Fuzzy-Wavelet RBFNN Model for Freeway Incident Detection

Abstract

Traffic incidents are nonrecurrent and pseudorandom events that disrupt the normal flow of traffic and create a bottleneck in the road network. The probability of incidents is higher during peak flow rates when the systemwide effect of incidents is most severe. Model-based solutions to the incident detection problem have not produced practical, useful results primarily because the complexity of the problem does not lend itself to accurate mathematical and knowledge-based representations. A new multiparadigm intelligent system approach is presented for the solution of the problem, employing advanced signal processing, pattern recognition, and classification techniques. The methodology effectively integrates fuzzy, wavelet, and neural computing techniques to improve reliability and robustness. A wavelet-based denoising technique is employed to eliminate undesirable fluctuations in observed data from traffic sensors. Fuzzy c-mean clustering is used to extract significant information from the observed data and to reduce its dimensionality. A radial basis function neural network (RBFNN) is developed to classify the denoised and clustered observed data. The new model produced excellent incident detection rates with no false alarms when tested using both real and simulated data.

The Role of Non-Recurring Congestion in Massive Hurricane Evacuation Events

“…This algorithm was compared with the California Algorithm #8 and was found to produce very low false alarms (on the order of 0.07%), as compared to the California algorithm (on the order of 3.82%) under the same detection rate scenarios when tested with simulated data. Limited tests with real data gave 0% false alarms at 95% detection ratio for this algorithm, whereas the California algorithm produced 0.63% false alarms at a 90% detection ratio (Adeli & Karim, 2000;Karim & Adeli, 2002a).…”

“…The use of DWTs (Roy & Abdulhai, 2003) has also been explored for training the PNN, with encouraging results. Adeli and Karim (2000) proposed an algorithm using a Discrete Wavelet Transform (DWT) for noise reduction and feature extraction, followed by a fuzzy c-mean clustering to reduce the dimensionality of the input vector, finally using a Radial Basis Function Neural Network (RBFNN) to classify the input pattern as an incident pattern or a non-incident pattern. Sixteen consecutive data-points for occupancy and speed from the immediate past are used to form the input signal.…”

“…Unlike the previous algorithms (Adeli & Karim, 2000;Ghosh-Dastidar & Adeli, 2003;Karim & Adeli, 2002a, 2002b, 2001, which use the DWT mostly as a tool for de-noising the dataset, this approach implies the direct use of the extracted features in the detection of changes in traffic flow. The difference of occupancies between two stations is used as the input signal.…”

Customization of Automatic Incident Detection Algorithms for Signalized Urban Arterials

“…The relatively new FWRBFNN based freeway incident detection algorithm was developed by Adeli & Karim (2000). The algorithm uses Radial Basis Function Neural Network (RBFNN) architecture in conjunction with wavelet-based denoising and fuzzy-logic based classification techniques.…”

“…[To maintain comparability with other three algorithms, in this paper, the FWRBFNN algorithm uses upstream volume and occupancy as input instead of the upstream occupancy and speed as used by Adeli and Karim (2000). ]…”