Man and Machine – Automated Driving: Infrastructure and Investments
16/09/2015
The infrastructure is an issue of crucial importance for a functioning environment permitting automated and autonomous driving. What is the impact of infrastructure quality and of traffic density? Who will contribute the necessary investments?
First things first: Definition of infrastructure
An infrastructure in the context of autonomous driving refers to a dense network of transmitters and receivers communicating with one another. In addition to the active road users, passive peripheral equipment, such as traffic lights or traffic signs, are all “interconnected” to create an infrastructure with the highest density possible. Tracking is done via GPS and communication via cloud connectivity; both must be guaranteed at all times. The density and quality of the necessary traffic data, also in the sense of real-time availability, depend upon the peripheral equipment. One thing is clear: the higher the quality and density of information, the greater the validity and above all the greater the speed of independent decision making by the
automated driving system.
In the future it should become possible to detect neurological signals from the driver and to feed them into the system and thus to supplement the infrastructure.Status quo
So far, the infrastructure for automated driving is rudimentary in design. At present, road signs and traffic lights can only be detected by vehicles using integrated stereo cameras but no information can be transmitted back or forth due to the lack of transmitters and receivers. With a positioning inaccuracy of GPS systems of up to two or three meters, this represents a no-go criterion for megacities like Shanghai, where inner-city multiple-level highways are part of the cityscape. Just imagine the fatal consequences if you were driving on the top level of an overlapping highway and your GPS believed you were on the bottom level.
Today we try to solve this problem by combining data from different sources. However, the vast amount of data, redundancies, and different bus systems make real-time allocation of the vehicle difficult or very expensive.
The goal of OEMs and leading first tier suppliers should be to create a modular on-board sensor and ECU infrastructure that is applicable to all vehicle and customer categories. That alone, however, will not solve the problem of the lacking infrastructure. Initial small steps have been taken by the governments of the leading automotive nations and model regions. Yet these measures fall far short of what is required to create the necessary infrastructure, at least in metropolitan areas. USA
What Silicon Valley is for the computer and software industry, Michigan is for automated driving. The center is the perfectly linked MCity in Ann Arbor. Here, over a distance of nearly ten kilometers, real traffic conditions including failed traffic lights, potholes, construction sites, traffic jams etc. and even "reckless" pedestrians are simulated by students from the University of Michigan who volunteer in the service of science.
In this environment, leading OEMs and suppliers have been testing their latest developments in automated vehicles for almost a year now. More than one million test kilometers have been completed. The construction of this facility cost nearly USD 10 M; that is approx. USD 1 M per kilometer of newly “connected” infrastructure. The facility was funded primarily from donations and subsidies. What financing will look like on a large scale is still unclear today. Germany
The Federal Ministry of Transport has also started designing special test tracks for automated driving. A section of the A9 autobahn has already been authorized to be used as a test track. In contrast to Ann Arbor’s Mcity, autonomous vehicles drive here in ideal conditions on a well-maintained highway without potholes or unpredictable pedestrians.
The stated goal of OEMs and the Ministry of Transport is to create a “national test field” with a particular focus on Car2Car communication (i.e. the exchange between active motorists). The investment for new sections of the autobahn and a network for the “national test field” alone is estimated at approximately EUR 80 -100 M.
Needless to say, this project will only partly reduce the United States’ competitive edge because states such as Nevada already allow automated driving on all streets.
Pilot projects have already shown that the development of large areas with an appropriate infrastructure is an extremely costly undertaking. So what is a feasible scenario? Who will pay and how much? The bulk of the investment could be borne by the beneficiaries, namely the OEMs. The role of the government and its willingness to invest remain questionable.Outlook and overall investment
A number of concepts exist as to how automated driving can be accomplished in urban areas, for example by driving in dedicated lanes or by utilizing the above-mentioned connectivity. As an example: If one wished to equip the ten largest metropolitan areas in the United States for automated driving, the total cost of investment would be approximately USD 500 bn. For Europe, it would be roughly the same.Schlegel und Partner has analyzed investments from governments, OEMs and suppliers over recent years and concluded that, at most, one-third of the necessary investment would be forthcoming.
This is where new market players come into the game, especially Apple with their recent attention-grabbing Titan project, which pursues the development of an electric car. Even Google, Alibaba and Badoo seem to have discovered the lucrative automotive business for themselves. Although Schlegel und Partner does not anticipate that Apple will be active as a standalone OEM, the company will nevertheless play a key role in the market through the implementation of software, maps, and cloud connectivity, just like Google and the others.It is clear that an investment package can only be put together as a joint initiative involving ALL market participants (including the government) to ensure an extensive and sustainable infrastructure for autonomous driving. In our next newsletter, you will be able to read about the impact of Connectivity and Automated Driving on the drive train. © Schlegel und Partner 2015
An infrastructure in the context of autonomous driving refers to a dense network of transmitters and receivers communicating with one another. In addition to the active road users, passive peripheral equipment, such as traffic lights or traffic signs, are all “interconnected” to create an infrastructure with the highest density possible. Tracking is done via GPS and communication via cloud connectivity; both must be guaranteed at all times. The density and quality of the necessary traffic data, also in the sense of real-time availability, depend upon the peripheral equipment. One thing is clear: the higher the quality and density of information, the greater the validity and above all the greater the speed of independent decision making by the
automated driving system.
In the future it should become possible to detect neurological signals from the driver and to feed them into the system and thus to supplement the infrastructure.Status quo
So far, the infrastructure for automated driving is rudimentary in design. At present, road signs and traffic lights can only be detected by vehicles using integrated stereo cameras but no information can be transmitted back or forth due to the lack of transmitters and receivers. With a positioning inaccuracy of GPS systems of up to two or three meters, this represents a no-go criterion for megacities like Shanghai, where inner-city multiple-level highways are part of the cityscape. Just imagine the fatal consequences if you were driving on the top level of an overlapping highway and your GPS believed you were on the bottom level.
Today we try to solve this problem by combining data from different sources. However, the vast amount of data, redundancies, and different bus systems make real-time allocation of the vehicle difficult or very expensive.
The goal of OEMs and leading first tier suppliers should be to create a modular on-board sensor and ECU infrastructure that is applicable to all vehicle and customer categories. That alone, however, will not solve the problem of the lacking infrastructure. Initial small steps have been taken by the governments of the leading automotive nations and model regions. Yet these measures fall far short of what is required to create the necessary infrastructure, at least in metropolitan areas. USA
What Silicon Valley is for the computer and software industry, Michigan is for automated driving. The center is the perfectly linked MCity in Ann Arbor. Here, over a distance of nearly ten kilometers, real traffic conditions including failed traffic lights, potholes, construction sites, traffic jams etc. and even "reckless" pedestrians are simulated by students from the University of Michigan who volunteer in the service of science.
In this environment, leading OEMs and suppliers have been testing their latest developments in automated vehicles for almost a year now. More than one million test kilometers have been completed. The construction of this facility cost nearly USD 10 M; that is approx. USD 1 M per kilometer of newly “connected” infrastructure. The facility was funded primarily from donations and subsidies. What financing will look like on a large scale is still unclear today. Germany
The Federal Ministry of Transport has also started designing special test tracks for automated driving. A section of the A9 autobahn has already been authorized to be used as a test track. In contrast to Ann Arbor’s Mcity, autonomous vehicles drive here in ideal conditions on a well-maintained highway without potholes or unpredictable pedestrians.
The stated goal of OEMs and the Ministry of Transport is to create a “national test field” with a particular focus on Car2Car communication (i.e. the exchange between active motorists). The investment for new sections of the autobahn and a network for the “national test field” alone is estimated at approximately EUR 80 -100 M.
Needless to say, this project will only partly reduce the United States’ competitive edge because states such as Nevada already allow automated driving on all streets.
Pilot projects have already shown that the development of large areas with an appropriate infrastructure is an extremely costly undertaking. So what is a feasible scenario? Who will pay and how much? The bulk of the investment could be borne by the beneficiaries, namely the OEMs. The role of the government and its willingness to invest remain questionable.Outlook and overall investment
A number of concepts exist as to how automated driving can be accomplished in urban areas, for example by driving in dedicated lanes or by utilizing the above-mentioned connectivity. As an example: If one wished to equip the ten largest metropolitan areas in the United States for automated driving, the total cost of investment would be approximately USD 500 bn. For Europe, it would be roughly the same.Schlegel und Partner has analyzed investments from governments, OEMs and suppliers over recent years and concluded that, at most, one-third of the necessary investment would be forthcoming.
This is where new market players come into the game, especially Apple with their recent attention-grabbing Titan project, which pursues the development of an electric car. Even Google, Alibaba and Badoo seem to have discovered the lucrative automotive business for themselves. Although Schlegel und Partner does not anticipate that Apple will be active as a standalone OEM, the company will nevertheless play a key role in the market through the implementation of software, maps, and cloud connectivity, just like Google and the others.It is clear that an investment package can only be put together as a joint initiative involving ALL market participants (including the government) to ensure an extensive and sustainable infrastructure for autonomous driving. In our next newsletter, you will be able to read about the impact of Connectivity and Automated Driving on the drive train. © Schlegel und Partner 2015