Chapter 1
Introduction
Automated and autonomous driving will change our lives substantially! It will affect every aspect of our daily habits and possibly lead to greater changes than space travel, nuclear warfare, or the rise of the internet—at least on a personal, everyday life basis. It will lead to an increase of safety, comfort, and efficiency in many ways. The role of vehicles will change dramatically for all of us. On the other hand, we will finally face ethical, political, and technical challenges beyond today’s imagination.
Let’s take a look at the businessman—the traveling salesman for whom the vehicle is a professional tool to get in touch with customers and partners. Instead of wasting precious time behind the steering wheel of his vehicle and facing the risk of personal damage through accidents, the vehicle will take over the job. It will find its way on its own, and it will interact with the surrounding ecosystem and other vehicles in order to make the journey as comfortable and safe as possible. The traveling time is no longer lost—the businessman will be able to work in the vehicle and even meet and negotiate with business partners while being on the road. In a nutshell, the future is bright!
The perspective of the family man may be similar. His vehicle is a tool to get his kids to school, to go for daily shopping at the local markets, to drive to work and back home, and finally, to bridge long distances to spend the summer vacation in a remote location. An automated vehicle may take over the role of some non-existent or unaffordable private servant who is always at the service of his or her employer and takes the members of the family wherever it is required, but in a way, that is less dangerous, more efficient, and less costly. Finally, the kids will no longer face the dangers of a drunken or an inattentive driver. We believe that the family man will appreciate the benefits of the upcoming technology.
The driving enthusiast, on the other hand, may see things differently. What is the point of buying a $300K sports vehicle if the computer takes over control and operates the vehicle based on principles of being reasonable, cost efficient and “boring”? It defies the purpose! One may argue that computer control may not affect the world of a vehicle owner insisting on drive himself, but will he still be on the same road that is populated by vehicles with V2X networking and connectivity for which a distributed digital consciousness is dealing with all aspects of control and decision making? In the extreme case, the human being behind the steering wheel will be reduced to being an unreliable and unpredictable element in an otherwise “perfect” chain. The world of the driving enthusiast may no longer be the same once technology has taken over.
Critics of technology will obviously focus on the changes and potential dangers that will inevitably come with the introduction of automated and interconnected vehicles. Indeed, the number of open ethical and political questions is tremendous. Would we really want artificial intelligence to invade our lives at the anticipated extent? We may indeed place life-or-death decisions into the hands of a machine that may decide in microseconds who will be injured or who will die in an unavoidable accident situation. Will it be the driver, or the child who has unexpectedly run into the street? Will the computer make better decisions than the human being behind the steering wheel? In the end, at least, there is no one left to blame except the designer of the decision-making intelligence in the vehicle. All of us will still need to learn tough lessons in this space.
For technology entrepreneurs and corporations, finally, the business potential behind automated and connected driving is endless. It is certain that a move toward this technology will create massive employment for highly qualified professionals of information technology and data analytics.
The city is becoming a hostile place for the vehicle industry. Between traffic, environmental legislation, and parking challenges, many cities seem to have turned against the vehicle. Some heavily burdened municipalities worldwide have run out of alternatives to lower an excessive dust and nitrogen oxide load; for instance, Germany’s top cities, Stuttgart and Munich have considered a citywide driving ban for diesel vehicles in their cities. How can we adapt to regain cities’ favor, capture the interest of the urban consumer, and be a key part of the mobility mix? How do we embed mobility services like Uber, MyTaxi and Lyft that complement existing transportation options like the tram, railway, and bike?
Drivers and passengers increasingly presume access in their vehicles to connected services such as weather apps, transportation and travel information in the same way they can access these services with their smartphones. Real-time data delivery of web, radio, or video, and access to multimedia content, gaming, and social networks is already implemented or is frequently requested. As vehicle users continue to request remote services, usage data, and automated vehicle function support, the automotive industry has responded by investing heavily in a rapidly developing market of connected vehicles.
But we also perceive extremely divergent trends in how vehicles are used and viewed by their drivers and passengers. There is sometimes boredom for people who are riding in the vehicle, while the drivers and passengers may be extremely occupied and overwhelmed by complex situations, smartphone usage, and higher levels of automation inside and outside the vehicle. Both situations can very often become life threatening. These days, we are asked or ask ourselves why we want vehicles that drive themselves.
For most of us, learning to drive was a rite of passage to becoming an adult. This rite may soon change with networking vehicles and autonomous driving. The concept of not dealing with the rush-hour traffic does sound appealing, and whether we like it or not, autonomous vehicles are coming. Supported by computing and communications, autonomous vehicles offer opportunities to make our roads safer, improve fuel efficiency, and give us the added bonus of a more relaxing driving experience. That’s how it looks today from the driver and passenger’s seats.
The view from vehicle manufacturers, computing, communications, and other vehicle ecosystem-entering stakeholders might be quite different. For example, for computing stakeholders such as data center operators, the pervasive use of autonomous vehicles creates unprecedented challenges. We do see in the research, development, and test phases of autonomous vehicles that the amount of data coming from self-driving vehicles is huge. The autonomous vehicle, driven about 90 minutes a day, generates about 4 terabytes (TB) of data. Some vehicle trials are generating over 1 petabyte (PB) of data per month. By 2020, a single autonomous vehicle might produce 4 TB of data during 1 to 2 hours of driving. For comparison, a single drone flight captures up to 50 GB of data, and a fleet of 500 drones creating maps can record about 150 TB of data. Data centers are already one of the crucial building blocks for gathering, dealing, and analyzing that data for autonomous vehicle systems. The transport of vehicle data to the data center is a challenge in and of itself. To scale out support for, let’s say, 10 million vehicles by 2020, data centers will have to deliver extraordinary performance in computing, networking, and storage.
Ultimately, to pave the way for the widespread use of autonomous vehicles, all vehicle ecosystem stakeholders need to be uniquely positioned. This is demonstrated, for example, by the many European research projects for vehicle networking, connectivity, and communications as part of European road transport research advisory council’s “Vision 2050.” Among these projects are CARTRE, SCOUT, COMPANION, AMIDST, AUTOPILOT, ENABLE-S3, AutoNet, and CarNet. The European road transport research advisory council calls connectivity and networking key challenges for autonomous driving, asking for solutions to balance the fast evolution of and request for connectivity with the slower pace of vehicle and infrastructure development, and to deal with the growing demand for communications, bandwidth, and data. The networking technologies, which span from vehicle to cloud, will require many distinct systems and capabilities to work together seamlessly to deliver tomorrow’s self-driving vehicles. These are thrilling times for the vehicle ecosystem heading toward automated and autonomous driving solutions. But these times are by no means easy ones for the ecosystem stakeholders like vehicle manufacturers and their suppliers. It’s a cutthroat challenge for any vehicle manufacturer or solutions provider today to determine which features the public will want, need, accept, and buy with the vehicle. Aggravating that challenge, it becomes even more demanding to estimate the types of advanced driving solutions that vehicle drivers and passengers are striving for—technical edge, safety, the joy of driving, or best comfort.
Obviously, the vision outlined in the examples above will not occur over night. Rather, a gradual, step-by-step introduction of the technology is going to be applied. We currently see first elements in today’s high-end vehicles, such as driver assistance systems, (partly) autonomous driving within certain limitations, etc. These features will further evolve until vehicles will even be able to fully operate without any driver.
According to the U.S. Society of Automotive Engineers (SAE) and the German Association of the Automotive Industry (VDA), six levels with increasing degree of automation are defined for automated driving (SAE Standard J3016, January 2014) as illustrated in Figure 1.1. Starting from no automation (Level 0), there is driver assistance (Level 1), partial automation (Level 2), conditional automation (Level 3), high automation (Level 4), and finally, full automation (Level 5). At Levels 0 to 2, the human driver monitors the driving environment, while at Levels 3 to 5, an automated driving system monitors the driving environment. Increasing the level means raising the needs for communications both inside the vehicle and to the external world. The communications needs are strongly related to the degree of cooperation, such as cooperative sensing and manoeuvring. We use the SAE model throughout the book. Descriptions of the levels are included below in Figure 1.1 and Table 1.1.
Table 1.1: Levels of automated and autonomous vehicle driving according to Bundesanstalt für Straßenwesen (BASt), German
However, we believe that, aside from technology and business, there is a greater angle to the introduction of automated driving and networked vehicles that will be key— the gradual introduction of different parts of our road infrastructure. As a first step, we anticipate that autonomous driving will be made available on major highways linking the larger hubs of a given country. Such an environment indeed allows for a confined and reasonable level of deployment of roadside communications equipment and other required infrastructure, and is thus a reasonable business model. At the same time, the impact to the population will be considerable—once the driver has entered the highway, the artificial intelligence within the vehicle will take over, which is fed with information through a suitable wireless ecosystem. Long trips will become easier, and the problem of drivers falling asleep will no longer be relevant. Your vehicle will be able to take you all alone from Munich to Berlin, from Marseille to Paris, or from New York to San Francisco. The driver only needs to take care of the first miles to the highway, and the miles from the highway to the final destination.
Subsequently, the technology will be made available in areas outside of the major road infrastructure. First, major hubs such as large cities will be equipped, and perhaps eventually, smaller towns, villages, and countryside roads. It remains to be seen whether a full availability of automated driving will be available everywhere. While the deployment of the technology is straightforward along major roads, the full coverage of any residential area in any town and village may come at a cost-benefit level, which is unattractive—at least in the short- to mid-term. We will face a specific challenge during the deployment phase when some vehicles will rely on the new technology while others will not. We observe in particular that local drivers apply behavior on the road that is not written in a textbook or otherwise documented. An automated vehicle will need to adapt to such behavior. To provide an anecdote, one of the authors of this book recently experienced some local particularity in the beautiful south of Italy. On a two-lane road, some drivers honked their vehicle horns to forcibly create a third, imaginary middle lane. The vehicles on the left moved further to the left, the vehicles on the right moved further to the right, and a new middle lane was created, improving overall traffic efficiency. Who knows what other unwritten and undocumented ways of behaving on the road may exist across the world! It will be a clear challenge to the technology to fit into the existing ecosystem and to possibly adopt local habits. Therefore, we will further elaborate on the ethical, political, and technical challenges of the technology, and we will comment on the road ahead.
1.1V2X objectives
There are wireless-based vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) links (see Figure 1.2) to network or connect vehicles using local- and wide-area radio access network technologies. We use the term “networking” as an exchange of data between self-driving vehicle system components inside and outside of the vehicle. The term “connecting” refers to the setting up of communications links between these system components. The primary objectives behind V2V and V2I (referred to as vehicle-to-everything, or V2X) are to prevent accidents and save lives by alerting drivers to the hidden dangers that can’t be sensed by on board equipment. Other names sometimes used for V2X are “cooperative connected vehicles,” or “cooperative intelligent transportation systems” (CITS). By sharing data from any V2X-equipped vehicle within a specified radius, the driver can be alerted to the most common causes of accidents in time to take action. For example, for V2I, these include safety red light violations, curve speed warnings, weather alerts, and signals for work zone safety, bridge height, and pedestrian crossing. V2V safety applications and services are emergency brake light warnings, forward collision warnings, traffic alerts, emergency vehicle notifications, and road hazard detections. V2X convenience uses are eco-driving, parking information, truck platooning, speed harmonization, queue warning, and insurance pricing. And there's a lot more we need data services for— gathering behavioral data of the vehicles to improve them by applying machine learning, upgrading software and security features, or having vehicles exchange data with one another through wireless links.
V2X communications shouldn’t be mixed up with autonomous and automated driving vehicles, which is a secondary target for this technology. As we can see from many tests of autonomous and automated vehicles, the capabilities of self-driving vehicles could be extended optionally by implementing communications networks both in the immediate vicinity (for collision avoidance, for instance) and far away (for congestion management, for example). But by networking vehicles with the traffic infrastructure, providing additional data for the internal data processing, one could no longer regard the vehicle’s behavior or capabilities as autonomous. The vehicle is rather automated in this case. In other words, if vehicles are networked and cannot drive without the network, then they are not autonomous—they are purely remote controlled. A vehicle that couldn’t drive when the access to the networks went down or went out of coverage is going to be unsafe and unusable.
In principle, self-driving vehicles are possible without networking vehicles with everything, even for high and full automation. There have been test pilots of self-driving vehicles based only on on-board sensor actor systems, like the Audi A7 550-mile piloted drive from Silicon Valley to Las Vegas in 2015 (Audi Press release). This pilot used high-resolution digital maps, which were downloaded offline from a server. V2X is specified to provide vehicles, drivers, passengers, and other stakeholders with additional data that will be integrated with data from the many other vehicles’ on-board sensors and actors. V2X is an enabling technology that will make autonomous and automated driving vehicles much safer by helping them to extend their viewing and visibility range. Besides communications performance, reliability, security, and privacy are must-haves for V2X communications.
In this book, we start looking at the major ecosystem stakeholders for networked vehicles, and their basic usage scenarios and use cases identified so far, for specifying the requirements of research and development platforms being capable of up to Level 5 automated driving. As previously mentioned, Level 5 automated driving means, for example, running a fully automated robot taxi with an auto pilot, including self-navigation, collision avoidance, automated valet parking, highway chauffeuring, lane change management, steering, and throttle and brake actuation control. These research and development platform architectures consist of hardware architecture and system enabling software such as device drivers, computer vision libraries, mission and motion planning software, perception detection and localization software, and sensor and actuator software drivers. All these system components need to communicate internally and externally with each other reliably, safely, and securely via communications hardware and software. V2X communications has to provide the right performance for the various use cases to connect platforms comprised of components like cameras, LIDAR, and GNSS with computing and storage systems, for example, to accurately determine a vehicle’s location. Already, today’s advanced driver assistance systems (ADAS) like automatic cruise control (ACC), lane departure warning (LDW), and pedestrian detection (PD) form the backbone of tomorrow’s mobility. Vehicles communicate with each other and with infrastructure. Vehicle-to-vehicle (V2V) communications allows vehicles to exchange relevant information such as local traffic data (for example, nearby accidents) and their driving intentions. Vehicle-to-infrastructure (V2I) communications is used to optimize the road network usage and thereby helps to reduce environmental pollution.
1.2V2X and D2D networking and connectivity
The explosive growth in data traffic demand on wireless communications due to the popularity of networked and connected vehicles can be satisfied with several local- and wide-area network and access architectures. One of them is the exploitation of device-to-device communications (D2D) for connecting vehicles in close proximity within a common communications range. In D2D communications, vehicles communicate with each other without intermediate nodes. The inherent lower delay in communications that is required in some of the traffic safety use cases or for collision avoidance systems can be advantageous. By using D2D or V2V, respectively, rather than using V2I and relying on infrastructure, vehicle grouping—including group handover—can be done with minimal signalling from the network. Device-to-device communications has been introduced in 3GPP Release 12 as a side link communication, but only a few years later, the same standard has been optimized to vehicle communications requirements. Strict latency and communications requirements have been introduced. The D2D side link channel has been modified to meet those requirements and enable new use cases.
V2X communications is the transmission of data from a vehicle to any unit that may impact the vehicle, and vice versa. V2X communications consist of four types of communications: vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), vehicle-to-network (V2N), and vehicle-to-pedestrian (V2P). This data exchange can be used for safety, mobility, environmental, and convenience services and applications, including driving assistance, vehicle safety, speed adaptation and warning, emergency response, navigation, traffic operations and demand management, personal navigation, commercial fleet planning, and payment transactions. The basic component of a V2X system is the vehicle (V) and its connectivity to any other intelligent transportation system (ITS) component. V2X consist of transceivers located on vehicles, on the roadside infrastructure, in aftermarket devices, or within any other mobile devices like smartphones.
3GPP TR 22.885 defines roadside unit (RSU), vehicle-to-everything (V2X), vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), and vehicle-to-pedestrian (V2P). A RSU comprises an eNB or stationary UE transmitting or receiving data from another UE using a V2I application to support V2I services. V2X is a communications service where a UE transmits and receives V2V application data via 3GPP transport. V2I service is a specific V2X service between an UE and a RSU by means of a V2I application. V2I service is another specific V2X service between an UE and the communications network—in particular, a LTE network. V2P service is a specific V2X service between UEs using a V2P application and V2V service is the V2X service type where UEs communicate via a V2V application.
In this book, we use “networking vehicles with everything” as synonymous with V2X communications. Examples for V2X use cases are vehicle-to-vehicle (V2V) communications for platooning, convoying, and vehicle safety like collision avoidance; pedestrian-to-infrastructure (P2I) for traffic light control and warnings; and vehicle-to-infrastructure (V2I) for smart intersection control and phased traffic light, dynamic environmental zones, and real-time vehicle localization. Vehicle-to-everything (V2X) is any communications involving a vehicle as a source or destination of a message. Depending on the nature of the other communications endpoint, several special cases exist: vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I) (road infrastructure, which may or may not be co-located with cellular infrastructure), vehicle-to-network (V2N) (e.g., a backend or the internet), vehicle-to-pedestrian (V2P), and so on.
1.3Technical challenges for V2X
Networking and connecting vehicles internally and externally have been inspiring research and development topics for decades. The tremendous progress in computing, communications and sensors actors could enable fully networked vehicles and in particular automated and autonomous vehicles starting in big quantities in 2020. Sensor actor technology, massive and robust signal and data computing, and enormous advances in wireline and wireless communications technologies will make networking and connecting vehicles with everything happen. The autonomous and automated vehicle platform is based upon sensors and actors, computing, storage, communications, and software. We derive some major technical performance indicators of networking and connecting vehicles, taking into account the data sheets of state-of-the-art vehicle platform components and most challenging use cases.
The way forward to autonomous and automated driving with increasing automation starts from data sharing and raising, and ends with decision sharing. At level 1, the vehicle gains awareness with status data like, “I’m a vehicle” and responsiveness regarding its location and traveling direction. At level 2, sensor data like, “It’s raining here” or “I just passed a RSU” are added. At level 3, the vehicle acts cooperatively, requesting the driver to confirm data and system actions in use cases like lane merging or traffic jam assist. At level 4, the vehicles work in collaborative or cooperative ways in use cases like “Let’s platoon” or slot-based intersections. Finally, at level 5, the fully automated vehicle relies on massive sharing and fusion of sensor data and distributed computing.
1.3.1Sensors
The up-to-date technology for highly autonomous and automated vehicle driving in controlled and in real environments is quite advanced. Vehicles use state-of-the-art sensors like radar, LIDAR, ultrasound, GNSS, and video camera systems (Figure 1.3) jointly with high definition (HD) maps, which enable on board computing platforms to identify obstacles, relevant traffic signage, and appropriate navigation paths. Long-range radar is taken for adaptive cruise control (ACC), forward collision warning, and night vision, and LIDAR for object detection like pedestrians and collision avoidance. Video cameras are deployed for lane departure warning, traffic sign recognition, surround view and park assist, short- and medium-range radar for forward collision warning, pre-saving, back-up obstacle detection, stop-and-go, low speed ACC, cross traffic warning, blind spot detection, and rear collision warning. Ultrasound is sometimes used for park assist.
We view V2X communications as an enabler to enhance any sensor in the vehicle. While sensors like radar, LIDAR and video cameras are actively surveying the environment around the vehicle, V2X wireless communications with non-line-of-sight capability, is able to extend its range far beyond, to any range needed by communicating with other vehicles, infrastructures, and pedestrians. We look as well at other in-vehicle sensors, which play a major role to ensure the main purpose of the vehicle—to drive. Position sensors in the vehicle are used, for instance, to measure the angles of the engine throttle plate, the chassis height link bar, the fuel level via float arm, and the steering wheel angle. Vehicle pressure sensors are used to determine the pressure of the engine manifold absolute, ambient barometric, evaporative fuel system leak, brake fluid, chassis adaptive hydraulic suspension, air conditioner compressor, or the common-rail fuel injection. Temperature sensors are there for engine coolant, fuel, brake, and steering fluid levels. Mass airflow sensors measure steady state and transient mass flow of air into a vehicles’ engine. Torque sensors estimate the steering wheel torque for electric power steering (EPS), driveshaft (transmission-out) torque, and clutch shaft (engine-out) torque. Linear acceleration inertial sensors measure the vehicle stability and chassis adaptive suspension systems, vehicle frontal, side, and rollover crash sensing and engine knock detection. Angular rate (Gyro) inertial sensors are there for vehicle electronic stability control (ESC), active chassis suspension, rollover protection of side curtain airbags, and vehicle navigation systems. Chemical and gas composition sensors measure gas, oxygen, and oil characteristics. Comfort and convenience sensors estimate solar radiation, twilight, dimming of mirrors, fluid levels, rain detection, temperature, and noise. Driver, passenger, and security sensors are there to determine occupants’ weight, size, position, seated weight, seatbelt tension, buckle status, seat position, and intrusion detection.
Nowadays, in vehicle tests, long-distance radar, LIDAR, short-range radar, and video data get uploaded to servers and clouds for analysis. In commercial use, this enables the application of deep learning and KI for vehicles in an ongoing process now. The other above-mentioned sensor data might additionally to get stored offline for big data analytics and telemetry as a service.
1.3.2Computing
A vehicle today is a distributed computing system with—when all optional equipment is installed—50 to 150 electronic control units (ECU) and distributed storage up to several GBs which run vehicle control, navigation and telematics and infotainment and accomplish vehicle performance and behavior. ECUs are connected with several in-vehicle networks with high real-time reliability and safety requirements. There is an ongoing transformation of the computing architecture from distributed ECUs toward a more centralized and domain-oriented electronic control (CEC) system to reduce the large number of ECUs and, in particular, to save space in order to support autonomous and automated driving. The storage requirements are changing, as well, and are driven by HMI, IVI, and data recording.
Computer vision, cognitive computing, deep learning, and AI for autonomous and automated driving require massive on-board computing power for real-time decision making, since vehicles have to make critical decisions in real time independently from any cloud. Nevertheless, cloud-computing is also required for collective and deep learning data acquisition in vehicle tests as well as common use—therefore, vehicles are becoming part of a wider networked and connected ecosystem. Use cases are ADAS sensors, big data, and analytics applications in which vehicle sensors are used to create, maintain, and improve a data source for services and applications (for example, dynamic HD maps). Vehicles—in particular, next generation autonomous vehicles—effectively become mobile data centers. The vehicles themselves will generate and process massive amounts of data from on board sensors, but will also take in large quantities of data from the network, including ultra-high definition maps and near real-time information to help navigate and detect what’s coming around the next corner or to avoid upcoming traffic congestion.
Vehicle computing platforms evolve with telematics, transmission control units (TCU). We identify three major architecture concepts delivering sufficient processing power for autonomous and automated driving, telematics and IVI which are getting prepared for V2X communications as well.
1.3.3Networking
Today’s cellular communications systems are not designed to handle the massive capacity required to support millions of self-driving cars when they hit the road. The way toward autonomous and automated driving involves the evolution of current communications technologies, as well as the development of new ones such as 5G. Networking and connectivity are important components of the upcoming autonomous and automated vehicles ecosystem where secure, reliable V2X communications with low latency is mandatory. V2X technologies involve the use of several wireless technologies to achieve real-time two-way communications among vehicles (V2V), between vehicles and infrastructure (V2I), and pedestrians and vehicles (PDI). The convergence of sensors, computing, storage, and V2X networking and connectivity will stimulate autonomous and automated driving.
V2X networking and connectivity can be supported through different wireless technologies. V2V networking and connectivity is likely by means of dedicated short-range communications (DSRC) technology, developed with safety critical use cases in mind and based on the IEEE 802.11p Wi-Fi standard. IST-G5 and WAVE as of Table 1.2, specify dedicated short range communications (DSRC) for an ad-hoc 2way network in dedicated licensed spectrum bands with 7 channels. DSRC has got low latency below 50 milliseconds and a range up to 2 kilometers with data rates from 6 to 27 Mbps. It provides signed messages using a public key infrastructure (PKI).
Table 1.2: V2X networking and connectivity stacks as of IST-G5 and WAVE protocol stack
| IST-G5 (Europe) | WAVE (USA) | |
| Security IEEE | ||
| 1609.2 | C2C-CC | SAE J2945 |
| CAM, DENM, SPAT, MAP | SAE J2735 (BSM, SPAT, MAP, etc.) | |
| GeoNet / Decentralized congestion control | DSRC WAVE short message protocol | |
| TCP/IPv6 | TCP/IPv6 | |
| IEEE 802.11p MAC EU Profile | IEEE 802.11p MAC DSRC | |
| IEEE 802.11p PHY EU Profile | IEEE 802.11p PHY DSRC |
This is the path the United States’ National Highway Traffic Safety Administration (NHTSA) and other regulators and stakeholders are currently assessing. For instance, General Motors the largest vehicle manufacturer in the U.S. said in March 2017, that all of its new 2017 Cadillac CTS vehicles would come armed with DSRC based V2V networking and connectivity. And Volkswagen in Germany said in June 2017 that it will fit first VW models with pWLAN technology which is based on IEEE 802.11p in 2019.
At present, the vehicle industry’s safety stakeholders generally are not fully committed to whether cellular technologies can be used besides infotainment for safety critical ADAS and autonomous and automated driving functions. This is due to performance indicators like cellular network coverage, reliability and liability, security and the need for mobile operator subscriptions, and roaming agreements. Issues like these compete unfavorably against—LTE with DSRC technology, for instance—in avoiding collisions and other safety applications as of today.
That’s where next generation new radio communications technology comes in. It’s expected to deliver more capacity, ultra-low latency, faster speeds and vehicle-to-vehicle (V2V) connectivity for the era of autonomous vehicles. With the increase in mobile traffic created by making everything smart and connected—including vehicles—networks need to transform to handle the additional connections. By leveraging innovations from cloud data centers, the network infrastructure is becoming agile, software-defined and flexible to enable the efficient and ultra-low latency connections needed for autonomous driving. Autonomous vehicles will need to act on near-instantaneous updates from around the corner or down the road–there’s no time to send or receive data from server hundreds of kilometers away when your vehicle is platooning with 100 other vehicles at 100 km/h down the highway. With 5G, networks will deploy computing resources at the very edge of the network in cellular base stations and towers that will deliver road status updates to connected vehicles in milliseconds.
With the millimeter wave spectrum and advances in wireless and antenna technology, 5G is expected to deliver multi-gigabit speeds for mobile uses. The industry expects 5G speeds to eventually be capable of up to 10 gigabits per second, which is over 600 times faster than today’s fastest average LTE speeds in the U.S. Self-driving vehicles will opportunistically connect to 5G cells when available and needed, then seamlessly fall back to 4G LTE to maintain network connectivity. 5G aims to deliver multiple models of connectivity—including direct vehicle-to-vehicle connections, as well as vehicle-to-infrastructure or vehicle-to-network connectivity. This flexibility in 5G design anticipates the varied situations that self-driving vehicles will encounter in, around, and between cities.
We show that there are technical challenges on top of the communications link for networked vehicles in the dynamic system architecture through the reconfiguration of a vehicle communications system after delivery, the reuse of system and software components, and building blocks during development, tests, and trials. There are also challenges in the operation of vehicles and the traceability of requirements and proof of their realization in variable and dynamic systems.
If we consider necessary updates of hardware and software features for networked vehicles, harmful influences on other system components must be absolutely excluded. This applies not only to direct functional relationships but also particularly to the resource requirements and the real-time behavior of the networked vehicle system, as well as aspects that could compromise the functional safety. Particularly high requirements arise when functions of different criticality, such as in-vehicle infotainment functions and vehicle functions, are executed on the same hardware. For this purpose, technologies are required that allow close interlinking and sharing of resources between the mentioned system parts, but do not endanger the functional safety of the vehicle functions.
Certain use cases of V2X are very challenging when it comes to vehicle functional safety according to ISO 26262, which requires that if and when a vehicle misbehaves, it is not hazardous to any life. Vehicle networking and connectivity are going to require additional hardware and software features in terms of modular hardware and software (with embedded security, in particular) for autonomous and automated driving. In order to ensure that these features fulfil the ISO 26262 requirements, a common validation and certification procedure is a prerequisite—the communications between all involved vehicles, infrastructures, and pedestrian components and functions is required to be evaluated how V2X networking and connectivity reacts in the case of misbehavior.
We need to ensure the best use of V2X networking and connectivity, so it is necessary to take into account the potential system impacts which cannot be anticipated a priori for all use cases. This requires new solutions that allow monitoring, plausibility checks, and safety-related interventions during vehicle operation via networking. These solutions must ensure compliance with clearly defined safety criteria across functional and system boundaries. In addition, appropriate services and interfaces will be required to enable early identification of anomalies, which may require the permanent collection and centralized evaluation of diagnostic data in real-time.
The introduction of new and open interfaces to the networked vehicle, and the possibility of reloading and adapting software during runtime, create new threat scenarios regarding data security. Therefore, appropriate precautionary measures must be taken in terms of processes, architectures, protocols, but also services at hardware and software levels. In the short-term, this affects the unintended accesses to the nonvolatile memory. This can be solved with currently available evolved security mechanisms such as authorization of flash memory, detection of software manipulation, and the transmission of successful security patterns from computing and communications like firewalls or virus scanners. In the mid-term, the focus shall be on the threats at run-time, particularly through the communications interfaces created by networking and the reloading of software. There are proposals to use security standards (such as IEC 62443) with their specified automotive safety integrity levels (ASIL) for compliance to safety-critical V2X components.
An increasingly important aspect is the confidentiality of private data (privacy). For driver and passenger protection, basic services are required, which are exemplarily secure communications channels, efficient encryption, reliable anonymization, and pseudo-authentication. The privacy services shall be adapted to altered threat scenarios—for example, by updating cryptographic procedures. With regard to the prevention or detection of manipulations via communications interfaces or by reloaded software, services are required for the certification of software, the reliable authentication of accessing third parties possibly in conjunction with a trust center, intrusion detection, and for the quarantine and isolation of suspicious software units.
For more than ten years, cellular networking and connectivity in vehicles was seen in a first phase as a solution to calling for roadside or emergency assistance, as in the case of General Motors’ OnStar or Daimler’s mbrace® service. These services are still vital and are part of a broader set of upcoming safety services and applications. But in the last few years, there has been a second phase of vehicle networking that is driven by smartphones and vehicle infotainment. Applications like Google Maps, Twitter, and Facebook were integrated into in-vehicle infotainment (IVI) systems along with weather, video, and music streaming applications. Audi integrated vehicle-to-cloud-to-infrastructure (V2C2I) networking and connectivity to its Audi connect infotainment systems, enabling traffic light signal phase and timing (SPAT) application (currently only working in Las Vegas in the U.S.), which informs drivers of the phase status of an approaching traffic light. It went further by implementing Wi-Fi, Bluetooth, and NFC-enabled remote access to vehicle data, diagnostics features, door lock, and even remote engine start. The most current advances are speech recognition, parking reservations, and digital assistants delivered via connected smartphones. BMW, Daimler, and Volvo introduced vehicle-to-cloud-to-vehicle (V2C2V) networking and connectivity to enable communications between their vehicles in 2017. On the opposite end, neither Alphabet (Google) nor Tesla Motors—front-runners in automated driving—are currently leveraging wireless cellular or Wi-Fi technology for collision avoidance and automation.
BMW’s roadmap for deploying autonomous technology shows achieved SAE level 2 advanced driver assistance that still requires hands-on direction from a human driver at all times. Level 3 automation will be achieved with its iNext vehicle by 2021. BMW’s iNext is going to be technically capable of Level 4 or 5 autonomy by 2021 for development and tests, but not for production. BMW uses high-definition maps to extend vehicles’ reference data further than the range of sensors. Together with V2I communications, this allows a vehicle to navigate itself to a stop in a safe location if the driver is asleep or otherwise unable to immediately take control. Artificial intelligence is seen as another key technology for safe self-driving vehicles, but will need to overcome significant challenges associated with AI data processing and communications. BMW expects fully self-driving vehicles between 2020 and 2030.
When 5G technology approaches, it will transform wireless vehicle networking and connectivity. This transformation of wireless networking and connectivity will come to play a decisive role in many vehicle domains, including autonomous and automated driving use cases. It’s becoming evident that the vehicle is increasingly dependent on tight integration between networking and connectivity, sensors, actors, storage, and computing. V2X will be enhanced by the capabilities provided by 5G as another option to DSRC for low latency collision avoidance, V2V, V2P, and V2 networking and connectivity. The challenge is to make cellular networking and connectivity a safety and mission-critical technology with the prospect that general wireless technology (cellular, in particular) is becoming an essential part of autonomous and automated driving vehicle solutions.
1.4Society, ethics and politics
We are clear on one thing—the vehicle is the most favored toy. Many of us are distrustful of leaving our personal rights and control over the road to the vehicle. For us, the vehicle is not just a means of transportation from A to B. No—our vehicle means, among other things, passion, status, freedom, fun, frustration, or even individuality. Who doesn’t enjoy being able to speed up his or her own style and be the master of the situation? Will this be taken away from us with a networked, autonomous vehicle? If it is only a matter of pure movement, public transportation is available to most of us. So, how will the world work with automated vehicles in the future? Can an autonomous vehicle honk, even if there is no immediate danger? And how will a vehicle convoy on Munich’s Leopoldstraße look like on the occasion of winning the Football World Cup in 2034? True, the driver has both hands free to swing the flags. But a driving speed of exactly 50 km / h is likely to cloud the festive mood. Or is a vehicle convoy with step speed, a hazard warning system, and horns planned in the common mode? What about the times when a manually controlled vehicle with a supposed traffic-rowdy person at the wheel winds through the traffic jam, while the autonomous vehicles avoid and circumvent a potential accident? Through networking and connectivity, will all vehicles get informed accordingly and form an alley-free ride for the driver? Driving schoolchildren in Germany are allowed to park at the push of a button. According to the driving license regulations, all equipment and systems available are basically approved. Driving with a modern “computer on four wheels” is not comparable to a vehicle without any comfort. The driver will quickly notice the progress when they enter their own car. The question of what has enabled the development on the vehicle side is answered quickly: software (in conjunction with sensors and actuators) and the wiring system. These two “actors” of a vehicle lie concealed behind covers or the carpet of a vehicle and would have actually earned a more prominent place.
An INRIX study from May 2017 shows that for 56 percent of German drivers, the integrated technology is an important factor that influences the purchasing decision as much as the vehicle performance. According to a representative survey conducted by Bosch, only 33 percent of Germans show interest in autonomous driving. They fear, for example, loss of control or even hacker attacks. In China, the survey is an astonishing 74 percent. But where does this big difference come from? Does the autonomous driving fit our lifestyles and us? Do we perhaps underestimate the safety aspect of road traffic, which is driven by autonomous and automated vehicles?
How do the Germans perceive the driverless car? Results of the study on the autonomous driving of TÜV Rheinland in May 2017 demonstrate this. A total of 1,400 drivers were surveyed across all age groups. The representative online survey is surprisingly positive, but the scepticism increases with increasing age. Since the agreement of the Bundestag and the Federal Council it has been clear, the autonomous vehicles will come—it is only a matter of time. The automotive groups and suppliers are eager to develop products and systems. The number of today’s driver assistance systems is growing steadily, and more and more test vehicles are in use around the world. But are the Germans also willing to use an autonomous vehicle and let themselves be driven? According to the first survey results, the German vehicle drivers were not yet convinced. The Figure 1.4 below shows survey results regarding the area of application. According to this survey, an average of 76 percent of the drivers interviewed can imagine a computer as a driver or are ready to use a driverless vehicle. A more detailed view of the supporters shows, however, clear differences, depending on the driver and use, as seen below in the Figure 1.5.
The rejection of autonomous driving appears to be the least where the technical challenges are highest: in urban traffic. It is striking in the analysis that, in particular, the younger generation of technology is very open-minded. Scepticism in technology increases with age. The seniors would, however, benefit from autonomous driving. There would be no need to ask for a new driving license or for mobility in old age. However, many people also see problems that are associated with autonomous driving. More than two-thirds are concerned about the difficulty of clarifying the debt issue and the liability in case of accidents. The initially important question about the alternatives in the case of unavoidable accidents has moved backward in prioritization. More concerned are the interviewees that hackers or cyber criminals supposedly manipulate the independent car. In addition to the above-mentioned questions about the basic readiness for autonomous driving, the focus of the study was also on the importance of system auditing and data protection, which around 90 percent of drivers see as important.
The automotive industry is racing to develop vehicles that are more and more autonomous, but the road ahead is tricky to navigate. Drivers are used to being in control. If the stakeholders push too hard and too fast, drivers won’t go along for the ride. It’s very much about the agency and control in the cockpit and how drivers and passengers want to interact with their connected vehicles. Today, drivers trust their vehicles because they themselves are in control. We can imagine that vehicles support some of our needs and give us choices. We can aspire that vehicles could predict our needs and make decisions for us. But the more our vehicles take control, the less we, as drivers, trust them. We don’t want to get overwhelmed by too much data and urgency. The vehicle has to be highly personalized and adaptive, and has to intelligently determine the right options to the user at the right time. But all in all, this is a very treacherous path that may eventually lead to a reduction in personal freedom and trust, as a certain number of drivers and passengers feel today.
Networked vehicles and autonomous driving are going to be the next big disruptive innovation in the years to come. We consider it as being predominantly technology-driven and therefore, we suppose it to have massive societal effects in many areas. When we radically reconstruct the transportation and travel infrastructure by networking vehicles with everything, we are also going to alter how our municipalities and neighborhoods look. We are going to transform the population settling in our rural and urban areas, and our economy, society, and culture.
A Nuance and DFKI study (Nuance, DFKI: Cognitive and Conversational AI for Autonomous Driving May 2017) about the needs of intelligent and collaborative assistants to effectively engage passengers in networked self-driving vehicles shows that the top five activities in a networked autonomous vehicle would be listening to the radio (64%), relaxing (63%), talking on the phone (42%), browsing the internet (42%), and messaging (36%). If driving with passengers, the percentages would change—drivers and passengers would be naturally having more conversations (71%) or listening to the radio (58%), rather than talking on the phone (only 19%) or messaging (23%). Integrated, multimodal user interfaces leveraging voice, touch, and visual cues are considered more agreeable and effective than visual cues, leading to faster reactions than simply vibrations or haptic alerts. Drivers trust audible and haptic responses from the automotive assistant more than visual cues alone.
The interaction of human and machine casts new ethical questions in the time of digitization and self-learning systems. The autonomous driving of vehicles could even be requested if it is thereby possible to reduce the number of accidents to zero. But in highly automated road traffic, dilemma situations cannot be completely ruled out. The approval of automated driving systems also depends on the considerations regarding human dignity, personal freedom of decision, and data autonomy. To deal with these issues, for example, the ethics commission of Germany’s Federal Ministry for Economic Affairs and Energy published guidelines (Bundesministerium für Verkehr und digitale Infrastructure. Ethik-Kommission Automatisiertes und vernetztes Fahren. June 2017) for partially and fully automated driving systems for the improvement of the safety of all involved in road traffic in June 2017.
These guidelines concluded that the protection of human beings has priority over all considerations of usefulness. The responsibility for the introduction and approval of the networked driving vehicles remains with the state and under official control. In the case of legal structuring, the individual’s right to free development must be taken into account. In dangerous situations, the protection of human life must always have priority over the prevention of damage to property and animals. In case of unavoidable inaccuracies, people must never be qualified according to personal characteristics such as age, gender, or physical or mental constitution. It is not software that decides, but the producer or operator of the autonomous vehicle driving system. Many dilemma situations in which the question is to live or to die can neither be standardized nor programmed. In any driving situation, it must be clearly regulated and recognizable who is responsible for the driving task—the human being, or the computer.
The report supplements the legislation for networked and automated driving launched in June 2017. Now the computer can partly take over the vehicle control. The last responsibility remains with the vehicle driver. The driver is obliged to resume the driving responsibility without delay if the highly or fully automated system prompts them to do so, or if they recognize, owing to obvious circumstances, that the requirements for the intended use of the highly or fully automated driving functions no longer exist.
Let’s assume that more than 75 million vehicles are sold annually, and the number of vehicles is around 1.2 billion. Then it takes around 15 years for the whole vehicle armada to turn over. For example, if vehicle makers start producing nothing but fully autonomous cars in 2020, we will still see a mix of manual and semi-autonomous vehicles until 2035 or later. The growing vehicle production rates per year and the popularity of used vehicle marketplaces make this turnover rate even much longer. Consequently, regulation authorities, insurance, and politics will have had to catch up long before this point in time. Safety and security standards have to be put in place, and the question of liability in cases of accidents has to be answered. After the turn over, municipalities will look drastically different. Sidewalks and bicycle routes could go away, as pedestrians, bicycles, and vehicles share the roads. There could be no street parking, since parking areas could be dynamically allowed everywhere. We could also get rid of traffic signs and infrastructure, which could be exchanged with computing and communications devices that only need to communicate with vehicles.
There are major legal and policy challenges surrounding the networked autonomous vehicle ecosystem. For example, we see an urgent call for action in regulation and standard setting for critical event control, driver responsibility, ownership and maintenance (Teare, 2014), civil and criminal liability, corporate manslaughter (Browning, 2014), insurance, data protection, and privacy issues (Khan, et al., 2012). Regulation, certification, and testing are required in homologation, periodical main examination, and data protection of the collected data. Regular monitoring of data protection and ensuring the reliability of autonomous vehicles will further increase the acceptance of potential users. Networked autonomous driving vehicles require an adaptation of existing legal frameworks needed for traffic and vehicle regulations. The adapted regulations accelerate mass production and the commercialization of networked autonomous vehicles. So, we perceive many development plans and initiatives of worldwide public authorities with the current objective to pave the way for a step-wise introduction of networked automated vehicles. Many of these are built around standardization, regulation, testing, safety, or networked autonomous vehicle technology developments.
These activities from various stakeholders, including governments in Asia, Europe, and the U.S., are supporting or even advocating vehicle communications. The U.S., which is ahead of other countries in developing and regulating driverless vehicles, plans to set its specific rules (Bryant Walker Smith: Automated Vehicles Are Probably Legal in the United States). In the U.S., several federal states have already passed laws authorizing networked autonomous vehicles testing on their roads (Walker S., 2014). Legislatures in California, Nevada, Michigan, Florida, and Tennessee have passed bills enabling automated driving (CIS Automated Driving: Legislative and Regulatory Action). The National Highway Traffic Safety Administration in the United States provides an official self-driving vehicle classification very similar to SAE levels. Instead of SAE 6 levels it differentiates between the levels: no automation, function-specific automation, combined function automation, limited self-driving automation, and full self-driving automation.
China is one of the most ambitious areas when it comes to networking vehicles with everything. The HD maps have issued a roadmap for having highway-ready, self-driving vehicles in 2021 and autonomous vehicles for urban driving by 2030 (Li Keqiang, Tsinghua University). China may utilize wireless data communications technology (like LTE or 5G) that is already used in many vehicles to access the internet, and adopt it for vehicle-to-vehicle communications rather than the dedicated short-range communications (DSRC) standards developed in the U.S. and Europe.
Europe started adapting the Vienna Convention on Road Traffic and the Geneva Convention on Road Traffic (Reuters, 2014) in order to be able to allow networked autonomous vehicles, but legal issues and uncertainty are still there. Japan and major European countries cooperate to urge the U.S. to adopt common standards compiled by a United Nations expert panel as part of the World Forum for Harmonization of Vehicle Regulations. The regulations will contain principles, such as safety provisions, controlling autonomous passing to highways, and holding human drivers accountable for any accidents.
In March 2017, Germany announced that the German Road Traffic Law (StVG) will get complemented in that vehicles with automated systems (highly automated or fully automated) will be allowed to be used in traffic on public roads as such, and that the vehicle driver is allowed to hand over the vehicle control system to the technical system in certain situations. The autonomous driving system can be manually overridden or deactivated at any time by the driver, who is obliged to take over the vehicle control immediately when asked by the system.
We still perceive societal, ethical, and political issues that must be answered as we move closer to commercial services of networked autonomous vehicles. How should engineers build ethical choices into automated features and advanced computing and communications algorithms? What are the factors in the algorithms, if there are any that would lead its system to turn one way or another? There are a wide variety of ethical issues that remain, and system architects have to make choices regarding how to deal with them. Learning how to treat these complicated societal and ethical issues is a major challenge facing the way forward (J. F. Bonnefon, A. Shariff, I. Rahwan, “The Social Dilemma of Autonomous Vehicles”). In this book, we do not elaborate further on this important topic, because we think its importance requires a book of its own.
1.5Outline
We treat in Chapter 2, the plethora of networking vehicles applications and use cases and provide insight into the many currently somehow disarranged views of computing, communications and vehicle stakeholders. We look at how large computing and communications stakeholders like Intel, Qualcomm and others define and specify vehicle networking and connectivity scenarios and use cases thereof. We reveal the view of big vehicle manufacturers like Audi, BWM, GM, Toyota and Volkswagen. And we look finally at the current hype in autonomous and automated vehicle networking and connectivity which is somehow climaxing in multi-stakeholder organizations like 5GAA, 5G-PPP, 3GPP, SAE, ETSI, GSMA and NGN. We provide important comments and offer the main use cases and scenarios from our view.
In Chapter 3 we show communications requirements for vehicle networking and connectivity as seen at this time by communications, computing and vehicle standards and regulation developing organizations mentioned in the previous chapter. Every camp has noticeably its own assessment and understanding of what kind of communications technologies are preferably to be implemented. The regulation of the spectrum resources in terms of frequency bands is another topic which needs to be solved soon, taking the development cycles of the vehicle industry and the time schedules of regulation bodies into account. We describe the challenges of dealing with these requirements to fulfill the future of autonomous and automated vehicle usage scenarios and propose a way forward.
We start with the state-of-the art technologies for vehicle networking and connectivity in Chapter 4 and look how the main vehicle platform components, sensors and actors, computing and communications building blocks are going to change for autonomous and automated vehicles. In particular we show the reader how these changes impact the communications technology opportunities which are currently available, together with the research and development challenges necessary to make cellular networking and connectivity a safety and mission-critical technology. LTE-A evolution and 5G cellular systems have the potential of supporting challenging and upcoming use cases that require low-latency, high reliability or high safety. Cellular V2X is able to work together with DSRC communications to enhance V2X communications.
At first glance, autonomous and automated driving does not need V2X communications. We take a second look and explain why V2X communications is just adding another sensor to the vehicle to support improved situational awareness, provide redundancy and make other sensors more reliable. V2X communications offer long range, data for collaborative driving and non-line-of sight capabilities. Based on the relationship of vehicle dynamics and communications requirements, we derive data throughputs for specific sensors and applications and map it against the performance of wireless communications technologies. We show that the ratio between uplink and down-link is symmetrical. V2X technologies seek to address, aside from automated driving and advanced driver assistance systems (ADAS), situational awareness, mobility services, and convenience services. There is the evolution of current ADAS, enabled by cameras and sensors, toward automated and autonomous driving where vehicle networking and connectivity augment ADAS and support it.
We elaborate in Chapter 5 on D2D and 3GPP release 12 side link communications and how it addresses the low-latency, high-reliability V2X use cases, for example in a complementary manner to DSRC. 3GPP LTE-based V2V achieves substantial link budget gains due to frequency division multiplexing (FDM) and longer transmission times. Hybrid automatic repeat request (HARQ) retransmissions are an option to realize higher link budgets. Turbo coding and single-carrier FDM implementation increase further link budgets. Combining sensing of the radio resources with semi-persistent transmissions at the system level exploits the periodic nature of V2V traffic and does not cause carrier sensing overhead. It results in better spatial reuse of radio resources.
We highlight in Chapter 6 the most widely implemented convenience service which is infotainment. Infotainment solutions for example in relationship with smartphones and software apps or the Genivi platform already provides today plentiful networking and connectivity solutions for V2X in many vehicles. Vehicles are integrating Android Auto, Apple CarPlay, Baidu CarLife or MirrorLink into their infotainment system. We look, in particular, at solutions to stream data into the vehicle and to upgrade and extend platform features via software using V2X and how these solutions get integrated with telematics, safety and connectivity planning. The current infotainment systems evolve into connected services, digital commerce, vehicle diagnostics, predictive maintenance, vehicle tracking and insurance.
We provide answers in Chapter 7 regarding the introduction of dynamically reconfigurable systems in networked vehicles which requires urgently an evolution of the existing framework conditions. In particular, the use of external software components for the dynamic expansion of a networked vehicle system poses challenges to the proven protection of these components. This needs to be done on the basis of a comprehensive component and building blocks specification, regulation and testing and, if necessary, must be assured by the certificate of a trustable stakeholder. Corresponding contractual frameworks between the suppliers of external components and building blocks, manufacturers and owner, driver or passenger of a vehicle can accompany the use of the components and building blocks from copyright and other legal aspects. We use wireless communications examples to illustrate the offered principles and solutions.
We give our assessment on the current status of vehicle networking and connectivity and the evolution of local- and wide-area communications technologies in Chapter 8. With a wide-ranging view on major vehicle ecosystem stakeholders, we explain why we think that V2X communications are a critical component of the networked and connected vehicle of the future. We motivate why the wireless communications ecosystem stakeholders shall engage in early efforts to assess the approaching capacity and coverage needs of networked connected vehicles now in a tight collaboration with vehicle industries.
References
2020 Roadmap, European New Car Assessment Programme (EURO NCAP) June 2014.
3GPP TR 22.885 Study on LTE support for Vehicle-to-Everything (V2X) services. V14.0.0 21.12.2015.
5G Americas V2X Cellular Solutions. October 2017.
Adriano Alessandrini, Andrea Campagna, Paolo Delle Site, Francesco Filippi, Luca Persia, Automated Vehicles and the Rethinking of Mobility and Cities, In Transportation Research Procedia (2015), Vol. 5, pp. 145–160.
Audi: Mission accomplished: Audi A7 piloted driving car completes 550-mile automated test drive. Press release 04.01.2017.
Autoware: Open-source software for urban autonomous driving. https://github.com/CPFL/Autoware.
Bryant Walker Smith, Automated Vehicles Are Probably Legal in the United States, 1 Tex. A&M L. (2014) Rev. 411.
Brandon Schoettle and Michael Sivak, A Survey of Public Opinion about Autonomous and Self-Driving Vehicles in the U.S., the U.K., and Australia. UMTRI-2014-21.
Bundesministerium für Verkehr und digitale Infrastructure. Ethik-Kommission Automatisiertes und vernetzes Fahren. Juni2017.
CIS Automated Driving: Legislative and Regulatory Action https://cyberlaw.stanford.edu/wiki/index.php/Automated_Driving:_Legislative_and_Regulatory_Action#State_Regulations
ETSI TR 102 638 V1.1.1, Intelligent Transport Systems (ITS); Vehicular Communications; Basic Set of Applications; Definitions.
ETSI TS 102 637-1 V1.1.1, Intelligent Transport Systems (ITS); Vehicular Communications; Basic Set of Applications; Part 1: Functional Requirements.
European Road Transport Research Advisory Council (ERTRAC) Automated Driving Roadmap v5.0 21.07.2015.
European Road Transport Research Advisory Council (ERTRAC) Automated Driving Roadmap v7.0 29.05.2017.
German Federal Government, Strategie Automatisiertes und Vernetztes Fahren, Strategic Paper, Jun 2015.
Gräter: Rollout of automated driving in cities, infrastructure needs and benefits. Conference connected and automated driving April 2017. http://connectedautomateddriving.eu/wp-content/uploads/2017/03/Day1_B02_Graeter_EC_CAD.pdf
https://itpeernetwork.intel.com/data-center-holds-keys-autonomous-vehicles/ iMobility Forum, Automation in Road Transport, Version 1.0, May 2013.
J. Carlson, ADAS Evolving: New Developments in Hardware and Software on the Road to Autonomous Driving, HIS Automotive, Spring Media Briefing, Detroit, 19 Mar 2015.
Jean-François Bonnefon, Azim Shariff, Iyad Rahwan, The Social Dilemma of Autonomous Vehicles, In Science, 24 June 2016.
Nuance, DFKI: Cognitive and Conversational AI for Autonomous Driving, May 2017.
Projekt Sim-TD, Sichere intelligente Mobilität, Testfeld Deutschland, BMBF-Projekt FKZ, http://www.simtd.de
Roadmap on Smart Systems For Automated Driving, European Technology Platform on Smart Systems Integration (EPoSS) (2015).
Rudin-Brown, C. M. & Parker, H. A., 2004a. Behavioral adaptation to adaptive cruise control (ACC): Implications for preventive strategies. Transportation Research Part F, 7(2), p. 59−76.
SAE, Taxonomy and Definitions for Terms Related to On-Road Motor Vehicle Automated Driving Systems, J3016, SAE International Standard (2014).
SafeTRANS: Eingebettete Systemin der Automobilindustrie – Roadmap 2015 – 2030. 21.09.2015
SCOUT and CARTRE http://connectedautomateddriving.eu
State of California: Proposed Driverless Testing and Deployment Regulations 10.03.2017 https://www.dmv.ca.gov/portal/dmv/detail/vr/autonomous/auto
Statista: Number of cars sold worldwide from 1990 to 2017. https://www.statista.com/statistics/200002/international-car-sales-since-1990/
Tracking Automotive Technology http://analysis.tu-auto.com/
United Nations: Report of the sixty-eighth session of the Working Party on Road Traffic Safety. http://www.unece.org/fileamin/DAM/trans/doc/2014/wp1/ECE-TRANS-WP1-145e.pdf
Volkswagen: VW will enable vehicles to communicate with each other as from 2019. https://www.volkswagen-media-services.com/en/detailpage/-/detail/With-the-aim-of-increasing-safety-in-road-traffic-Volkswagen-will-enable-vehicles-to-communicate-with-each-other-as-from-2019/view/5234247/6e1e015af7bda8f2a4b42b43d2dcc9b5?p_p_auth=afnpxKt928.06.2017
Weltbestand an Autos: http://www.live-counter.com/autos/