European Roadmap Smart Systems for Automated Driving

European Roadmap Smart Systems for Automated Driving

1 Introduction Smart components and their systems integration,

traditional strengths of the European high technology industries, more and more develop into key enabling technologies (KET) for innovative products and applications [2]. This trend has been most obvious in the automotive sector for many years, where driver assistance systems enabled breakthroughs in road and passenger safety, energy efficiency and emission reduction [3]. Further along this path, higher degrees of road vehicle automation appear to be feasible soon [4, 5]. Automated driving (AD) will, on the long term, contribute to the reduction of road fatalities, increase productivity and social inclusion, and add value in terms of energy efficiency and the protection of the environment. As European car manufacturers and automotive suppliers have been successful in developing and implementing advanced driver assistance systems, the trend towards AD is based on an excellent knowledge foundation. It thus serves the objectives of increasing the competiveness of the European industry on global markets. Furthermore, AD, particularly high and full automation, represents a promising application of the Internet of Things (IoT) in the mobility sector [6]. Therefore, the aim of this roadmap is to share information about state-of-the-art efforts of the European industry, and to state what research actions have to be taken when in order to meet the milestones along the path towards implementing AD. Social and legal challenges that have an effect on complete system implementation and usage of AD in the future are carefully taken into account, and the worldwide developments in the field are reviewed. The recommendations given by this roadmap are expected to influence the whole chain of economic value added. Therefore, this roadmap is kept open to contributions from all involved stakeholders.

1.1 Automation Levels As a foundation for a deeper analysis

the levels of automation and the criteria for their definition have to be considered (see Figure 1) since these may be a source of confusion when developments on AD are discussed. There are three such fundamental criteria to be considered when defining the level of vehicle automation. The first important criterion refers to the controlling functions, i.e., the ability of the system to take over none, either longitudinal or lateral control, or both at the same time. The second criterion is related to the human driver and whether he is allowed to dedicate his attention partially or completely to other activities except driving. The third criterion considers performances of the vehicle and its ability to independently "understand" the processes that appear during driving. According to SAE International road vehicle automation can be classified into six different levels [7, 8]. Levels 0-2 take into account the human driver as the main actor responsible during driving. In case of faults, the human driver has less than one second to react and he or she isn't allowed to divert his/ her attention towards any other activities except driving. While the European suppliers of automotive smart components and systems invented and further improved driver assistance systems for lateral and longitudinal control of levels 0 and 1 in recent years, systems for partial automation of level 2 are currently under demonstration and in the early market place phase [4]. The most advanced solution, a combination of driver assistance systems like adaptive cruise control (ACC) and lane departure warning (LDW), is applied in high-end vehicles today [4]. For higher levels of automation, as Levels 3-5, complicated driving and decision making processes will be adopted by the vehicle in a stepwise manner. For level 3 or conditional automation, the vehicle is becoming aware of its surroundings. The reaction time for the human driver increases to several seconds, i.e. the vehicle will alarm the driver with a request to intervene, if necessary. For automation levels 4 and 5, the reaction of the human driver 3 extends to the couple of minutes, as the vehicle is becoming able to react independently during the entire drive. Level 3 of automation thus allows the human driver to do other activities while driving, whereas, levels 4 and 5 consider a complete adoption of the driving process by the vehicle while the driver is even able to fall sleep. There are however, other definitions for vehicle automation levels available and commonly used in practice. The National Highway Traffic Safety Administration (NHTSA) in the US uses five different subclasses instead of the described six. A vehicle having the driver only and no other assistance systems or automation classifies as level 0 and the fully automated classifies vehicle as level 4. In other words, with this arrangement no difference will be made between "high" and "full" automation as by SAE (see Fig. 1). Both, the conditional and high automation, assume that the human driver does not have to permanently monitor the system, but in necessary cases, he will be requested to take over the control with a certain time buffer. In this document we will mostly concentrate on automation Level 3 and higher, where the vehicle’s smart systems take control over the vehicle, even in critical situations.

1.2 Predictions

on Automated Driving Automated driving has attracted much attention recently mainly due to spectacular announcements by several players from the automotive and IT sectors. Also, a number of roadmaps or position papers were published announcing the importance of this topic. According to the VDA the greatest advantages of automated driving, compared to standard vehicles, are the increase in safety and the possibility of fluent, uninterrupted traffic [9]. The eNOVA Strategy Board on Electric Mobility and CLEPA emphasize the importance of R&D on key enabling technologies that would seriously contribute to an evolution of AD in Europe [10, 11]. Also, the electrification of vehicles is expected to leave space for synergies with an idea of automation of transport. Already in 2012, European Tier 1 suppliers predicted that the implementation of highly automated driving will be possible from 2020 and fully AD to start from 2025 [12]. The usage of partial automation should already be available from 2016 for "stop and go" situations on freeways at the speed of 30km/h. Similar predictions were made in the ITS roadmap of CLEPA that forecasts the implementation of highly automated driving between 2020 and 2025 [4, 10]. The German VDA expects the implementation of the level 2 automation on a short term, and the level 3 on a mid-term [4]. Even though the research progress is enormous and would respond to the predicted terms, there are significant legal Figure 1: Levels of automated driving as defined by e.g. SAE. For comparison, definitions of automation levels of NHTSA are also given. The latter one comprises high and full automation levels towards level 4 (high automation). 4 boundaries that need to be amended. Also, considerable safety issues of AD are a challenge that can only be bridged by further development of environment monitoring, perception, and driver assistance enabled by smart components and systems. Intelligent Transportation Systems (ITS) are seen as an important enabler of AD in many of the roadmaps. CLEPA suggests development of technical solutions for cooperative systems and automated driving vehicle technologies [10]. The iMobility Forum published the first version of their roadmap on "Automation in Road Transport" in May 2013, where the emphasis was put on the analysis of all possible applications of novel technologies [13]. The authors provided contributions to the commercial aspect of future technologies with respect to various levels of automations. Even though, as in the case of electrification of road transport [14], AD can be applied to all traffic participants like bikes, motorcycles, cars, trucks etc., the roadmap at hand concentrates only on the automation of passenger cars. This will simplify the analysis and enable constructive planning of tasks and timeframes, delivering the sphere of activities assignable to other systems. An extension to other vehicle classes and even a transfer of the concepts developed to other application domains, e.g. in manufacturing, or agriculture will be of great benefit.

Read more : http://www.smart-systems-integration.org/public/documents/publications/EPoSS%20Roadmap_Smart%20Systems%20for%20Automated%20Driving_V2_April%202015.pdf