The present study examines expert-defined parameterizations of faults in steer-by-wire systems with regard to controllability and subjective experience of normal drivers. Steer-by-wire systems become increasingly important as prerequisite for automated driving Level 4. A total of six different types of faults and six vehicles were included. This led to the conduction of six individual studies. Each study considered the combination of one vehicle and two different types of faults, respectively, leading to a total of twelve cases. The participant´s task was to repeatedly complete a specified course on a test track consisting of three maneuvers while experiencing a specific fault at a predefined time. Controllability was assessed by means of lane departure from the specified driving corridor within a defined time interval after fault injection. Based on Brockmann (2009) controllability was assumed if 100 % of participants did not leave the driving corridor after fault application. 100 % of participants had to comprise at least n = 20 valid data sets. Neukum scale (Neukum & Krüger, 2003) was used to evaluate subjective perception. Results show that in nine out of twelve cases fault parametrization defined by experts could be controlled by normal drivers. Although the majority of cases in the present study indicate expert judgements to be generally reliable, three cases exhibit potential for the necessity of further tests and readjustments of fault parametrization.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Evaluation methodology of normal driver`s controllability performance challenged with expert-defined steering faults injected in steer-by-wire systems

  • Lotte Saupp,
  • Julia Pelzer,
  • Jörn Lützow,
  • Daniel Wegener,
  • Lutz Eckstein,
  • Stefan Ladwig

摘要

The present study examines expert-defined parameterizations of faults in steer-by-wire systems with regard to controllability and subjective experience of normal drivers. Steer-by-wire systems become increasingly important as prerequisite for automated driving Level 4. A total of six different types of faults and six vehicles were included. This led to the conduction of six individual studies. Each study considered the combination of one vehicle and two different types of faults, respectively, leading to a total of twelve cases. The participant´s task was to repeatedly complete a specified course on a test track consisting of three maneuvers while experiencing a specific fault at a predefined time. Controllability was assessed by means of lane departure from the specified driving corridor within a defined time interval after fault injection. Based on Brockmann (2009) controllability was assumed if 100 % of participants did not leave the driving corridor after fault application. 100 % of participants had to comprise at least n = 20 valid data sets. Neukum scale (Neukum & Krüger, 2003) was used to evaluate subjective perception. Results show that in nine out of twelve cases fault parametrization defined by experts could be controlled by normal drivers. Although the majority of cases in the present study indicate expert judgements to be generally reliable, three cases exhibit potential for the necessity of further tests and readjustments of fault parametrization.