapping content and technology, particularly towards self-driving cars. The maps you have seen for the last several years or the last decades are what we called navigation maps. They are meant for human navigation and other types of endpoints. One wants to go to the store, they type it in and then the map directs the human there. HD Live Map is a very different type of product, it's still a map but it's meant to be a machine to machine map. It's geared towards, self-driving vehicles and we believe, in terms of self-driving, you need to have a high definition map in the vehicle to help the automated driving system understand how to safely and comfortably operate.
When we think about self-driving cars, there are typically three main ingredients to make it work, you have the perception side and this is all the sensors that are out of vehicle(self-driving vehicle) so there's the camera, the lidar, the radar, the ultrasonic sensors, everything helps in perceiving the world around it.
The second main ingredient in path planning is the software, the AI brain that sits in the automated vehicle that decides what to do.
The third main ingredient is an HD Map. The HD Map provides the whole bunch of content that's needed for self-driving cars. For example, it provides a global representation of the world around the self-driving vehicle. These vehicles have sophisticated sensors but they cannot see around physical objects, so if you are driving on the highway and there is a truck next to you, the sensors cannot see through the truck to understand the other side of it. So, an HD Map provides visibility beyond the sensor horizon whether there is an object or not beyond the sensor range of around 100, 200, 300 or 400 metres ahead of the vehicle. It also provides important semantic information on how to navigate the roads.
What kind of data is present in these HD Live Maps and what is your coverage?
We build the map data providing essentially a very high-fidelity representation of the road. So, the map contains a set of layers of the information. The first is what we call the road model. It includes everything like speed limits on the roads, lanes you can turn, in which direction, etc. So, if I am in the right lane, can I make a left turn at the intersection or not? It has all the rules of the road that help govern legal driving manoeuvres. And then we provide essentially the lane model. That is the second layer. What the lane model has is essentially a precise 3D blueprint of the road surface. How wide is the road, how wide are the lanes, what colour are the lane markings, how wide is the shoulder? So, if the car must pull over in an emergency, it can pull over on the side of the road. And the final layer is the localisation layer. What that contains are landmarks around the road. Things like poles, trees, signs, overhead sign face. This feature allows the vehicle to localise and precisely position itself on the road network because GPS is not good enough. You have a GPS in your vehicle but it is only good up to five meters of accuracy. Five meters on a roadway is a huge distance. Further, the localisation lets you precisely position yourself on the roadway. HERE Technologies have built this content on a continental scale. Today we have full coverage on all the highways in North America, all the highways in Western Europe. Then we are working with numerous car companies and technology companies to integrate onto their self-driving systems.
What challenges do you think, autonomous and connected vehicles and all these smart mobility solutions are facing globally?
I think from a connected vehicle standpoint, privacy is a big issue. Once your car is connected to the internet, this privacy issue is that we must deal with in terms of data from those vehicles it is a personal item that you own. You want to make sure that privacy is maintained. For HERE, privacy is a primary concern and it reflects in our design and implementation of the technology around our connected services and vehicles.
With automated driving, there are three big challenges at the global stage, including India as well. The first one would be legislation. How do you legislate and how do you have liability in terms of a legal framework and when you have self-driving vehicles, you can potentially be involved and because of an accident. I think that is not even properly explained in the United States and Europe. There is still legislation being developed by the government to understand how this will be applied. The number two issue, as discussed earlier, is user acceptance. Do customers want self-driving cars or are they comfortable for it? The third one is technology issue. When is the technology going to be good enough to be reliable or with a high degree of confidence to deploy this technology because having an automated work on a test track where you mapped and work for hours, hundreds and hundreds of hours, is one thing? But then when you take it to a different city or a different environment, you must certify the system all over again. There are significant technology challenges that we still need to solve.
But what we have seen is billions and billions of dollars being poured into this industry because it fundamentally changes the way we do transportation. There is a significance of the invention of consumer aeroplanes where everybody can buy a ticket and fly. A self-driving car will be more significant than that.
What do you think about High Bandwidth 5G Spectrum in terms of future of mobility?
I know 5G is a large strength that we are seeing, HERE Technologies is working with many of leading providers of 5G technology in the various market across the world. We see 5G being important for connected vehicle, one because of the bandwidth increases by many orders of magnitude. For a self-driving car, it lets you do in-vehicle information processing so, instead of having to take data from the vehicle for self-driving car and move it to the cloud to process, you can do a lot of that processing on the edge because of 5G infrastructure which allows for edge computation. So, we are doing quite a bit of R&D and quite a lot of effort with these leading telecoms to understand these used cases, both from connectivity but also computation, distributed computation perspective.
How are you handling the competition from Google Maps?
There is quite a lot of competition from Google Maps. We take pride in differentiating ourselves in a couple of important dimensions, one is essentially the core automotive offering that we have is really strong in terms of an embedded navigation system, in offering content that can be integrated into a variety of different endpoints. Fundamentally that means is if you want to use Google maps, you have to use the entire stack, you have used their website, you have to opt into their tracking and advertising. With HERE technologies, we provide content separately. The biggest differentiator between us and Google is that we provide just the content only. however, we also provide location services and we do provide services and embedded software for customers that want that, but we don't require somebody to take the whole end to end stack. We provide content separately, so if you have your own rendering service or routing service or search, you could build that yourself on the top of our content or you can take our services and our content together. That's one of the areas where we differentiate ourselves. In such terms we monetize our content and our services unlike Google. They monetize the user and their whole business is predicated on advertising and monetizing users intended behaviour, we don't do that. The third biggest area is we operate as an open and independent platform so we allow sharing of data, pooling of data and it's much more of an open eco-system that you would see with Google in terms of the way they operate their business and their monetization models.
Minister Nitin Gadkari had said that he will not allow autonomous vehicles in India, where you think India stands in terms of connected Vehicle?
I think a connected vehicle is a global trend and I don't think India will be excluded from that. Connected vehicle and autonomous vehicles are very different but I think connected vehicle are in trend because users are going to demand access to services and content that can only be served from a cloud infrastructure. With autonomous vehicles, I think India possesses a very interesting and challenging technical problem. I am from the United States and in terms of infrastructure, it is much different from what you see in Asia. I am in Mumbai today and here when my team and I are driving around, we always think of the complexity of making a self-driving car work in India. You can make it work in India, you can make it work anywhere.
Tell us about HERE Technologies's Indian operations?
We have a very large presence in Mumbai. We have over 4000 employees, not only on the operations side but also in technology. So, I am sitting in our technology office in the NESCO centre in Mumbai, where we have an incredible workforce of highly talented engineers working on all parts of our technology stack, from infrastructure to technical operations, all the way to machine learning. I think we are driving state-of-the-art not just in India but on a global stage, using our workforce here in India. We are very proud of that and continue to grow our presence in India. We see a wealth of talent and opportunities to expand our workforce.
