In previous research, measures like collision probability, driving errors and response times have been used to describe and quantify human safety performance. Human performance in traffic can be studied by analysing accident statistics or driving behaviour in real or simulated traffic environments. To measure human performance, task demands have to be defined and human capability has to be determined. In summary, driving covers various subtasks which require specific abilities (see Fig. If driving a vehicle for the first time, stabilisation might be more regulated based on knowledge or rules.
BENCHMARK AUTOMOTIVE DRIVERS
For example, to stabilise a vehicle, experienced drivers will mostly regulate their driving behaviour based on highly automated skills. Depending on the task, drivers regulate their behaviour based on knowledge, rules or skills. The different subtasks demand different competences. Stabilisation requires drivers to keep the vehicle in a steady state by, for example, braking and steering. Navigation refers to the planning of an appropriate route, whereas guidance involves adapting driving behaviour to the course of the road, traffic conditions and traffic rules. The driving task consists of navigation, guidance and stabilisation according to the three-level model of the vehicle driving task.
Task demands are composed of environmental factors (e.g., weather and road conditions), the vehicle and the driving task itself. The driver’s capability results from skills and abilities, and is influenced by human factors (e.g., emotions, inattention or stress). If not, control is lost and an accident is highly possible.
BENCHMARK AUTOMOTIVE DRIVER
If capability excels task demands, the driver controls the situation. In case of driving, the task-capability interface model considers human performance as a function of the driver’s capability and the demands of the driving task. In general, human performance refers to the potential of a person to successfully perform a task. Section 5 discusses the results and the presented approach before drawing conclusions in Sect. 3 and human performance is quantified in Sect.
BENCHMARK AUTOMOTIVE SIMULATOR
The driving simulator study is described in Sect.
2 human performance is defined, followed by a short summary of existing approaches to measure human performance (2.1 and 2.2) and an introduction of the approach presented in this article (2.3). This article is structured as follows: in Sect. To facilitate HAVs’ safety assessment, this work presents a scenario-based method to quantify the limit of human performance, which was developed as part of the project PEGASUS. However, appropriate methods to quantify and compare human and HAV performance are rare. Following the request that HAVs should only be released if they are statistically safer than human drivers made by the German Federal Ministry of Transport and Digital Infrastructure, human performance should be used as a benchmark for HAV performance (at least in Germany).
But before HAVs become part of everyday traffic, their safety performance-especially when confronted with our constantly changing environment-has to be tested and verified. Highly automated vehicles (HAVs SAE level 3 or higher ) allow the human driver to attend to a task other than the driving task. Autonomous driving is associated with many potential advantages, e.g., increased traffic efficiency and safety benefits.
0 kommentar(er)
