Sensing technology for autonomous vehicles


[Music] [Music] [Music] [Music] [Music] [Music] [Music] [Music] [Music] [Music] [Music] [Music] Yes. So, my name is
Henry as a technical expert within Volvo Cars. So, I’m going to talk a little bit about The Drive Me Project. There are some updates. Maybe some
of you have seen this presentation already before I have updated it’s likely with the newer things. It’s not so technical, so it’s more about how
you know the focus of the project and so on. And then we end up with the technical parts in the app. Looking also a little bit of the new cars that we
have been manufacturing. Okay, thank you. Now, I need to use this one. Okay, so anyway The Volvo Philosophy and cars in general is all about freedom right. So, this is really the focus I
think also for autonomous vehicles in a way. So, it is still about freedom
when it come to autonomous vehicles because its freedom to do whatever you like when you drive and that’s really
important because you can basically keep your life more time and that is a very precious thing today. Today the
fact is that when you look on the roads like here in the US it’s really
really crowded. People are sitting in cues every day, every week, every year and of course they do a lot of other things. And during that time when they do other things which is about 20% of
the time if you looked deep into you know monitoring systems, we have had
in research vehicles on the road and we can see that approximately 20% of
the times, people are not looking on the road. And that we did up with these active safety systems but of course people would like to actually grab more time and then we go in to autonomous
driving. So, what are the Global Megatrends? It is a lot about the global organization and also the number of
mega cities that this coming. I mean you can see the number of megacities in EU is not really growing but for sure
they are in China and in Southeast. So, and that also there are qualities
that made you healthy issue. For sure when you drive as a driver, you are not really good on using your throttle and brakes and so on, so by actually
taking that into the autonomous vehicle you can do a much much better
driving situation and that also it gives you less exposed. The desire for individual mobility is still there, so we believe our core manufacturers, I think that we will still have you know, we will have these vehicles that
we can drive on the road, we can do whatever we like in some areas but in
the cities we believe that will take some while to actually perform these
type of things. So, it will be the major roads around the cities where we first start to do autonomous driving. Let’s go to the next one. So, the Volvo vision of course “New Volvo’s shouldn’t crash” I mean that’s really important
that we set this target many years ago that in fact we shouldn’t have
any crashes. This is actually true in one of the car models already. Well not crashes, sorry, killed, I would say. This is really our problem today because our kids, I imagine you have kids. These guys and girls they really
love to do things with their friends on the net using Facebook or any other of these social applications. My girl, she’s not fun to, well she just took
her driver’s license. She had real problems because you know she had a
problem of concentrating on the road. Because when something happened, she
wanted to immediately react like she usually does and that is really a problem for the youngsters. So, we need to help them. So anyway, the main benefits
is Safety. Environment. Traffic flow of course. Lower and more effective
infrastructure investments, that’s really important for many of the new
cities because if we apply autonomous vehicle, what we get is that we could reduce or actually increase the traffic flow with about 50% and that’s major thing for the infrastructure departments of the cities, it’s very important. And of course, use of time. So anyway, we started this Drive Me Project a
long time ago, actually 2013. We are now in 2016 and our target is to have
the cars on the roads by 2017. As you can see, we have also the traffic
authority is with us in this project and that’s pretty important because
we are breaking some new ground and we need some hands to hold because we
are in some cases in a legal, well we have some legal problems to some extent. And we don’t know even at this hour, we don’t know really to what extent we can make the vehicle operating in those conditions where we allow
it to operate autonomously without having the supervision of the driver. So, this is I mean, we are working on this. So, let’s look at this short
movie. [Plays Movie]. Sorry, you can turn off the video now. I can do that, sorry. This is quite long actually.
It’s available on YouTube. Anyway. So what, this was kind of the mission from the start of the project and what we
have done with was looking into where are the use cases? And the use cases
is usually when you go to work and back from work of course. But you sit
in these long queues, traffic is not very fast. So, we decided on having a
target of launching 100 autonomous vehicles on the road in Gothenburg, in this ring road as you can see on this
map here. Where actually people is moving a lot, but to work and back
from work. And this road will have no special things, it would not be really adapted but we would have a very very
good map of this road that would be continuously updated, that’s very important for these cars. Of course, the cars
need to be able to move even without a map but the map is very crucial for
autonomous driving. There’s a lot of companies that are going into this
area right now, we can see that. The real importance of this product is that we intend to use normal customers drivers like you and me to actually use these cars. We also will have a lot
of I mean it’s a research project so the important part is to find out what drivers are doing in the car but also
externally. Because you know how are the other traffic behaving when they have an autonomous vehicle by the side? This is really unknown to us right now. And a little bit importance of this area or these roads is that it’s roads that have two lanes. And they
also have a septic guardrail, so it is kind of a road that is not really you know you don’t have pedestrians. You don’t have all these very very hard
things to take care of. So, it is a very simplified way of actually start
executing autonomous driving and understands what’s happening. And the city or actually the speed of these roads
is between 70 and 90 km/h and the target for our vehicles is to go quicker than 70 km/h but this is something
that is really hard to do and I will like talk a little bit about that later in the technical part. So the research areas we have is of course the societal and economic impact because the states or the road authorities wants
to know, you know what can we gain by these autonomous vehicles. That’s
also the, well of course that’s part of the infrastructure effect but it’s
also important to understand what are the suitable traffic situations. I mean the [inaudible 12:24] was to be you
know if we have snow and everything, how to deal with this. I mean we will probably have some kind of center that is so to say, “Okay now it’s okay to
go, now it’s not okay to go on the road” as well autonomous function. But also the customer expectations is of
course very important for us. I mean, this will be of course free to read about I mean the project reports would be available so to say for everyone.
The legal expects is of course very, it’s actually quite troublesome because legal part is moving quite slow but
as you have heard, our president say that we will actually, when we release a completely autonomous vehicle where
drivers does not need to make a surveillance then we actually as Volvo or as Car manufactures in general, we’ll
have to take on the responsibility. It’s like the product you know the product sales and that’s why if it was a crash or to say then it must be
like this. And that’s really the case already, now right? That if you have a product and you get in a situation, then it’s up to the manufacturer of the product to deal with it. So anyway,
the technical challenges is really really tough. It’s important to have something in all situations so we need to take care of all types of different objects on the roads and so on. You
need to have partly redundant control units. You may not need to be fail operational but you need to be able to detect that you have a failure and take control of it in a safe manner. This
also means that you need to have redundant brake system. You need to have redundant steering. You need to have
redundant signaling paths between the different control units. And you need
to have trusted power distribution where that means that you, well if one power fails down the power should
work. And also, some kind of safety critical HMI that works in collaboration with the driver in a safe way.
You shouldn’t be able to do what you know think
that the car is autonomous while it’s not autonomous, that’s really one safety critical part that we have to solve. And actually, you heard also
that we are cooperating with Uber to actually share some of the costs of these platforms. But actually, Uber is not using our technology I mean they
have their own technology for the sensing and everything that is clear.
This ASIL, Functional Safety Goals, I don’t know if you heard about this
but it’s something that is used today in mid-sensors and in all types of
sensors for autonomous vehicles, it’s becoming very important. Already within the active safety systems, these, the ASIL levels is important. And
normally, well when the driver is in control you have ASIL level B. But if you go up to full autonomous driving then
you need to have ASIL level D which is basically, well you need to be able to take care of any type of error
basically in the control. And if we look on the services because this is a sensor conference. We have a lot of different objects that could be on the road. I mean, as we know we have elks
that’s a problem. You could have you know we shift the wheels like you do
maybe here in Belgium also but those actually lay on the road once in a while or you could have people you know
changing tires and walking out here on the road. You need to take care of all these because, I mean if the driver isn’t looking ahead, then of course the vehicle needs to look ahead. Ah, there it is. The Sensor Fusion we talked, I have heard that a little bit in some other talk so it’s really important. Sensor Fusion means that well basically that we take the sensors from different kinds and we fuse them together and we get a very high precision
on the output. And we can understand exactly for instance where your able vehicle is in comparison to everyone else. And also, especially in regards
to the environment because the environment as the road ahead, it’s very important for an autonomous vehicle to always know this is exactly where I am. And you can think about this as a problem because if you have a GPS which is you know basically what you will have in your mobile phones. If you look into your GPS and you started off and
you fine usually that you are quite well you know, you are with in maybe
10 meters of precision but sometimes like in commerce you don’t even have a GPS right? So, you need to keep track
exactly where you are because if you don’t know where you are then you
don’t know if you can change the lane for instance to avoid an accident or to avoid something that is on the road or another car that is blocking the road for a while. So it’s really important to have a very good positioning
system. On Road Obstacle Detection, I’ve already talked about that, we
haven’t for seen everything yet. I mean we have a long list of different
things that we need to take care of but how to take care some of these we
haven’t solved yet. So, it’s in a bit lucky that you know within this research project we have still limited amount
of cars that we will start to use these cars under surveillance. That’s important because we have to log everything that is happening, we need to re-use,
well actually look on that data, recimilate it and make changes in the control algorithm because we will find things all the time. So anyway, the Sensor Technology, what we have in this car, we have a number of forward looking cameras. We have surround vision cameras that also have object detection and of course the forward-looking
camera also have object detection. We have nine cameras us all together looking outside of the vehicle. We also have a forward-looking radar. We have
four corner radars. We have forward looking laser scanners and a high
definition map. I’ve already talked a little bit about the importance of these. That’s also of course the importance to have cloud connection because these cars can communicate to each other or
actually to the cloud and then the cloud can tell the other vehicles that here is a problem on the road, let’s
be cautious or maybe don’t even drive their autonomously. And there’s
also driver monitoring because it’s important for both for the project to
understand what the drivers are doing but also for the control system. You know, is the car, maybe the car is not in full autonomous mode but in off autonomous mode than it’s good to know what the drive is doing because that gives a hint on the time to take over. So if you look into the Sensor and
Control Related Components, we have basically the vehicle control unit here, Including also Sensor Fusion part. There’s also an autonomous drive ECU
which is actually in these cars it’s IPX two unit which is a very powerful. You might say AD computer for a car
and it’s very useful for performing research task like this because it’s a tremendously powerful computer. Through these two that are connected through actually easy net. There’s also a cloud connection and there’s an HD Map which we have actually moved into the
AD ECU now, so we are actually you know already trying to bring down the
number of units in these cars to make it more streamlined. And we have the
Sensors that I talked about before, it’s this [inaudible 22:09] it’s not looking like this anymore but similar. There’s the lighter, it’s actually
the same lighter on the Audi S8 that is coming next year. There’s surround view cameras. There’s a driving monitoring. There’s the radar. Surround view radars. And there’s Ultrasonic sensors. We’re actually more ultrasonic sensors than on a normal car so we have also
ultrasonic sensors along the side of the car you know in order to detect for instance, we cannot move the [inaudible 22:47] or something like that. Clip the car. Let’s look on this
more movie. Can we have some sound? Okay, nobody’s here? Can you have with the sound please. Thank you. Yeah so this was. Actually, I mean if you look into this film you see a lot of about sensors and everything and we are just putting these together right now I mean it’s in approximately half a year, we will have films that actually shows exactly what’s happening on the road
in these situations. So that’s pretty interesting. Let’s hear. What is really interesting now is that the first vehicle has come out of our factory
and it looks like this. Number one, what’s really interesting is that,
as you can see it looks exactly like the normal vehicle. The integration of the lidar and sensors here at the front is basically invincible. We have
invincible installation of radars in all areas of the car. The only thing that is not really invincible is these reference cameras because they are
used to monitor the drivers and we actually have seven of these reference cameras to look on what the driver is
doing during the time he’s commuting. This I find quite nice and that is
the installation of the side camera looking real to understand whether you have traffic coming from the rear in
your lane. And here you can see the camera, what is this? This is actually cleaning system. So, we have managed
to also package also in this area we have managed to package also a cleaning system because this is really important for these types of cars. And here
you can see this box basically and that’s the computing part of this car. It’s on the surface of the rear of what you say? On the car, sorry. So, in this you have logging equipment. You
also have these extremely powerful [inaudible 28:35] and some other things. But if you actually go to get the
Borg and find these types of vehicles you will see something like this. And
what is this? Well it is actually a driving vehicle but it has under development you might say but it has this huge Lidar that is rotating on the top. And we use this as, this is actually
the reference sensing system that we have and we compare continuously.
We don’t use this for navigation but we use it for understanding errors within the other sensors. So, it’s really pretty important component when we
develop these cars. The Human Machine Interface is very important also because as i said, the human machine interface brings you I mean as a driver you
need to understand whether you’re in control or the vehicle is in control. And it was even like this that our first decision was that the driver would not be able to take over unless the car said it was okay to take over. Now
we see that there are some German regulations may be coming that actually demands that the driver needs to be
able to take over in anytime. So, we don’t know when really when we end up
here but that’s a very important part on these vehicles in order to start auto pilot you need to press these two buttons or actually do like this for
quite some time I think it’s two seconds and then you actually start
up the autonomous part. So, this is really the last slide, you know the journey isn’t over, that’s really what we say. I think that we in the whole industry is eagerly waiting to understand what this technology brings to us and
also not only the industry but also all our customers within the industry. And we can for instance see today that we need to have longer range sensors in order to go with a higher speed
than 70 km/h or 90 km/h to actually monitor the road ahead and understand whether you have an object on the road or not. Because if I mean, there’s a tremendous amount of work that is in front of us as an industry I believe.
So, I really look forward to the next coming years both in development
understanding what’s happening here. Looking on all the other initiatives
that we see on the market, different car manufacturers pursuing different things. And you know, in 2025 we may have this car that is actually doing what we do here. I mean people talk about 2021. I think that we can have that, we can have a certain amount of autonomy on certain areas in the world but
not everywhere. 2025 I believe we can come much closer to where we need to
be so to say to be able to bring this technology to most of our customers
on many roads. So, it’s an exciting future thank you very much. Thank you.