AI Knowledge (5421000)
Autonomous and connected driving will be a central component of the mobility concept of the future. One of the key factors in the implementation of automated driving functions is currently seen as the use of algorithms that employ artificial intelligence (AI) methods. The aim of the present project is to create a comprehensive system for incorporating knowledge into the training and validation of AI functions. The results will contribute significantly to increasing the functional quality, data efficiency, plausibility and validation of AI-supported functions in the automotive sector. In particular, the demand for rule conformity with all nationally applicable legal standards (StVO, StVG, etc.) is already a central demand of the legislator and is at the centre of the discussion in licensing regulations and the ethical debate on the future of automated and connected driving. The integration of normative knowledge in AI-supported functions is therefore equally groundbreaking and requires an adequate design of approval regulations and formulation of requirements for automated vehicles of SAE Level 3 and 4. The project AI Knowledge is creating a system to map these requirements in a holistic process.
Highly automated driving (AD) offers improved safety together with a multitude of previously unimagined possibilities such as reduced disparities, stress, and more meaningful activities while driving. However, realistic AD systems involve multiple AD levels where the tasks accomplished by driver and automation differ, leading to novel driver roles in which he/she needs to cope with these transitions. Therefore, AD level transitions require well-designed human machine interfaces (HMIs) to allow the driver to establish accurate situation awareness. The project investigates how fluid HMIs improve and allow for appropriate performance of the driver role during active driving automation and transitions. HADRIAN aims at safe transitions and high trust in automation. In addition to investigations on human-machine-interaction, BASt also contributes to HADRIAN with analyses regarding infrastructure constraints and opportunities.
VV-Methoden pursues the goal of developing methods for the safety verification of automation level L4/L5-vehicles on the basis of test scenarios, with a focus on the inner-city area and the use case "urban intersection". Central elements of the project are: A methodology to minimise the number of necessary test scenarios for L4/L5 driving functions, based on a comprehensive understanding of the cause-effect relationships in potentially accident-prone scenarios. A methodology that enables the reuse of test results and the exchange of tests between manufacturers and suppliers for the safety verification of an overall system via the testing of modular subsystems. The development of an L4/L5 assurance infrastructure to demonstrate the procedures developed in the project and additionally enable a seamless connection of different verification procedures. This requires the development of a continuous verification chain and the consolidation of interfaces in the infrastructure. The project has 9 sub-projects. In SP9, the BASt will contribute to the development of the assurance platform within the framework of the creation of an architecture and role model. In this context, the economic, legal and technical framework conditions will be examined that enable the collaboration of all system-relevant partners of a safeguarding platform.
Systematization of suitable non-driving related activities for automated driving of heavy goods vehicles (82.0737)
Continuous monitoring of the automated driving system and the environment by the driver is no longer required during automated driving at SAE Level 3 or higher automation levels. In future, drivers are allowed to carry out non-driving related activities (NDRA) during automated driving. If a situation cannot be controlled by the automated driving function, the system will display a take-over request and the driver must resume vehicle control after a short lead time. Therefore, there is a need for systematization and assessment of possible NDRAs, which includes criteria on suitability for automated driving at SAE Level 3 and Level 4. Criteria which are to be used for the assessment include aspects of human-machine interaction and road safety and, with regard to commercial vehicles, have to take economic benefits into account. From the perspective of human-machine interaction and road safety, criteria focus on the effects of NDRTs on driver state (availability to take over control) and driver’s takeover performance. One of the main issues during long term automated driving at higher automated levels is the increase of fatigue due to monotony. The aim of the project is a comprehensive systematization (categorization, assessment, catalog) of possible NDRAs for automated driving in trucks (SAE Level 3 and 4, heavy goods vehicles over 7.5 tonnes).
VV methods project (5419004)
The project focuses on the development of a methodological approach for the safety verification of highly automated and autonomous vehicles of SAE level 4/5 in urban environments. The project builds on the results of the predecessor project PEGASUS, which examined automated vehicles of SAE Level 3 on motorways. The extension of the scope of application of automation to urban areas is accompanied by an increasing complexity of the road traffic aspects to be considered. Exemplarily, the application case of the inner city intersection is considered, which is characterized by crossing and turning traffic, the effect of traffic lights, the design of separate lanes for turning as well as a multitude of different road users - especially vulnerable road users like pedestrians and cyclists. In the project, BASt formulates requirements for the project results of VV methods from the point of view of a regulation-related federal research institution. In the context of the homologation of urban SAE-Level 4/5 vehicles, a profound understanding of the hazards in a complex urban traffic environment is essential. Based on a fundamental understanding of the interrelationships between traffic and hazards, risk factors can be identified and generalised, thus allowing the scope of testing for the homologation of the systems to be derived.
EU research project HADRIAN (5419007)
The EU research project HADRIAN (Holistic Approach for Driver Role Integration and Automation Allocation for European Mobility Needs) aims at the safe use of a vehicle equipped with several different systems of different automation levels. The safe use of an automation system requires the driver to know which parts of the driving task can be performed by a system, where possible system limits lie and which tasks the driver may still have to perform. The overall goal of the EU project HADRIAN is to research a safe human-machine interface that ensures that drivers are always aware of their role, even when switching between functions of different automation levels. System boundaries of automated driving are mostly triggered by situations in the vehicle's environment, so that there is a correlation between the vehicle-internal driver involvement and the road infrastructure, which is considered in the project. BASt and 15 other European partners from industry, science and the public sector are working together in the project, which is funded by the EU within the Horizon 2020 Programme.
The aim of the project is the development, production and testing of an individual, electrically driven micromobility vehicle. In particular, it should allow pedestrians to safely and comfortably carry heavy and large-volume goods. The micromobility vehicle will be able to follow users automatically. However, it is also intended to offer autonomous delivery of goods. Together with partners, BASt will prepare traffic safety related requirements for the micromobility vehicle and carry out functional and vehicle tests of the hardware. BASt will also carry out the analysis and evaluation of the legal licensing and operating requirements.
Fundamentals of communication between automated vehicles and road users (82.0701)
Since vehicle automation is developing only slowly, which means step by step, it can be assumed that in the future both non-automated and automated vehicles will be in road traffic, which means so-called mixed traffic. The aim of the research project is to consider traffic situations that must be assessed as critical because of future mixed traffic, because today's forms of communication such as gestures or looks cannot be transferred to the communication of road users with automated vehicles. First of all, relevant forms of communication between road users will be analysed, which today occur in different traffic situations (city, country, motorway), and forms of communication will be developed which have proven themselves and can be transferred to future mixed traffic. It is also important to identify and evaluate new forms of communication where today's forms of communication do not work due to mixed traffic.
EuroNCAP HMI-AD (5417004)
The safety of automated driving greatly depends now as in the future on the safe interaction between the driver and his automated vehicle. It was therefore decided to include the aspect of the man-machine interaction in the assessment process in the EuroNCAP Roadmap 2020 to 2025. The aim is to make it possible to assess a safe interaction between the driver and the automated driving function using the current human factor research methods. In view of the absence of a method up to now, it was decided to approach the object of assessment from a best-practice point of view. For this purpose, standardised assessment methods are being developed which incorporate test subject investigations in addition to expertise. Continuous automation functions available on the market are to be considered at a use case level, starting with SAE Level 2.
EU project L3Pilot (5417006)
This project is funded as part of the EU Horizon 2020 assistance programme and aims to test the functionality of automated driving in passenger cars as a safe and efficient means of transport and to research and promote new service concepts based on automated driving for comprehensive mobility. Four sub-goals have been derived from the general objective: create a harmonised European test environment for pilot studies on automated driving, coordinate preparations for data capture, conduct and assess pilot studies on automated driving functions and on cooperative automated driving as well as initiate and promote the market launch of automated driving. The coordinator of the EU project is Volkswagen AG, Germany. As part of the project, the Federal Highway Research Institute (BASt) is conducting a study to examine the long-term influence of automated driving on user acceptance and is involved in the assessment of the effects of automated driving on safety.
PEGASUS FAT project (5416002)
The PEGASUS project (project for the establishment of generally accepted quality criteria, tools and methods as well as scenarios and situations for the release of highly automated driving functions) is intended to close important gaps in the area of testing through to the release of highly automated driving functions. The aim is to develop an approach to test automated driving functions to facilitate the rapid introduction of automated driving on the roads. As an associated partner in the project, the Federal Highway Research Institute (BASt) participates with an "Evaluation of automation risks of a highly automated driving function" and receives funding from FAT. BASt aims to identify automation risks which may arise from the interaction between a highly automated driving vehicle and the driver. The influence of the traffic density in transfer scenarios to the driver is firstly considered here in a field study. The driving tests are intended to uncover possible safety risks during transition and make a first assessment of the influence of the traffic situation. In a further step, an extended application scenario is to be defined in coordination with the PEGASUS consortium and FAT, which is similarly to be tested in the field. BASt’s own test vehicle is to be used for the empirical test studies to realistically reflect an automated driving function.
The European research project AdaptIVe aims to improve efficiency and safety in the road traffic sector. AdaptIVe strives to achieve an optimal interplay between drivers and driver assistance systems with the help of enhanced sensor technology, cooperative vehicle technologies and integrative strategies. To achieve this interplay, the degree of automation adjusts dynamically to the traffic situation and the abilities of the driver. In order to make automated driving a reality, the focus is not limited to the technological aspects of the project, but also includes legal aspects in order to ensure its successful market launch. The project will examine automated driving functions for motorways and for urban traffic, as well as specialist areas of application such as fully-automated parking. The Federal Highway Research Institute (BASt) is involved in several subprojects, such as the legal assessment of the compatibility of driver assistance systems with European road traffic law, product liability law and data protection law. The BASt will also work to develop a common evaluation framework for automated applications. This work will concentrate on developing new methods for the assessment of safety improvements as the result of automated driving functions.
Infrastructure Requirements for Automated Driving: A Research Project Covering the Basic Issues (82.0623)
Whilst the automation of vehicles is being contemplated as a way of breaking the closed loop system between driver and vehicle, a lack of clarity remains regarding the infrastructure requirements which might be necessary to enable automated driving to become a reality. This project, covering the basic issues, seeks to identify the infrastructure requirements necessary to implement these highly-automated systems and their technical requirements. The study will examine systems which might be expected to exist on the roads in reality in the foreseeable future.