Table 3: Comparison of tunnel incident detection systems.
Category |
VDS-Based Algorithmsa
|
Video Incident Detection Systemb
|
AADS (This Study) |
Performance |
DR |
85-95% |
14-85% |
80-95% |
FAR |
0.03-0.35% |
15-85% |
2.6-3.5% |
MTTD |
156-306 s |
10-120 s |
1 s |
Operations characteristics |
Calibration |
Difficult |
Difficult |
Easy |
Installation height restraint |
No |
Yes |
No |
Influence of occlusion |
No |
Yes |
No |
Tunnel incident types |
Fire |
X |
O |
X |
Crash |
X |
X |
O |
Wrong-way maneuver |
X |
O |
X |
Debris |
X |
O |
X |
Stopped vehicle |
X |
O |
X |
Queue |
X |
O |
X |
Slow vehicle |
X |
O |
X |
Skid |
X |
X |
O |
Pedestrian |
X |
O |
X |
Change in traffic flow |
|
X |
X |
Note: DR = Detector Rate; FAR = False Alarm Rate; MTTD = Mean Time To Detection; O = Detection is possible; X = Detection is impossible.
aSource: Comparison and Analysis Tool for Automatic Incident Detection, Transportation Research Record, No. 1925, 2005.
bSource: Video Incident Detection Tests in Freeway Tunnels, Transportation Research Record, No. 1959, 2006.