Neural network incident detection on arterials using fusion of simulated probe vehicle and loop detector data

This paper describes the development of neural network models for Automatic Incident Detection (AID) on arterials, using simulated data derived from Inductive Loop Detectors (ILDs) and probe vehicles. This study extends previous research by comparing the performance of various neural network architectures for data fusion and by providing a comparison of model performance for various probe vehicle penetration rates and detector configurations. Data from 108 incidents was collected from ILDs and probe vehicles at two locations on a previously validated network for two detector configurations. Configuration 1 was similar to a freeway link, while Configuration 2 conformed to the standard configuration on road networks. The best performance obtained for Configuration 1 was a Detection Rate (DR) of 59% for a False Alarm Rate (FAR) of 0.5%, for a probe vehicle penetration rate of 20%. The best performance obtained for Configuration 2 was a DR of 86% for a FAR of 0.36% for a probe vehicle penetration rate of 20%. Satisfactory performance can also be obtained using ILD data alone (DR = 86% for FAR = 0.41%). Inclusion of speed data further improves performance (DR = 90% for FAR = 0.5%) and its use when available is highly recommended. This research demonstrates the feasibility of developing a neural network model for detection of incidents on arterials using loop and probe vehicle data. Options for further study are also presented.

[1]  Kyriacos C. Mouskos,et al.  Transportation Operations Coordinating Committee System for Managing Incidents and Traffic: Evaluation of the Incident Detection System , 1999 .

[2]  Hussein Dia,et al.  Development and evaluation of neural network freeway incident detection models using field data , 1997 .

[3]  Hussein Dia,et al.  Assessment of incident-induced impacts on the performance of an arterial network , 2001 .

[4]  Timothy Masters,et al.  Practical neural network recipes in C , 1993 .

[5]  N. Cottman.  Modelling the impacts of intelligent transport systems using microscopic traffic simulation , 2002 .

[6]  Frank S. Koppelman,et al.  USE OF VEHICLE POSITIONING DATA FOR ARTERIAL INCIDENT DETECTION , 1996 .

[7]  John N. Ivan,et al.  REAL-TIME DATA FUSION FOR ARTERIAL STREET INCIDENT DETECTION USING NEURAL NETWORKS , 1995 .

[8]  John N. Ivan,et al.  Data Fusion of Fixed Detector and Probe Vehicle Data for Incident Detection , 1998 .

[9]  H. F. Dia,et al.  Incident detection on arterials using neural network data fusion of simulated probe vehicle and loop detector data , 2004 .

[10]  Jose C. Principe,et al.  Neural and adaptive systems , 2000 .

[11]  John N. Ivan,et al.  Vehicle-based versus fixed-location measurements for traffic surveillance in IVHS , 1995, Other Conferences.

[12]  John N. Ivan,et al.  Incident detection using vehicle-based and fixed-location surveillance , 1997 .

[13]  John N. Ivan.  Neural network representations for arterial street incident detection data fusion 1 1 The contents o , 1997 .