U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
| SUMMARY REPORT |
| This summary report is an archived publication and may contain dated technical, contact, and link information |
|
 |
| Publication Number: FHWA-HRT-17-025 Date: December 2017 |
Publication Number: FHWA-HRT-17-025 Date: December 2017 |
As described in chapter 2, CACC experiment 1 included two conditions in which a crash occurred ahead of the vehicle string in which the participant was driving. The participant could not see the crash and therefore could not anticipate that the vehicles ahead would brake hard with a maximum of 1 g sustained deceleration. This crash avoidance event was experienced by the group referred to as the CACC with crash avoidance group. The other group that experienced this event did not have any type of cruise control and was provided with a multifunction display to assist in maintaining a 1.1-s gap. This other group was referred to as the control group.
Five of the 11 participants in the control group collided with the vehicle ahead. Only 1 of 12 participants in the CACC with crash avoidance group collided with the vehicle ahead. The difference in collision experience between the groups was large and statistically reliable. This finding indicates that CACC systems configured as in experiment 1 could be effective in reducing crashes precipitated by the rapid deceleration of vehicles not within the driver’s field of view.
Any one of several factors may have contributed to the difference between the crash probabilities of the CACC-equipped group and the control group. The CACC with crash avoidance vehicles (and all other platoon vehicles other than the lead vehicle) began a 0.4-g deceleration 0.1 s after the platoon lead vehicle initiated a 1-g deceleration. Simultaneously with the onset of the 0.4-g deceleration, the CACC-equipped vehicles sounded an audio alarm. The control group had neither auto-braking nor an auditory alarm. The first cue for control group drivers that they needed to start braking was the looming of the vehicle ahead as it began to decelerate at 0.4 g. The brake lamps on the vehicle ahead were delayed by 1.9 s because these lights were not activated until the simulated driver ahead began to manually brake, which initiated a 1-g rate of deceleration. In addition, the control group was observed to spend more time gazing at the center-stack display, so distraction could not be ruled out as a cause of the higher probability of a crash among control group drivers.
This experiment was conducted to determine the source of the CACC crash probability reduction. This was accomplished by the following:
The independent variable factorial design for this experiment is depicted in table 8.
Auditory Alarm |
0.4-g Automatic Braking |
|
|---|---|---|
No |
Yes |
|
No |
ACC |
CACC-B |
Yes |
CACC-A |
CACC-AB |
In this experiment, the focus was entirely on the final crash avoidance event, which was described in chapter 2. Dependent measures were the following:
If collision was avoided by steering out of lane rather than braking, then reaction time was defined as steering wheel torque greater than 1.125 lbf following onset of deceleration by the vehicle ahead.
The same FHWA driving simulator was used for this experiment as was used for the experiments described in chapters 2 and 3. However, the visual projection system was updated for this experiment. Three projectors were used to provide a 200-degree (horizontal) by 40-degree (vertical) field of view. Each projector provided a nominal resolution of 4,096 by 2,400 pixels. The updated projection system provided a substantial increase in resolution, brightness, and contrast but a slightly narrower horizontal field of view (200 degrees rather than 240 degrees).
The simulator scenario was nearly the same as that described in chapter 1 for the two groups that experienced the crash avoidance event. The only change was that the crash avoidance event occurred after 20 min of driving rather than 38 min.
The study had 112 participants, 28 in each of four groups: control (ACC), CACC with auditory alarm and 0.4-g braking (CACC-AB), CACC with auditory alarm but no braking (CACC-A), and CACC with 0.4-g braking but no alarm (CACC-B). Individuals who participated in the CACC experiment 1 were excluded from participation in experiment 3. To roughly balance the groups on participant age, half the recruits in each experimental group were under the age of 47. Each condition and age grouping consisted of equal numbers of males and females.
The following subsections describe the results of the experiment for crashes, reaction time, and adjusted TTC.
Table 9 shows the number of crashes and crashes avoided by each group. Also shown in the table are maximum likelihood estimates of crash probability and the 95-percent confidence limit for those estimates. The probability of a crash was lower with the full CACC system (CACC-AB) compared with the other groups, which did not differ from each other in crash probability. The effect of condition was significant (
2(3) = 10.6, p = 0.01). Post hoc testing showed that only the CACC-AB group significantly differed from the ACC group (p = 0.003).
nc = not computed.
The reaction times to the onset of the crash event are shown in figure 15. Three participants in the CACC-B group never reacted and therefore were not included in the reaction time analysis. The condition effect was significant (p < 0.0001). Post hoc testing showed that the ACC group’s mean reaction time did not differ significantly from the CACC-AB group mean but that all the other group mean comparisons yielded significant differences.
Note: Error bars represent estimated 95-percent confidence limits of the means.
Figure 15. Graph. Reaction time from onset of braking by platoon-lead vehicle.
The TTC findings are displayed in figure 16. The findings are based on a sample size of 92 participants. The remaining 20 participants had uninterpretable adjusted TTC estimates. Three of those 20 had no reaction and never applied the brakes. The remaining 17 participants had uninterpretable adjusted TTC values because they were decelerating at a rate less than that of the lead vehicle (also decelerating) at the time of impact, thereby generating adjusted minimum TTC values of negative infinity. None of the participants with full CACC capabilities (i.e., CACC-AB) had to be excluded from this analysis, and only one participant in the CACC-B group had to be excluded. The participants who failed to brake at all or were decelerating less at the time of collision than the preceding vehicle were evenly distributed between the ACC and CACC-A groups.
Note: Error bars represent estimated 95-percent confidence limits of the means.
Figure 16. Graph. TTC results.
The effect of condition was significant (p = 0.04). As can be seen in figure 16, the CACC-AB group had a substantial positive adjusted TTC (i.e., on average, members of this group have almost 0.6 s extra time to respond to the collision event). The ACC and CACC-B groups had significantly lower mean adjusted TTC values than the CACC-AB group. The CACC-A group mean was not significantly different from any of the other three group means.
Experiment 3 reinforces the main conclusion of experiment 1; a full CACC system (i.e., as configured for the CACC-AB group) has the potential to provide a substantial safety benefit. The control condition in experiment 3 did not have an in-vehicle display or the requirement to frequently monitor the speedometer—two potential explanations for the high crash rate of the experiment 1 control condition. Nonetheless, the crash rate for the ACC condition in experiment 3 was nearly identical to that in experiment 1. This suggests that it was CACC automatic braking and alarm that provided the apparent safety benefit in both experiments. Removing either the alarm or the automatic braking from the CACC system diminished or eliminated the safety benefit of the full system.
It is not clear from these results why the absence of an auditory alarm (ACC and CACC-B) condition resulted in an increased crash risk. The CACC-B group had the longest reaction times and had the three incidences of no response. The ACC group also had no alarm, yet it reacted as quickly as the group with full CACC. Perhaps this is an example of over trust in the system. The CACC-B braking force was twice that of the ACC braking (0.4 g versus 0.2 g), so it is conceivable that the CACC-B group felt the system responding and trusted the automated response until it was too late to recover. The mild braking in the ACC condition may have been easier to perceive as inadequate compared with the more aggressive braking in the CACC-B condition.
The CACC-A group, which received an auditory alarm but had no automated braking, responded more quickly than any group but still had a high crash rate. The extra time provided by the 0.4-g automated braking to the CACC-AB group appears to have been the key to enabling that group to respond more slowly while retaining an average of a 0.6-s cushion in extra time available.
The alarm mitigated the apparent over trust of the system, and the automated braking feature provided drivers with the extra time they needed to respond. Whether the combination of alarm and automated braking would be effective with other CACC implementations, such as with shorter gaps between vehicles or more aggressive automated deceleration, remains to be explored.
