The new human-machine team.
MR GERN, EXPERT KNOWLEDGE ABOUT AUTONOMOUS DRIVING IS SPREADING RAPIDLY AROUND THE WORLD. HOW ARE YOU AND YOUR COLLEAGUES KEEPING PACE?
Digitalisation, autonomous driving and connectivity have a decisive influence on our future vehicle concepts. Daimler therefore maintains a potent global research and development network with around 16,000 staff members. Mercedes-Benz Research & Development North America has been in existence for over two decades in Silicon Valley. With a team of around 20, we were the very first OEM to set up a base there. We now have over 240 employees. We research the latest developments in the fields of autonomous driving, advanced user experience design, machine learning and business innovation. My team has been out on the roads of California since 2014. We were one of the first manufacturers to be granted a licence to trial self-driving vehicles in California. Yet our focus is on the autonomous driving of not just tomorrow, but also the day after tomorrow. For example, one of the main ideas behind our future vision of the self-driving vehicle, the F 015 Luxury in Motion, is a continuous exchange of information between vehicle, passengers and outside world. The passengers can interact intuitively with the vehicle using gestures, eye-tracking or touch. We know that, in urban traffic of the future, the space will be shared between humans and machines. Therefore, the F 015 is able, among other things, to communicate audibly and visually with its environment and, for example, to project a zebra crossing onto the road. These are all ideas that are thought up here in Sunnyvale.
WHAT IS MERCEDES-BENZ’ APPROACH TO AUTONOMOUS DRIVING?
Mercedes-Benz is adopting a twin-track approach to the development of autonomous driving. With the further development of existing driver assistance systems, such as Active Distance Assist DISTRONIC, which is capable of semi-autonomously following the vehicle in front in a traffic jam, we are adopting the so-called evolutionary approach. This will be gradually further developed. The second approach is the development of fully self-driving vehicles. Let us imagine, for example, that, one day, you can leave your vehicle at the entrance to a multi-storey car park and the vehicle then goes off and parks itself. Of course, it would also “pick you up” when you wish to continue your journey. Or imagine a car2go that makes its own way to you when you need one. We are also working on this revolutionary approach.
WHAT ARE THE NEXT STEPS?
The next big step will be the one from partially automated Level 2 to highly automated Level 3, mainly in the motorway scenario, in which there are neither traffic lights nor pedestrians or cyclists and in which the carriageways are separated by crash barriers. Almost all vehicle manufacturers are currently working in this direction. This requires further advances in terms of sensors. LiDAR is extremely important in this regard. It is excellently suited for localisation and object recognition. The goal is the intelligent combination of the various sensor principles. Connectivity is in need of further optimisation, and the sensors themselves also have more potential. For instance, sensors are still highly dependent on the weather, and there is also room for improvement on the cost and design fronts.
APART FROM THAT, AUTOMOTIVE MANUFACTURERS ARE INCREASINGLY EVOLVING INTO MOBILITY SERVICE PROVIDERS.
That’s right. This is once again the revolutionary approach. Here, the focus is on robotaxis as well as on autonomous carsharing vehicles. Mobility is a service that needs to be as low-cost and convenient as possible. This approach will presumably also decisively change the face of local public transport. For the end customer who is no longer the vehicle owner, the looks of the vehicle will then play a minor role.
CAN AN AUTOMOTIVE MANUFACTURER STILL MAKE MONEY FROM THIS?
The costs of such a system are dependent on the business case. A LiDAR system costing several thousand euros could pay off here. This kind of mobility is especially attractive in large cities where parking space is expensive and limited. However, the scenarios are diverse and highly complex. The revolutionary approach is being adopted by some traditional vehicle manufacturers as well as by competitors from the IT industry, among whom there is a very clear shift of emphasis from the mechanicals and the hardware to the software. At the same time, it is very important to have an understanding of the vehicle as a system as well as of issues such as safety, ruggedness, etc. For Daimler with car2go as the world’s largest carsharing company, this field is a highly attractive one.
' The focus is on robotaxis as well as autonomous carsharing vehicles. '
Axel Gern, Head of Autonomous Driving, MBRDNA
SO DAIMLER IS SIMULTANEOUSLY GOING DOWN BOTH THE EVOLUTIONARY AND THE REVOLUTIONARY PATHS.
The company has realised that it needs to be successful in both areas. In Silicon Valley, our exclusive focus is on realising the disruptive approach, particularly in the inner-city area. The teams in Germany are responsible for the sensors, actuators and system integration in the vehicle. In Sunnyvale, we are working on the “intelligence of the system”, i. e. on sensor fusion, situation recognition and vehicle manoeuvering.
WHAT IS SILICON VALLEY’S CONTRIBUTION?
Silicon Valley is a special place where innovation, inventiveness, boldness and capital come together. Here, the training in areas such as Artificial Intelligence and robotics is excellent. Since September of this year, therefore, we have been cooperating with the Udacity online university on the development and structure of a nano degree programme. With the research subjects of deep learning, computer vision, robotics, position determination and sensor fusion, the programme prepares the students for a job as a Self-Driving Car Engineer. Our goal is to train good people to work in the field of autonomous driving.
' New technologies open the way to a host of additional possibilities for interaction between vehicle, passenger and environment. '
HOW FAR HAS SITUATION ANALYSIS COME? A FAMOUS EXAMPLE IS THE OLD LADY AT THE ROUNDABOUT TOWARDS THE END OF THE BERTHA BENZ JOURNEY WHO WANTED TO WAVE THROUGH THE SELF-DRIVING S-CLASS, A GESTURE THE VEHICLE WAS UNABLE TO INTERPRET.
As far as situation analysis and assessment for inner-city traffic is concerned, it is vital to recognise and appropriately respond to all possible situations. In some situations, such as the previously mentioned example, interaction with other road users is very important, as we have demonstrated with the F 015. New technologies and forms of communication open the way to a whole series of additional possibilities for interaction between vehicle, passenger and environment.
The goal is to shape the relationship between human and car in such a way that the autonomous vehicle carries out those tasks the human is unwilling or unable to perform, such as permanent 360-degree all-round vision. The human, on the other hand, decides the speed at which the car should move or which is the more scenic route. This leads to the development of a new kind of “team relationship” between vehicle and driver that makes the human’s life easier. And, of course, detail-rich map systems form the basis for all self-driving vehicles. To this extent, our participation in the largest map-maker HERE is strategically very important.
WHAT ROLE IS PLAYED BY THE MERCEDES-BENZ DRIVING SIMULATOR?
In late 2010, Mercedes-Benz inaugurated the world’s most advanced “moving base” driving simulator in Sindelfingen. With its 360° screen, high-speed electric drive and 12-metre-long rail for longitudinal and transverse motion, it is the most powerful moving simulator in the automotive industry. It makes it possible to realistically simulate highly dynamic driving manoeuvres, such as changes of lane, and thus to intensively research the behaviour of driver and vehicle in road traffic. It is also safely possible to investigate the interaction between the autonomous system and the driver, e. g. when, in critical situations or on leaving the motorway, the task of driving reverts from the vehicle to the driver.