Introduction
C-ITS systems that warn the motorcycle rider of an upcoming danger only work if the rider is interpreting the warning correctly and reacts accordingly. So far however, there is little knowledge about how long a rider reaction towards a warning takes. Additionally, the question arises whether reactions from the passenger car domain can be applied to PTWs.
This whitepaper describes a dynamic motorcycle riding simulator study by CMC, which investigated motorcycle riders’ reaction times towards a warning on the dashboard. This study is a first step towards empirical evidence in this domain.
Such knowledge can bridge the gap between results from the accidentology side to the use case and test case specific strategies. The latter focus on the decision on how an application’s display/ alert principle should be designed (e.g., advisory notification, crash warning, active intervention).
This whitepaper describes a dynamic motorcycle riding simulator study by CMC, which investigated motorcycle riders’ reaction times towards a warning on the dashboard. This study is a first step towards empirical evidence in this domain.
Such knowledge can bridge the gap between results from the accidentology side to the use case and test case specific strategies. The latter focus on the decision on how an application’s display/ alert principle should be designed (e.g., advisory notification, crash warning, active intervention).
What has been investigated
The dashboard warning was a generic visual warning which can act as a benchmark to improve upon in the future.
Reactions in an urban and a rural scenario were tested. These did not include imminent crash warnings, but advisory warnings with 3 seconds between warning onset and the potentially critical situation becoming visible.
The results of this study can be used in the following ways:
1. Based on the temporal evolvement and Time-to-Collision of different accident scenarios, the results help to better estimate for which C-ITS applications running on a PTW, a purely visual warning could be appropriate, and for which ones not.
2. Any OEM’s individual HMI solution, i.e., rider notification concept, should result in faster reaction times and less missed warnings in this test setup than the conservative rider notification assessed in this study.
3. The distribution of rider reaction times can clarify to which extent results from passenger car research are applicable to the PTW domain and serve as an input to parameterize rider behaviour models in traffic simulations necessary for the effectiveness estimation of (C-ITS) safety applications.
Reactions in an urban and a rural scenario were tested. These did not include imminent crash warnings, but advisory warnings with 3 seconds between warning onset and the potentially critical situation becoming visible.
The results of this study can be used in the following ways:
1. Based on the temporal evolvement and Time-to-Collision of different accident scenarios, the results help to better estimate for which C-ITS applications running on a PTW, a purely visual warning could be appropriate, and for which ones not.
2. Any OEM’s individual HMI solution, i.e., rider notification concept, should result in faster reaction times and less missed warnings in this test setup than the conservative rider notification assessed in this study.
3. The distribution of rider reaction times can clarify to which extent results from passenger car research are applicable to the PTW domain and serve as an input to parameterize rider behaviour models in traffic simulations necessary for the effectiveness estimation of (C-ITS) safety applications.
Important outcomes
The following interesting outcomes could be observed:
• In 16.7% of cases, the purely visual warning was not recognized at all.
• Among the other cases, the average time between onset of the notification and gaze towards the dashboard was about 1 second already.
• The average time between notification onset and ‘throttle off’ was about 2 seconds.
• The average time between notification onset and ‘initiate braking’ was about 2.5 seconds.
• The mentioned reaction times were shorter in the urban scenario compared to the rural one, where the situation was perceived as less critical.
Another interesting observation could be that, in the more time-critical urban scenario, all riders who had seen the warning, initiated braking before the obstacle became visible. In combination with the favourable evaluation of the test riders after the experiment, this shows a good potential for the safety benefit of C-ITS applications.
In comparison to driver reaction times in passenger car studies, more missed warnings were observed, reaction times seem longer and reaction time distributions seem wider; hence there is a clear need for PTW-specific reaction time studies.
• In 16.7% of cases, the purely visual warning was not recognized at all.
• Among the other cases, the average time between onset of the notification and gaze towards the dashboard was about 1 second already.
• The average time between notification onset and ‘throttle off’ was about 2 seconds.
• The average time between notification onset and ‘initiate braking’ was about 2.5 seconds.
• The mentioned reaction times were shorter in the urban scenario compared to the rural one, where the situation was perceived as less critical.
Another interesting observation could be that, in the more time-critical urban scenario, all riders who had seen the warning, initiated braking before the obstacle became visible. In combination with the favourable evaluation of the test riders after the experiment, this shows a good potential for the safety benefit of C-ITS applications.
In comparison to driver reaction times in passenger car studies, more missed warnings were observed, reaction times seem longer and reaction time distributions seem wider; hence there is a clear need for PTW-specific reaction time studies.