Live Mapping Is One Of The Underpinning Technologies Of Autonomous Driving - MTC With Jason Jameson

We spoke to Jason Jameson, SVP & GM for APAC Here Technologies, on how they are spearheading the journey to autonomous driving with their location and mapping technologies.

Q: Hey Jason, welcome to Making The Cut. Let's start by talking a little about HERE Technologies; tell me a little bit about what you do and what excites you about your job at HERE Technologies.

We were the company that invented the first digital map 35 years ago. We're a Silicon Valley startup and have since worked to create this digital representation of the world. We're owned by several companies, including the major German car companies and embedded across cars worldwide to help with vehicle navigation and things like advanced driver assistance systems on the path to autonomous vehicles. So in-vehicle navigation is a good example of the technology that we do. Q: Perfect! Tell us about some of the challenges in autonomous driving technology. Why are we still waiting to see as much of it? A lot of this has been done in what we call autonomous zones. So that the cities that have proceeded with autonomous vehicles have designated areas where autonomous cars can drive, other things that are slowing down the decision-making here are really the legal framework. Because the idea of having a car that drives itself means where does the responsibility start and stops between the car and the driver? And then there are all the other factors, vehicles on the road that are not autonomous add a complicated random dimension to what truly autonomous vehicles need to deal with. Many vehicles are starting to have these semi-autonomous functions, so it's a steady step towards full autonomy. Q: At the recent CES, we heard that BMW is launching a fully autonomous 7 Series; the CTO himself, Frank Weber, says that this is a level three autonomous vehicle and in autonomous driving technology, there are five levels, zero being no autonomy at all, up to level five. However, there is a disparity between levels two and three. Level three is very advanced technology when it comes to autonomous driving and of course, he is saying this is a level three BMW self-driving 7 Series. When do you see the entire car population reaching at least 50% in autonomous vehicles? It's amazing technology. Level three is where the driver is no longer making decisions; the vehicles make all the decisions for the driver. The reason it's level three and not level five is because it's conditional full autonomy, which means that the vehicle will drive itself under certain conditions. While that technology's there, it's dependent on many other technology foundations. So cars are increasingly becoming computer data centres on wheels, there's so much computing power and to do full autonomous, it's very expensive and you need very advanced sensors; you also need a high-definition map because the sensors only see so far. A high-definition map with precise positioning allows you to see exactly where the vehicle is on the road and around corners, beyond what the sensors can see so that the vehicles can predict and preempt what might happen. The other factor is things like traffic and weather conditions, all need to be considered as part of autonomous driving, so we're still some way off and it's hard to position ourselves with the current data. When you start getting into vehicle-to-vehicle communication, the vehicle sensors can determine, "Hey, these are roadworks." They'll recognise those cones that are mapped out there to block off the lane. So then the vehicles start talking to each other. Our platform already captures planned maintenance, so we know when roads will be blocked in advance for the traffic calculation and so on. Q: If we want to create this whole reality where you have an autonomous automobile environment, what role will HERE Technologies play in helping to make this reality for our planet? There is this thing that we are doing now with high-definition maps, which is a very precise sub-one-centimetre mapping stored in the car so that when it's driving, it has the sensors and maps telling us where it is. BMW, for example, is using our high-definition live map as one of the underpinning technologies for autonomous driving, so that's the role here that we play. We'll continue to evolve in many other areas; there is also R&D going on in spaces that I mentioned earlier, so it will keep evolving, but you know, that technology stack is enormous and there's so much to do. Q: One of the major challenges is that the ethics of tech is always far behind the tech itself. So when it comes to autonomous driving technology, what are some of the ethical considerations you must consider? When the vehicle is in charge of making decisions and in emergencies, there starts to be a calculus around what you do if you are challenged to protect the occupants or save lives outside. First, would you get into an autonomous vehicle that is programmed to kill you in certain situations? If you knew that was going to happen, the decision it would make. Maybe what you'll have is a dial that can say, "Here's your ethical spectrum" and you can select "altruistic" or "selfish". AI can have that sort of tuning. The thing is, it still has a lot to do with the regulatory framework. Ethics will be between the car companies and the purchasers based on the uses of the vehicles people choose. Still, the framework to accommodate the fact that it's not a human making those decisions needs to catch up; at what point does the liability sit with the manufacturer, the programmer, the driver, the other parties, etc.? Q: Speaking of a regulatory framework, for example, if we look at a traffic offence in some countries, having an accident with a jaywalker is a traffic offence. What if the autonomous vehicle is unable to avoid such an accident? Who is going to be held responsible? Is it even building towards those kinds of situations? It's not there at this stage. The existing laws are there and right now, the driver is responsible. In the same way, if the plane has some incident on autopilot, ultimately, the plane's captain is accountable, but then there's a secondary liability, etc. Q: When discussing the regulatory framework in this area, how about HERE Technologies? How are you positioned in this? Can you be a mediator? Where do you sit in this ecosystem? We certainly play a part in that; we have very close relationships with governments, particularly in places like Germany and elsewhere, where our technologies can help with safety. For example, seeing things like the legislation in Europe recently around intelligent speed assist, which is legislation for all new vehicles in Europe that require the car to be able to inform the driver about changes in speed zones and, in some cases, have the driver react to that change in speed zones, so we've actively worked with the government around those sorts of safety measures. As autonomous vehicles evolve, we'll also continue to work with the government in those areas. Q: So let's think about it from an ethical consideration, if there's no clear, best-possible scenario and the computer has to make certain decisions based on probabilities, based on available information. How do you think humans, in general, will take issue with that? People want to know when they're in the car, particularly about ethics and sustainability, whether you are in a vehicle that will make good decisions that will not be detrimental to your safety or others. You want that nice, happy medium where everyone gets cared for. So it will be an adjustment process for people; there will always be some unknown process that hits us. It's very hard to say more than that, but the fact is that technology will help play a role in minimising the risks and help reduce the opportunities for errors to creep into or prevent these unknown situations from happening. There will always be a human oversight on this. AI is artificial intelligence, but I like to think of it as augmented intelligence. It's designed to make your life easier but not to replace you using your own brain. There's a lot around that, and there will always be a role for humans and while they might self-learn and tweak their algorithms, we still provide those parameters that set the boundaries. Even something like ChatGPT is tweaked and tuned continuously to make sure it stays within the mark. Q: So, Jason, help paint a picture for us. How is the future going to look like when we talk about fully autonomous driving vehicles lining our roads, how is that going to look like? It's not going to be that different. I mean in the sense that you'll be able to, firstly, how do you get from A to B if you want to go out, or if you want to go shopping etc. and I think it'll be a matter of being able to tell which destination you're going and let the car pick you up from your home, which, you may not even own the car, it might be just an autonomy service. I'm sure people will want to own vehicles, but it will take you to the destination, find a parking spot, and drop you off at the door. Then, when you're ready to come out, you may not even have to summon the vehicle, it knows that you have left the shop, and you can walk out the front and the vehicle will be waiting for you. Because clearly, your mobile or your watch or any wearables you wear will speak to it. When you're driving, the vehicles will be talking to each other. So they'll have a hive-like knowledge of what's going on in the vicinity because there'll be a lot of macro-level dynamics, like, traffic at certain times of day will be well known because of people coming and going. The map and the location technologies will be at the heart of this.

---

This content is brought to you in collaboration with BMW Asia.