E-Mobility

What is slowing down the rollout of self-driving vehicles?

autonomous_vehicle_slowIn the general belief, Self-driving cars have always been seen as futuristic. Let’s just remember how it was represented in the 80s show, Knight Rider, picturing a superhero saving lives thanks to an autonomous vehicle called Kitt. Now you don’t have to be called Michael Knight, driverless technology does exist already but isn’t available yet for the general public. This is just a matter of time until we see cars without drivers; Safety, convenience, traffic congestion and cost reductions are all important benefits which would make driverless cars a success. That’s why we need to understand what is making us wait for these intelligent cars?

What is going on?

Thanks to the investment of innovative giants such as Google, Apple and naturally Tesla, lots of self-driving prototypes have already been tested successfully. Especially, Tesla which has already equipped production cars with his own autonomous solution. The system is however restricted for regulatory reasons and can only work with human supervision (Hands have to be kept in contact with the steering wheel to work). Google and Apple have a strategy to create self-driving software’s with the objective to resell licences to traditional carmakers. These mobility-as-a-service strategies could well be offered to a multitude of personal drivers, taxi (e.g Uber) or car sharing companies. Apple is clearly focusing on this approach instead of releasing cars. Self-driving parts provider such as Nvidia could be a major beneficiary of this change as this market could well weight billions in a near future.

What are the slowing down elements of autonomous vehicles?

As we generally agree that this technology will be part of our life, let’s analyse what could slow down the rollout of autonomous cars.

System/Software capabilities

The mobility platform for driverless vehicles is core and would be the main trigger for a successful self-driving car rollout. System reliability, performance and security are key!

Reliability

An autonomous car driving on public roads have to be 100% reliable. We are used to working with computer systems which do crash from time to time. Consequences of a system failure in an office are almost a casualty but could be dramatic on the road. Let’s simply imagine what could happen if the autonomous self-driving system would shut down while cruising on a highway with 120 km/h or even worse in a crowded city. The potentiality of a system breakdown cannot be tolerated. The solution would be to build up a sort of backup system which could be activated on emergencies.

Performance

Performance of the system shall include a tremendous amount of parameters to manage unpredictable outdoor conditions like ice, snow storm or to avoid wild animals like deer, elk, caribou or Kangaroos. That seems crazy for Londoner but any Canadian or Australian would surely agree on this! We shall also imagine that these cars would drive around people with unpredictable attitude. For example, in cities with the anticipation of kids, bicycles or addicted smartphones users.

Security

The system shall be of course highly encrypted to avoid any piracy or kidnapping to happen. Especially, if we consider that future cars could be sold without a steering wheel and controlled by a central unit.

Regulation

In simple terms, we do need a driving licence to drive on the road and so would a driverless car have too. Governments are going to have to legislate about the usage of driverless cars or more precisely a certified system standard to ensure a sufficient quality level. A step by step approach could be preferred with half automated cars driving autonomously on highways than on secondary roads to end up with a full rollout in cities.

Responsibilities (insurances)

As of today, you are responsible for the way you are driving, that makes you automatically the insurance holder of your car. If we suppose that the car would drive on its own, the insurance possessor shall be the car maker or self-driving system rather than the driver. The reason is that the so-called car OEM (Original Equipment Manufacturer) will become the driver and you the passenger.

Traditional car maker (OEM)

“Spass am Fahren”, Driving pleasure is the advertisement title of BMW which clearly target car “pilot”. Therefore, how can they be interested to push self-driving cars while they are selling the idea of driving lifestyle? Existing carmaker could become the big losers of this transition if they do not invest massively in this technology. Thankfully the capitalisation of these carmakers and R&D capabilities are mighty so we can believe that they do work on it… As they did for electric vehicles… This would mean that they are building up prototypes without really pushing for a proper rollout. This is understandable if we consider that autonomous vehicle would reduce the number of cars on the road. Why would your family need two cars when one could fit all jobs? Driving you to the train station or bringing kids to school, one car could manage it all. Furthermore, why would you need to own a car if you can use a sharing service and have it in demand? This is not a surprise if BMW and Daimler Benz are investing in car sharing companies and this is probably also the model to follow because car sales are going to fall down. They should also not underestimate the tech giants which are going to roll out autonomous cars. Tesla proved that it was possible to rollout competitive production e-cars, Google could well be the next autonomous vehicle leader. Key is to follow customer demands, don’t underestimate the technological development and make correct partnerships.

Drivers

Last but not least, YOU! Do we want to become a passenger after being used to be in control for decades? Do we want to own a taxi? These thoughts are clearly coming from today’s driver generation. Growing up dreaming about Ferrari F50, Lamborghini Diablo or Porsche 911 and considering a driving licence as an achievement, it’s definitely hard for this generation to imagine losing this “driving feeling”. Nevertheless, if we consider that road restrictions of all sorts are growing with increasing traffic, driving pleasure is also decreasing. Moreover, the concept of owning a car would also lose its interest with the ease of use of a sharing concept. This would most certainly start in cities due to parking limitations and slowly be moving to suburbs. Broader acceptance would just be a matter of time.

Let’s be optimistic

Once the self-driving technology will be proven 99.99% reliable and secure, it will just be a matter of years until we see cars without drivers on the road. Regulation, insurances would have to be adapted but road safety and service benefits will overcome any potential reluctance. Car maker along with new players are going to follow this trend because it simply makes sense.

Let’s now focus on what could speed up the rollout of electric cars, any thoughts?

7 Comments

  1. Andreas Schiffler

    Why does one even need to own a car? Considering a fleet of self-driving Ubers make self-ownership of a personal vehicle look like an antiquated concept, especially for city dwellers who often don’t even have a regular car anymore. The key technology and regulatory allowance is the capabilities to move a vehicle around without a driver. Good reliability and security in this space would come at a high price. The capital investment one needs to get a driver-less vehicle rolling with a minimum of “5 nines” AI will likely not be made by individuals.

    • Comment by post author

      Thanks for the comment! It is indeed obvious that the ownership concept of a car might almost disappear with self-driving vehicles. That’s why I believe that traditional carmakers do start invest in car sharing companies or even “taxi” services. The future of the privately owned car as a transportation system might be luxury goods. However, once some driveless platforms will be certified, the technology could be available at an affordable cost if delivered on a larger scale… I wouldn’t write off private cars but it will probably become a minority of cars on the road and sharing-services a majority, yes!

  2. Self driving is just one more modern gadget that does not change the fundamentals of transportation. Traffic congestion has to be solved once driving is made easier and cheaper via increased safety. Shifting the focus from what is under the dashboard to what is over the land is a more advanced way to consider what is coming next. Being an independent person, I want to maintain my independence so the Uber crowd can stay on their lawn and the rest of us will stay in ours. Advancing roads to be guideways and shifting fossil fuels or electric power grid to off-grid power from PV solar is what aerodynamic guideway vehicles can do as well as reducing the number of vehicle to vehicle interactions of today’s 2D roadways using a 3D guideway scheme. Paying for this with incredible savings in energy and lack of crashes with a pay as you go toll plan keeps the anti-crowd from getting any foothold. If no tax incentives are given and the guideway users pay for every last penny of the infrastructure then detractors will still complain as that is their nature but they won’t have a case. Keeping givernment out and being based on the individual beats the collective thinking that permeates the uber rich (example – Bill Gates) who desire to control us. Self-driving is a subset of advances coming soon to a garage by you.

    • Comment by post author

      Thanks for your detailed comment! I do however believe that self-driving is more than a gadget. This would reduce drastically the number of cars on the road and optimise also the speed and other parking space benefits. I do however share with you the fact that I want to stay independent too and being a simple client of a sharing company would not fit for me. That’s why I believe that sharing companies like Uber won’t be the only one option but they are going to be key in the future. For the rest, with the shift of fossil fuel to electric with the increased usage of solar technology, I can only share this feedback too.

  3. Reliability is the key for the introduction of self driving cars. Today we are far away from reliability. Engineering processes are not sufficient, the OEMs most of the times don’t even know what’s going on in the many ECUs from different suppliers. Instead of looking at the car from a reference model it is split into pieces and the OEMs try to manage the interfaces. This led to a big mess with 70.000.000 requirements with 15.000 attributes at VW. Even IBM’s Watson could not help manage this problem eventhough the requirements management system is IBM’s DOORS. When we take a look at the airplane technology we will find redundancy with often 3 ECUs doing the same thing with a choice of 2 of 3 in case of electronic control of the plane. Impossible in cars for cost reasons. I was director development in electronics for 7 years at a railway OEM and we also used redundancy for safety critical issues. Still you won’t find a lot of metros or trams riding without a driver eventhough it is bound to a given track without other vehicles or persons on it. And even this doesn’t work today if you take a look at the new Berlin-Munich ICE track and the french ETCS controlling the train. In addition there is a lack of engineers and a lack of experts for real time system development. We have to realize that there is a big difference between the development of smartphone electronics (unreliable, unsafe, unsecure) and a vehicle control system. And an even bigger difference in the development of PC-software and the real time system development (we do both). Different engineers, different worlds, different cultures. Everyone who ever developed a multimedia system with both internet and CAN connection knows what I mean….

    • Comment by post author

      Thanks for your very detailed and interesting comment! Yes, system performance and reliability is really the number 1 problematic of a potential rollout. You are absolutely right about the comparison with trains or trams which are generally not yet “self-driving”.. Apart of basic airport train shuttle.
      That’s why I rather do believe in a step by step approach with specific driverless road portion. Next to that, I wonder if Machine learning, AI technologies with building up of experience could be a solution for this. What do you think?

  4. Another comment to the guys arguing with shared economies. If this would be such a perfect idea we would all live in hotels instead of renting an appartment or building a house. I was working in Shanghai for three years and after some months I engaged a driver and rent an appartment because it just doesn’t work with taxis and hotel rooms in real. Much too much effort, time and stress to manage that. Shanghai is the city with the most taxis but if it starts raining you will never get one. The same will happen with automomous taxis. They will start to optimize the number of vehicles so the number will decrease in total and you won’t get one when you need it the most.

Leave a Reply

*

code