Researchers in Holland have constructed an evaluation system for the design of automobile automated driving system, and the reliability of this evaluation system has been tested through different driving situations and the experimental research on different system design schemes.
Beukel and Voort from University of Twente have published this latest research in Applied Ergonomics (an authoritative journal of ergonomics). And a driving simulator with 180° visual angle and three screens is applied in the experiment, through which the driving situation on the expressway can be simulated.
The three indicators involved in this evaluation system of the research institute refer to the strength of situation awareness (SA), degree of accident avoidance (AA) and acceptability of the scheme (CA) during the driving process.
In the research, we found 24 users with at least one year of driving experience to complete the simulated driving task under different situations (attention situation, intervention situation) and different system design schemes (low, medium and high level of information richness).
Introduction to the Two Situations
Attention Situation: no accident will occur; additional attention of the user is required, but there is no need for the user to intervene in the system to operate;
Intervention Situation: the user has to intervene in the system to avoid accidents.
Introduction to the Three Specific Schemes
Scheme A: only provide the alarm;
Scheme B: provide the alarm and the tip text will appear on the interface;
Scheme C: provide the alarm and the tip text and graphical warning message will appear on the interface; then the location and surroundings of the car will be displayed in real-time.
The research result shows that:
1. Scheme C, which has a high level of information richness, is the optimal in attention situation, but it is the worst in intervention situation.
2. Scheme A, which has a low level of information richness, is the worst in attention situation.
3. The usability of Scheme C has obtained the highest evaluation from the user.
The researchers explain the results as follows:
1. Scheme C can provide the most information so that the user feels it is available. But on the other hand, it will take a lot of time to digest such information, so the user will be influenced and cannot timely take action to avoid accidents in intervention situation.
2. If the user cannot determine which kind of information is really helpful, he/she will select the system with the most detailed feedback as the "safest choice", showing that the potential problems may exist if we only consider the conclusion obtained from the subjective indicator data of users' self-evaluation.
3. The evaluation system of this research can quickly and effectively identify the worst performing design , but such function does not work on the optimal performing design.
1. When designing the automobile automated driving system, the design shall be simplified and the redundant information shall be reduced if the driving situation which needs intervention to avoid accidents frequently appears, while the system shall be designed to provide abundant information if the attention situation with little need of intervention appears frequently.
2. When testing, it is not recommended to only consider the subjective indicators. The reliability of the evaluation result can be improved based on the objective indicators.
3. The evaluation system of this research can quickly and effectively identify the worst performing design, which will facilitate the selection of scheme design phase.
Arie P. van den Beukel, Mascha C. van der Voort(2017).How to assess driver’s interaction with partially automated driving systems——A framework for early concept assessment.Applied Ergonomics,59,302-312.