Filtern
Dokumenttyp
- Konferenzveröffentlichung (14) (entfernen)
Schlagworte
- Error (14) (entfernen)
Institut
- Sonstige (11)
- Abteilung Fahrzeugtechnik (4)
- Abteilung Straßenbautechnik (1)
Ziel der Forschungsarbeiten zum Thema Durchsickerung von Straßenböschungen ist es, ausgewogene Beurteilungskriterien für einen umwelteffizienten und zugleich ökonomischen Einsatz von Ersatzbaustoffen und Bodenmaterial zu schaffen. Durch eine belastbare Datengrundlage zur Erfassung des Wasserhaushaltes des gesamten Bauwerkes können Instrumente entwickelt werden, um die Wirksamkeit der unterschiedlichen technischen Sicherungsmaßnahmen zu beurteilen. Um dieses Ziel zu erreichen, verfolgt die Bundesanstalt für Straßenwesen verschiedene Projektansätze zum Thema Durchsickerung von Straßenböschungen. Jeder Projektansatz hat seine Stärken und Schwächen. Durch großmaßstäbliche Versuche können die realen Verhältnisse gut wiedergegeben werden. Solche Versuche sind jedoch personalintensiv und langwierig. Schneller und preisgünstiger sind Modellrechnungen. Die Qualität der Berechnungen hängt aber stark von der Qualität der gewählten Parameter ab. Durch inverse Modellierung müssen die Eingabeparameter der Programme zur Modellrechnung an die Realität angepasst werden. Die realen Bedingungen werden am besten durch ein Pilotprojekt an einem realen Versuchsdamm wiedergegeben. Diese Versuche sind sehr aufwendig und sehr kostenintensiv, aber zur Validierung der Modellrechnung und der großmaßstäblichen Versuche erforderlich. In dem Betrag wird beschrieben, welche großmaßstäblichen Versuche durch die Bundesanstalt für Straßenwesen (BASt) durchgeführt wurden und wie sie von Anfang an durch Simulationsrechnungen begleitet wurden.
Per definition, SAE Level 2 (L2) Systems perform both the lateral and longitudinal vehicle motion control with the expectation that the driver completes the Object and Event Detection and Response (OEDR). Since every system performs also parts of the OEDR itself and this amount of OEDR also varies between different L2 systems depending on the intended system design, it cannot be taken for granted that drivers automatically understand their roles and responsibilities in interaction with the system. Especially highly reliable L2 systems performing a greater amount of OEDR while at the same time requiring only little driver input over time can make it difficult for drivers to correctly identify their role and responsibility. Until now, neither application-oriented assessment methods nor design guidelines for OEDR related system design features taking safety of human-machine-interaction into account are available. The objective is therefore to deliver a standardized tool for the assessment of human-machine-interaction-related safety of vehicles with L2 systems currently available on the market. To evaluate the impact of different system design aspects on safety of human-machine-interaction and also to be able to differentiate between system designs, a holistic, standardized and application-oriented assessment procedure is proposed. The novel tablet-based assessment tool focuses not only on available standards and guidelines but measures also concrete user behaviour and user understanding in interaction with the L2 systems. The aim is to gain further insights which cannot be measured directly by simple checklist instruments. For preparation, based on international standards, literature reviews and expert consultations, a first checklistbased expert-evaluation for currently available vehicles with L2 systems was developed. These assessments are focusing on different sources of user information (e.g. user manual), human-machine-interface design as well as the prevention of unintended use by different driver monitoring techniques. The checklist-tool was developed in cooperation with experts of different EuroNCAP test laboratories and validated in a common expert workshop to gain high level of standardization and agreement. However, to assess safety of human-machine-interaction holistically beyond these rather explicit forms of information design criteria, also implicit forms of drivervehicle-communication via vehicle dynamics, functional behavior or reliability play an important role and should be taken into account. Therefore, the main and novel methodological aim is to consider also interaction related processes regarding user´s understanding of roles and responsibilities when applying automated driving functions as well as user´s awareness of automation modes or traffic situations in the modular tablet-based assessment tool.
From literature well-known analyzes on risks, hazards and causes of accidents of older drivers are amended by the present study in which a comparison of the specific features of accident causes of older car drivers (older than 60 years) and of younger car drivers (under 25 years) is conducted. Mainly the question is pursued if specific errors, mistakes and lapses are predominant in the two different age groups. The analysis system ACAS (Accident Causation Analysis System) used hereby consists of a sequential system of accident causation factors from the human, the technical and the infrastructural field, whereupon for this study the influence of the human features on the accident development in two different age groups is of interest. ACAS is both an accident model and an analysis and classification system, which describes the human participation factors of an accident and their causes in the temporal sequence (from the perceptibility to concrete action errors) taking into consideration the logical sequence of individual basic functions. In five steps (categories) of a logical and temporal sequence the hierarchical system makes human functions and processes as determinants of accident causes identifiable. The methodology specifically focuses on the use in so-called "In-Depth" and "On-Scene" investigation studies. With the help of the system for each accident participant one or more of five hypotheses of human cause factors are formed and then specified by appropriate verification criteria. These hypotheses in turn are further specified by indicators in such manner that the coding of the causation factors by a code system meets the needs of database processing and are accessible to a quantitative data analysis. The first results of the descriptive comparison of the two age groups concern mainly differences in the functional levels "information admission/perception" (where the elderly drivers have more difficulties than the young ones) and "information processing/evaluation" (where the younger drivers show more problems). Concerning the cognitive function of "planning" the group of younger drivers seems to be more often involved in an accident because of excessive speed.
The effect of fatigue on driving has been compared to the effect of alcohol impairment in both driver performance and crash studies. However are crash characteristics and causation mechanisms similar in crashes involving fatigue to those involving alcohol when studied in the real world? This has been explored by examining data held in the EC project SafetyNet Accident Causation Database. Causation data was recorded using the SafetyNet Accident Causation System (SNACS). The focus was on Cars/MPV crashes and drivers assigned the SNACS code Alcohol or Fatigue. The Alcohol group included 44 drivers and the Fatigue group included 47. "Incorrect direction" was a frequently occurring critical event in both the Alcohol and Fatigue groups. The Alcohol group had more contributory factors related to decision making and the Fatigue group had more contributory factors relating to incorrect observations. This analysis does not allow for generalised statements about the significance of the similarities and differences between crashes involving alcohol and fatigue, however the observed differences do suggest that attempts to quantify the effect of fatigue by using levels of alcohol impairment as a benchmark should be done with care.
The misuse of CRS (child restraint system) is one of the most urgent problems in connection of child safety in cars. Numerous field studies show that more than two thirds of all CRS are used in a wrong way. This misuse could lead to serious injuries for the children. Surprisingly the quality of CRS use is coded much better in accident data (e.g. GIDAS) than the results of observatory field studies show. It is expected that misuse of CRS was not detected by the accident teams in a large number of the cases. An essential part in improving child seats and their usability is the knowledge of the relation between misuse and resulting injuries. For that the analysis and experimental reconstruction of accidents is an important part. For allowing an exact experimental accident reconstruction, it is necessary to have detailed information about the securing situation of the child and about the installation of the CRS in the car.
For a number of EU regulatory acts Virtual Testing (VT) is already allowed for type approval (see Commission Regulation No. 371/2010 of 16 April 2010 amending the Framework Directive 2007/46/EC). However, only a very general procedure on how to apply VT for type approval is provided. Technical details for specific regulatory acts are not given yet. The main objective of the European project IMVITER (IMplementation of VIrtual TEsting in Safety Regulations) was to promote the implementation of VT in safety regulations. When proposing VT procedures the new regulation was taken into account, in particular, addressing open issues. Special attention was paid to pedestrian protection as pilot cases. A key aspect for VT implementation is to demonstrate that the employed simulation models are reliable. This paper describes how the Verification and Validation (V&V) method defined by the American Society of Mechanical Engineers was adapted for pedestrian protection VT based assessment. or the certification of headform impactors an extensive study was performed at two laboratories to assess the variability in calibration tests and equivalent results from a set of simulation models. Based on these results a methodology is defined for certification of headform impactor simulation models. A similar study was also performed with one vehicle in the type approval test setup. Its bonnet was highly instrumented and subjected to 45 impacts in five different positions at two laboratories in order to obtain an estimation of the variability in the physical tests. An equivalent study was performed using stochastic simulation with a metamodel fed with observed variability in impact conditions of physical headforms. An estimation of the test method uncertainty was obtained and used in the definition of a validation corridor for simulation models. Validation metric and criteria were defined in cooperation with the ISO TC22 SC10 and SC12 WG4 "Virtual Testing". A complete validation procedure including different test setups, physical magnitudes and evaluation criteria is provided. A detailed procedural flowchart is developed for VT implementation in EC Regulation No 78/2009 based on a so called "Hybrid VT" approach, which combines real hardware based head impact tests and simulations. This detailed flowchart is shown and explained within this paper. Another important point within the virtual testing based procedures is the documentation of relevant information resulting from the verification and validation process of the numerical models used. For this purpose report templates were developed within the project. The proposed procedure fixes minimum V&V requirements for numerical models to be confidently used within the type-approval process. It is not intended to be a thorough guide on how to build such reliable models. Different modeling methodologies are therefore possible, according to particular OEM know-how. These requirements respond to a balance amongst the type-approval stakeholders interests. A cost-benefit analysis, which was also performed within the IMVITER project, supports this approach, showing the conditions in which VT implementation is beneficial. Based on the experience gained in the project and the background of the experts involved an outlook is given as a roadmap of VT implementation, identifying the most important milestones to be reached along the way to a future vehicle type approval procedure supported by VT. The results presented in this paper show an important step addressing open questions and fostering the future acceptance of virtual testing in pedestrian protection type approval procedures.
Recent findings from real-world accident data have shown that fatality risks for pedestrians are substantially lower than generally reported in the traffic safety literature. One of the keys to this insight has been the large and random sample of car-to-pedestrian crashes available in the German In-Depth Accident Study (GIDAS). Another key factor has been the proper use of weight factors in order to adjust for outcome-based sampling bias in the accident data. However, a third factor, a priori of unknown importance, has not yet been properly analysed. This is the influence of errors in impact speed estimation. In this study, we derived a statistical model of the impact speed errors for pedestrian accidents present in the GIDAS database. The error model was then applied to investigate the effect of the estimation error on the pedestrian fatality risk as a function of car impact speed. To this end, we applied a method known as the SIMulation-EXtrapolation (SIMEX) method. It was found that the risk curve is fairly tolerant to some amount of random measurement error, but that it does become flattened. It is therefore important that the accident investigations and reconstructions are of high quality to assure that systematic errors are minimised and that the random errors are under control.
Within the process of integrating passenger airbags in the vehicle fleet a problem of compatibility between the passenger airbag and rear-facing child restraint systems was recognised. Especially in the US several accidents with children killed by the passenger airbag were recorded. Taking into account these accidents the deactivation of a present passenger airbag is mandatory if a child is carried in a rear-facing child restraint system at the front passenger seat in all member states of the European Union. This rule is in force since the deadline of 2003/20/EC at the latest. In the past a passenger airbag either could not be disabled or could only be disabled by a garage. Today there are a lot of different possibilities for the car driver himself to disable the airbag. Solutions like an on/off-switch or the automatic detection of a child restraint system are mentioned as an example. Taking into account the need for the deactivation of front passenger airbags two types of misuse can occur: transportation of an infant while the airbag is (still) enabled and transportation of an adult, while the airbag is disabled, respectively. Within a research project funded by BASt both options of misuse were analysed utilising two different types of surveys amongst users (field observations and interviews, Internet-questionnaires). In addition both analysis of accident data and crash tests for an updated assessment of the injury risk caused by the front passenger airbag were conducted. Both surveys indicate a low risk of misuse. Most of the misuse cases were observed in older cars, which offer no easy way to disable the airbag. For systems, which detect a child seat automatically, no misuse could be found. The majority of misuses in cars equipped with a manual switch were caused by reasons of oblivion. Also the accident analysis indicates a minor risk of misuse. From more than 300 cases of the GIDAS accident sample that were analysed, only 24 children were using the front passenger seat in cars equipped with a front passenger airbag. In most of these cases the airbag was deactivated. When misuse occurred the injury severity was low. However, when analysing German single accidents the fatality risk caused by the front passenger airbag became obvious. From the technical point of view, there were important changes in the design of passenger airbags in recent years. Not only volume and shape were modified, but also the mounting position of the entire airbag module was changed fundamentally. Even if these findings do not allow obtaining general conclusions, a clear tendency of less danger by airbags could be identified. For future vehicle development a safe combination of airbags and rear faced baby seats seems to be possible in the long term. This would mean that both types of misuse could be eliminated. For parents an easier use of child seat and car would be the result.
As the official German catalogue of accident causes has difficulty in matching the increasing demands for detailed psychologically relevant accident causation information, a new system, based on a "7 Steps" model, so called ACASS, for analyzing and collecting causation factors of traffic accidents, was implemented in GIDAS in the year 2008. A hierarchical system was developed, which describes the human causation factors in a chronological sequence (from the perception to concrete action errors), considering the logical sequence of basic human functions when reacting to a request for reaction. With the help of this system the human errors of accident participants can be adequately described, as the causes of each range of basic human functions may be divided into their characteristics (influence criteria) and further into specific indicators of these characteristics (e.g. distraction from inside the vehicle as a characteristic of an observation-error and the operation of devices as an indication for distraction from inside the vehicle. The causation factors accordingly classified can be recorded in an economic way as a number is assigned to each basic function, to each characteristic of that basic function and to each indicator of that characteristic. Thus each causation factor can be explicitly described by means of a code of numbers. In a similar way the causation factors based on the technology of the vehicle and the driving environment, which are also subdivided in an equally hierarchical system, can be tagged with a code. Since the causes of traffic accidents can consist of a variety of factors from different ranges and categories, it is possible to tag each accident participant with several causation factors. This also opens the possibility to not only assign causation factors to the accident causer in the sense of the law, but also to other participants involved in the accident, who may have contributed to the development of the accident. The hierarchical layout of the system and the collection of the causation factors with numerical codes allow for the possibility to code information on accident causes even if the causation factor is not known to its full extent or in full detail, given the possibility to code only those cause factors, which are known. Derived from the systematic of the analysis of human accident causes ("7 steps") and from the practical experiences of on-scene interviews of accident participants, a system was set in place, which offers the possibility to extensively record not only human causation factors in a structured form. Furthermore, the analysis of the human causation factors in such a structured way provides a tool, especially for on-scene accident investigations, to conduct the interview of accident participants effectively and in a structured way.
The aim of this study was to evaluate the performance and accuracy of Event Data Recorders (EDRs). The analysis was based on J-NCAP crash tests from 2006"2007, with the corresponding EDR datasets. The pre-crash velocity, maximum delta-V and delta-V versus time history data recorded in the EDRs were compared with the reliable crash test data. The difference between the EDR pre-crash velocity and the laboratory test speed was less than 4 percent. In contrast, in several cases the maximum delta-V and delta-V versus time history data obtained from the EDRs showed uncertainty of measurement in comparisons with the reliable delta-V data. The difference in maximum delta-V in these comparisons was more than 5 percent in 10 of 14 tests and more than 10 percent in 4 of 14 tests. The EDRs underestimated the maximum delta-V in almost all tests. It was also concluded that the calculated acceleration from the EDR delta-V versus time history data showed good agreement with the instrumented accelerometer signal during the collision in almost all tests.