Automated driving systems
Automated driving: Addressing challenges and exploiting opportunities
Reliable automated driving systems that require no human operation or relieve drivers of certain responsibilities have the potential to dramatically reduce the number of traffic accidents. Such systems will never become tired or distracted or allow themselves to be influenced by emotions; all of these are factors that frequently play a role in accidents caused by human error.
The potential improvement of traffic safety is not the only benefit offered by automated driving systems. The technology can also enable efficient, resource-saving traffic and can also help to reduce emissions.
Automated driving systems also offer extensive potential for road freight transport in terms of safety, since most accidents in the road freight transport sector are also caused by human error. For example, automated driving systems can support truck drivers in demanding driving situations and on long, monotonous trips.
Despite all the benefits, care needs to be taken, as ethical issues and data-protection risks must also be considered as automated driving systems are developed further. Daimler does this as early as the product development stage. An important consideration here is the responsible use of artificial intelligence (AI). AI as a component of self-driving vehicles is particularly important with regard to machine learning, since among other things AI helps the system quickly and reliably identify objects and situations in or next to the roadway.
Along with safety, we believe that the consideration of social, ethical, and data-protection aspects will play a key role in gaining acceptance for automated driving systems.
Paving the way for automated driving
Our goal is to continue to develop automated driving systems. We are placing equal emphasis here on technical, legal, and ethical aspects, and to this end we have formulated the following key points:
- Daimler seeks to play a leading role in the field of automated systems. One example here is the new S-Class from Mercedes-Benz, in which the automated DRIVE PILOT and INTELLIGENT PARK PILOT are to be used for the first time.
Mercedes-Benz is making the leap to conditionally automated and highly automated driving
- Daimler supports the establishment at national and international levels of a reliable legal framework, technical standards, and ethical guidelines relating to the use of the new technology and therefore promotes the broad-based public dialog that is necessary.
- Daimler actively participates in the social and political dialog on the ethical questions that are arising in the context of the new technologies.
Uniform regulations create safety and trust
New technologies require legal certainty. That’s why Daimler is a member of German and international bodies and associations that promote the establishment of consistent legal standards for automated driving. We seek to support the development of a secure legal framework for the technical certification of automated driving systems, and we will continue to monitor the legislative process in this area in the future.
In Germany, the legal basis for automated driving systems is defined by the automated driving amendment to the Road Traffic Act (StVG), which went into effect in 2017. We welcome this amendment because it makes Germany one of the first countries to provide a legal basis for further technological developments. Beyond that, we also believe that the respective national legislation regarding traffic and behavior needs to be further developed in order to establish legal certainty in connection with the use of conditionally automated systems (SAE Level 3). Further changes need to be made to road traffic law if fully automated driving in particular is to become a reality. We therefore also welcome the initiative of the German federal government for the establishment of an approval framework for the operation of motor vehicles with autonomous driving features.
Other countries have in the meantime created legal frameworks or initiated legislative processes regarding the use of automated driving systems. If the technology is to be launched on the market, not only will amendments have to be made to respective national traffic laws; measures will also have to be taken to make it possible to approve and register conditionally and highly automated driving systems for actual use on the road. The United Nations Economic Commission for Europe (UNECE) has established conditions for this by implementing the Automated Lane Keeping System (ALKS) regulation in January 2021. This will make it possible to launch initial conditionally automated systems for use in traffic jam situations on highways in Germany.
In order to enable the cross-border use of automated cars, international harmonization of the relevant legal regulations will also be necessary. These should be as compatible as possible and include the same technological requirements.
This also involves the issue of how the data needed to ensure the proper operation of automated driving systems should be handled. One example is the technical standardization of the driving mode recorder for automated driving systems that is required by law in Germany. Among other things, this device records when an automated system was activated or the driver controlled the vehicle. Experts from the UNECE are currently exploring ways to establish an international technical standard for such a recorder. We support this effort and emphasize the importance and necessity of ensuring data security in such recording technologies. The ALKS regulation already includes technical requirements relating to data storage in the driving mode recorder when the vehicle is driven in the conditionally automated mode.
Responsible development of automated driving systems
The development of automated driving systems presents special challenges. This is why we make use of the instruments in our technical Compliance Management System (tCMS) during the development process. We have formulated specific behavior guidelines for automated driving systems, for example. Complex questions in this area are examined and answered in an interdisciplinary process that takes technical, legal, and certification criteria into account.
We also comply with ethical principles and additional internal rules and regulations such as the principles of our data vision and our AI principles for the responsible use of artificial intelligence systems. They also apply to automated driving with regard to the software aspects as well as to social aspects, for example, in relation to hardware. These are based on our corporate values and have also been incorporated into our Integrity Code. We also comply with our internal tCMS policy and ISO standards 26262 and 21448 for safety-relevant electrical/electronic systems in vehicles.
In addition, our product-development activities are guided by the German government’s Ethics Commission’s 20 ethical rules for automated and connected driving. We also take into account draft proposals and resolutions relating to planned regulations and standards, and thereby take account of the dynamic developments in the area of automated driving. Furthermore, we also comply with external guidelines such those formulated by AI4People, the Institute of Electrical and Electronics Engineers (IEEE), and the High-Level Expert Group on Artificial Intelligence (EU).
Combining expertise for automated driving systems
We use an integrated approach to answer the technical, social, ethical, and legal questions relating to automated driving. The participants include our experts from research and development, product safety, and quality management — and also since 2018 an interdisciplinary team at the Integrity and Legal Affairs executive division. The team works with engineers, legal advisors, and specialists in data protection, compliance, social sciences, and philosophy to assess the potential impact of new technical developments. It also increases awareness of complex social and legal issues, and develops and implements new solutions. The topics addressed include the responsible use of data in programming processes and the possible changes to behavior in urban environments that might be brought about by the use of new technologies. The objective of this approach is to increase both the safety and the acceptance of our products.
Comprehensive data protection is also important for ensuring public acceptance of automated driving systems. This is why we involve our data protection experts in our concept development processes at a very early stage. The goal here is to develop data-protection-friendly concepts in accordance with the privacy by design principle.
Open technology dialog strengthens trust
We promote an open dialog between business and consumer associations, government authorities, industry representatives, and society at large because we believe that a broad-based social discussion is a prerequisite for the acceptance of automated driving systems.
Since 2015, we have been using the annual “Daimler Sustainability Dialogue” to discuss ethical, legal, and social questions that arise in connection with automated driving systems. The most recent “Daimler Sustainability Dialogue” took place as a digital event on November 5, 2020. Participants in the Traffic Safety working group at the event discussed sustainable traffic safety concepts and other relevant topics. All agreed that traffic safety remains one of the key central issues for the future orientation of our sustainable business strategy.
The participants also talked about the future of traffic safety, above all from a social perspective. The discussion concerned not only the technical reliability of the systems but also the societal changes that might be brought about by the use of the new technologies — and the extent to which people will be willing to accept these changes. Such a dialog regarding their social impact is especially important in order to be able to determine how we can ensure that our technical systems will be accepted by society in the future.
Daimler will focus on the following issues in particular in 2021:
- Support for the creation of a competence center in Freiburg that will conduct research on future traffic scenarios and patterns in order to gain a scientific understanding of the impact the new technologies will have on society.
- Development of arguments for regulating the handling of data as a component of the Federal Ministry of Transport and Digital Infrastructure’s traffic safety program for 2030.
- Work on our internal scorecard for traffic safety in order to help us achieve our sustainability targets in this area.
Involvement in committees and associations: We support responsible decision-making
Daimler is a member of numerous national and international committees and associations, including the German Association of the Automotive Industry, the European Automobile Manufacturers’ Association, and the working groups of the UNECE. Within the framework of these memberships, we participate in consultation processes regarding new legislation and share ideas and information with political decision-makers.
- Daimler joined the Automated Vehicle Safety Consortium (AVSC) in April 2019. The consortium develops safety principles for automated driving, with a focus on safety tests before and during the use of automated vehicles, data processing and protection, and the interaction between automated vehicles and other road users.
- Since July 2019 we have also been participating in a workshop on “Ethical aspects of the standardization of artificial intelligence in autonomous machines” that is organized by the German Institute for Standardization (DIN). The workshop was part of our participation in a project with DIN and the German Commission for Electrical, Electronic & Information Technologies. The goal of this project was to create a roadmap known as “Ethical aspects of the standardization of AI.” This project, which was completed in April 2020, resulted in the publication in September 2020 of a white paper on “Ethics and Artificial Intelligence.” Among other things, the white paper recommends the use of several assessment criteria to ensure that AI applications do not infringe upon human autonomy.
- The ideas outlined in the white paper with the title “Safety First for Automated Driving” (SaFAD) that we published in July 2019 together with leading companies from the automotive and supplier industries are now being incorporated into international standardization processes. The twelve main principles presented in the white paper were used as a basis for the ISO Technical Report TR 4804, which we helped to produce and was published in 2020. Plans now call for the principles to be described in further detail as ISO Technical Specification TS 5083.
- Since July 2019 we have been participating in the research association for “Legally Viable and Efficient Homologation of SAE Level 4 and Level 5 Autonomous Vehicles” (VVM). This association basically picks up where the “Project for the establishment of generally accepted quality criteria, tools, and methods as well as scenarios and situations” (PEGASUS) that was funded by the Federal Ministry for Economic Affairs and Energy left off. PEGASUS was completed in 2019. The association has set itself the goal of developing systems and methods for the safety verification of highly automated and fully automated vehicles and driving functions.
- Since 2019 we have been participating in the ISO TC/241 WG6 through the DIN Standards Committee Road Vehicle Engineering mirror group. The topic of the working group is “The development of recommendations for ethical considerations in connection with autonomous vehicles.” The idea here is to firmly establish an ethical perspective in the development process for automated vehicle systems. The International Organization for Standardization (ISO) plans to publish the recommendations in the summer of 2022.
We seek to initiate a public discussion on technical standardization
Automated driving systems will only be approved for road use if they can meet very stringent safety requirements. This is why we are working hard to define the technical standards needed here. The results of the PEGASUS and VVM projects with regard to testing methods and the approval of automated driving functions were presented publicly on a regular basis. We support the continuation of these activities and their harmonization with international efforts in this area — for example within the framework of ISO activities.
In the United States we have published Voluntary Safety Self-Assessments (VSSAs) of our joint projects with our partner Bosch in Sunnyvale, California (SAE Levels 4-5), and of DRIVE PILOT (SAE Level 3) in Long Beach, California and Ann Arbor, Michigan. We use the VSSAs as a basis for interaction with government organizations and other stakeholders and for public discussions with them about our projects.
Mercedes-Benz is making the leap to conditionally automated and highly automated driving
Mercedes-Benz is determined to introduce technologies that enable the safe operation of conditionally automated vehicles and meet the demanding legal requirements for a Level 3 automation system. With DRIVE PILOT (SAE Level 3) and INTELLIGENT PARK PILOT (SAE Level 4), Mercedes-Benz is now taking the decisive step toward conditionally automated and highly automated driving. Both systems are planned to be used in the new S-Class, which means Mercedes-Benz now intends to offer its customers the possibility to turn the task of driving over to vehicle systems in a production model.
We expect that beginning in the second half of 2021, the new Mercedes-Benz S-Class will be able to drive in conditionally automated mode (SAE Level 3) with the new DRIVE PILOT in congested traffic, or in traffic jams, on suitable highway segments in Germany.
Mercedes-Benz is going a step further with parking, as a pre-installation of the INTELLIGENT PARKING PILOT system is available as an option for the new S-Class. This means the new S-Class is already prepared for driverless highly automated parking (SAE Level 4). This function may not be used yet, however — but it will be possible to use it as soon as national legislation permits it, parking facilities are equipped with the necessary infrastructure, and the associated Mercedes me connect service has been activated.
Automated commercial vehicles can increase safety and productivity
Partially and highly automated trucks offer several benefits for our customers, the economy, and society. These include fuel savings and greater efficiency in transportation processes, for example. We also expect that automation can improve the flow of traffic, particularly on highways. More than anything else, however, automated trucks can help increase road safety. Under certain conditions, redundant systems and numerous sensors and instruments can reduce the risk of an accident or mitigate the consequences of those accidents that do occur.
Daimler Trucks already offers partially automated driving systems (SAE Level 2). These systems are available in trucks from our Mercedes-Benz, Freightliner, and FUSO brands in our most important markets on three continents.
In commercial transport, highly automated driving (SAE Level 4) is the next step, as this level of automation could further increase safety, efficiency, and productivity. Daimler Trucks focuses on three principles in its research and development activities for automated trucks:
- The safety of vehicle occupants and other road users is our top priority.
- We develop our products in line with the requirements of our customers and in cooperation with them.
- A clear legal and regulatory framework for issues related to vehicle operation and liability must be established.
A major focus of development at Daimler Trucks over the next few years will be to develop highly automated trucks (SAE Level 4) to the series production stage.
We took a major step in this direction by launching a strategic partnership between Daimler Trucks and Waymo in October 2020. In the initial phase of cooperation, Waymo’s automated driving technology is being installed in a variant of the Freightliner Cascadia from Daimler Trucks North America that was developed especially for automated driving applications. Waymo and Daimler Trucks want to use highly automated trucks (SAE Level 4) to improve both road safety and fleet-customer productivity. The highly automated Freightliner Cascadia truck (SAE Level 4) equipped with the Waymo Driver system is supposed to be made available to customers in the United States in the next few years. Waymo and Daimler Trucks will also soon be examining possibilities for expanding their activities to other markets and vehicle brands.
The Autonomous Technology Group brings together all of Daimler Trucks’ expertise and activities related to automated driving worldwide. This global unit’s central responsibilities include the further development and implementation of our overall strategy for automated driving, including all research and development activities, and the establishment of the required infrastructure and network for vehicle operation.
Torc Robotics, a software firm located in Blacksburg, Virginia in the United States, is part of the Autonomous Technology Group. Torc Robotics is one of the world’s most experienced companies in the field of conditionally automated driving with highly sophisticated, roadworthy technology. Daimler and Torc Robotics are now jointly testing highly automated trucks (SAE Level 4) on selected public highways — including in Virginia and New Mexico. Previously we had tested the technology for months on closed-off tracks. These tests represent an important further step for Daimler Trucks in its effort to put highly automated trucks that are safe and reliable on the road around the world in the future.
Luminar Technologies Inc., one of the world’s leading suppliers of lidar hardware and software technology, is also part of the Daimler Truck AG development network. We are working closely together in order to enable Luminar’s lidar technology to be used in highly automated (SAE Level 4) applications for trucks moving at higher speeds, such as those commonly reached by trucks in the United States, for example. Activities here relate in particular to object recognition, data processing, and overall system performance. Here we are initially focusing on employing the technology on US highways in order to enable its safe commercial use in production vehicles.
How we assess the effectiveness of our management approach
The sound decisions made in our development projects form the foundation for ensuring product safety and technical compliance. Certain potentially feasible future developments are still not addressed in the external provisions and regulations regarding automated driving systems. Interdisciplinary expert and decision-making committees jointly define the requirements for the design of products that are used in automated driving systems.
In addition, all employees at the development departments can submit technical compliance questions to the responsible tCMS units, which then make their decisions within the framework of an interdisciplinary process. During the reporting year, the established tCMS units used this interdisciplinary process to deal with questions related to automated driving.