论文部分内容阅读
Abstract: In real-life freeway transportation system, a few number of incident observation (very rare event) is available while there are large numbers of normal condition dataset. Most of researches on freeway incident detection have considered the incident detection problem as classification one. However, because of insufficiency of incident events, most of previous researches have utilized simulated incident events to develop freeway incident detection models. In order to overcome this drawback, this paper proposes a wavelet-based Hotelling T2 control chart for freeway incident detection, which integrates a wavelet transform into an abnormal detection method. Firstly, wavelet transform extracts useful features from noisy original traffic observations, leading to reduce the dimensionality of input vectors. Then, a Hotelling T2 control chart describes a decision boundary with only normal traffic observations with the selected features in the wavelet domain. Unlike the existing incident detection algorithms, which require lots of incident observations to construct incident detection models, the proposed approach can decide a decision boundary given only normal training observations. The proposed method is evaluated in comparison with California algorithm, Minnesota algorithm and conventional neural networks. The experimental results present that the proposed algorithm in this paper is a promising alternative for freeway automatic incident detections.
Key words: Freeway incident, incident detection algorithms, Hotelling T2 control chart, wavelet transforms, feature selection.
1. Introduction
The automated detection of incidents on the freeway is essentially one of the primary functions of an Advanced Traffic Management Systems (ATMS), which is the important part of Intelligent Transportation Systems (ITS). Accurate and rapid incident detection is particularly important to improve the capacity of freeways because incident causes severe reduction in the traffic flow or a secondary incident, which may result in sacrificing many human life and cost. Numerous freeway incident detection algorithms were developed in the last several decades. These algorithms include California algorithm [1], Minnesota algorithm [2], neural network based algorithm [3-6], nonparametric regression [7], Support Vector Machine (SVM) [8], decision tree learning [9], and wavelet transform-based algorithm [10-13].
However, because of traffic data dependence and high False Alarm Rate (FAR) of existing algorithms, most of traffic information centers are not operating them. Recently, a wavelet transform, which is one of the promising techniques in the field of signal analysis and data mining, has been applied for traffic incident detection [12]. Because of the superior ability of denosing, wavelet transforms have been used to extract informative features from noisy original traffic measurements before using input vectors into incident detection algorithms. Several recent researches [14] based on the wavelet technique showed that the wavelet coefficients (finer and coarse level coefficients) were directly utilized in detecting changes in traffic measurements. For deciding decision-making rule for incident detection, the finest level coefficients were employed to detect significant and abrupt changes in traffic measurement while the coarse level coefficients were taken advantage to detect global incident trends[14]. The upstream traffic measurements of occupancies and traffic volumes will increase and downstream ones will decrease after the occurrence of incidents. The traffic speeds indicate an opposite traffic phenomenon. However, most of previous researches for freeway incident detection consider incident detection problems as classification ones where lots of incident traffic measurements are needed to construct an incident detection model. Their need for much of incident measurements has kept them from a wide deployment at ITS traffic management and control centers. Therefore, this research proposes a novel freeway incident detection algorithm of the wavelet-based Hotelling T2 control chart, which utilizes multi-resolution property in wavelet transforms and statistical quality control model with only normal training dataset. Unlike the existing methods, which consider incident detection problems as classification ones, the proposed approach is applicable even to the cases when only normal traffic measurements are available. The authors compare the performance of the proposed algorithm with that of existing popular methods based on both real-life datasets.
This paper is organized as follows: A brief review of wavelet transform is given in section 2; The proposed methodology is described in section 3, followed by experimental results in section 4; finally, section 5 concludes this paper with a brief summary and discussion in the future works.
References
[1] H.J. Payne, S.C. Tignor, Freeway incident-detection algorithms based on decision trees with states, Transportation Research Board (682) (1978) 30-37.
[2] Y.J. Stephanedes, A.P. Chassiakos, Application of filtering techniques for incident detection, Journal of Transportation Engineering 119 (1993) 13-26.
[3] H. Dia, G. Rose, Development and evaluation of neural network freeway incident detection models using field data, Transportation Research Part C: Emerging Technologies 5 (1997) 313-331.
[4] S. Ishak, H. Al-Deek, Performance of automatic ANN-based incident detection on freeways, Journal of Transportation Engineering 125 (1999) 281-290.
[5] A. Samant, H. Adeli, Enhancing neural network traffic incident-detection algorithms using wavelets, Computer-Aided Civil and Infrastructure Engineering 16(2001) 239-245.
[6] D. Srinivasan, X. Jin, R.L. Cheu, Adaptive neural network models for automatic incident detection on freeways, Neurocomputing 64 (2005) 473-496.
[7] S. Tang, H. Gao, Traffic-incident detection-algorithm based on nonparametric regression, IEEE Transactions on Intelligent Transportation Systems 6 (2005) 38-42.
[8] F. Yuan, R.L. Cheu, Incident detection using support vector machines, Transportation Research Part C: Emerging Technologies 11 (2003) 309-328.
[9] S. Chen, W. Wang, Decision tree learning for freeway automatic incident detection, Expert Systems with Applications: An International Journal 36 (2009) 4101-4105.
[10] A. Samant, H. Adeli, Feature extraction for traffic incident detection using wavelet transform and linear discriminant analysis, Computer-Aided Civil and Infrastructure Engineering 15 (2000) 241-250.
[11] A. Karim, H. Adeli, Incident detection algorithm using wavelet energy representation of traffic patterns, Journal of Transportation Engineering 128 (2002) 232-242.
[12] S. Ghosh-Dastidar, H. Adeli, Wavelet-clustering-neural network model for freeway incident detection, Computer-Aided Civil and Infrastructure Engineering 18(2003) 325-338.
[13] X. Wang, J. Zhang, A traffic incident detection method based on wavelet Mallat algorithm, in: Proceedings of the 2005 IEEE Mid-Summer Workshop on Soft Computing in Industrial Applications, Espoo, Finland, 2005, pp. 166-172.
[14] Y.S. Jeong, M. Castro-Neto, M. Jeong, L. Han, A wavelet-based freeway incident detection algorithm with adapting threshold parameters, Transportation Research-Part C: Emerging Technologies 19 (2011) 1-19.
[15] B. Vidakovic, Statistical Modeling by Wavelets, Wiley, Hoboken, New Jersey, 1999.
[16] D. Montgomery, Introduction to Statistical Quality Control, New York, Wiley, 2005.
Key words: Freeway incident, incident detection algorithms, Hotelling T2 control chart, wavelet transforms, feature selection.
1. Introduction
The automated detection of incidents on the freeway is essentially one of the primary functions of an Advanced Traffic Management Systems (ATMS), which is the important part of Intelligent Transportation Systems (ITS). Accurate and rapid incident detection is particularly important to improve the capacity of freeways because incident causes severe reduction in the traffic flow or a secondary incident, which may result in sacrificing many human life and cost. Numerous freeway incident detection algorithms were developed in the last several decades. These algorithms include California algorithm [1], Minnesota algorithm [2], neural network based algorithm [3-6], nonparametric regression [7], Support Vector Machine (SVM) [8], decision tree learning [9], and wavelet transform-based algorithm [10-13].
However, because of traffic data dependence and high False Alarm Rate (FAR) of existing algorithms, most of traffic information centers are not operating them. Recently, a wavelet transform, which is one of the promising techniques in the field of signal analysis and data mining, has been applied for traffic incident detection [12]. Because of the superior ability of denosing, wavelet transforms have been used to extract informative features from noisy original traffic measurements before using input vectors into incident detection algorithms. Several recent researches [14] based on the wavelet technique showed that the wavelet coefficients (finer and coarse level coefficients) were directly utilized in detecting changes in traffic measurements. For deciding decision-making rule for incident detection, the finest level coefficients were employed to detect significant and abrupt changes in traffic measurement while the coarse level coefficients were taken advantage to detect global incident trends[14]. The upstream traffic measurements of occupancies and traffic volumes will increase and downstream ones will decrease after the occurrence of incidents. The traffic speeds indicate an opposite traffic phenomenon. However, most of previous researches for freeway incident detection consider incident detection problems as classification ones where lots of incident traffic measurements are needed to construct an incident detection model. Their need for much of incident measurements has kept them from a wide deployment at ITS traffic management and control centers. Therefore, this research proposes a novel freeway incident detection algorithm of the wavelet-based Hotelling T2 control chart, which utilizes multi-resolution property in wavelet transforms and statistical quality control model with only normal training dataset. Unlike the existing methods, which consider incident detection problems as classification ones, the proposed approach is applicable even to the cases when only normal traffic measurements are available. The authors compare the performance of the proposed algorithm with that of existing popular methods based on both real-life datasets.
This paper is organized as follows: A brief review of wavelet transform is given in section 2; The proposed methodology is described in section 3, followed by experimental results in section 4; finally, section 5 concludes this paper with a brief summary and discussion in the future works.
References
[1] H.J. Payne, S.C. Tignor, Freeway incident-detection algorithms based on decision trees with states, Transportation Research Board (682) (1978) 30-37.
[2] Y.J. Stephanedes, A.P. Chassiakos, Application of filtering techniques for incident detection, Journal of Transportation Engineering 119 (1993) 13-26.
[3] H. Dia, G. Rose, Development and evaluation of neural network freeway incident detection models using field data, Transportation Research Part C: Emerging Technologies 5 (1997) 313-331.
[4] S. Ishak, H. Al-Deek, Performance of automatic ANN-based incident detection on freeways, Journal of Transportation Engineering 125 (1999) 281-290.
[5] A. Samant, H. Adeli, Enhancing neural network traffic incident-detection algorithms using wavelets, Computer-Aided Civil and Infrastructure Engineering 16(2001) 239-245.
[6] D. Srinivasan, X. Jin, R.L. Cheu, Adaptive neural network models for automatic incident detection on freeways, Neurocomputing 64 (2005) 473-496.
[7] S. Tang, H. Gao, Traffic-incident detection-algorithm based on nonparametric regression, IEEE Transactions on Intelligent Transportation Systems 6 (2005) 38-42.
[8] F. Yuan, R.L. Cheu, Incident detection using support vector machines, Transportation Research Part C: Emerging Technologies 11 (2003) 309-328.
[9] S. Chen, W. Wang, Decision tree learning for freeway automatic incident detection, Expert Systems with Applications: An International Journal 36 (2009) 4101-4105.
[10] A. Samant, H. Adeli, Feature extraction for traffic incident detection using wavelet transform and linear discriminant analysis, Computer-Aided Civil and Infrastructure Engineering 15 (2000) 241-250.
[11] A. Karim, H. Adeli, Incident detection algorithm using wavelet energy representation of traffic patterns, Journal of Transportation Engineering 128 (2002) 232-242.
[12] S. Ghosh-Dastidar, H. Adeli, Wavelet-clustering-neural network model for freeway incident detection, Computer-Aided Civil and Infrastructure Engineering 18(2003) 325-338.
[13] X. Wang, J. Zhang, A traffic incident detection method based on wavelet Mallat algorithm, in: Proceedings of the 2005 IEEE Mid-Summer Workshop on Soft Computing in Industrial Applications, Espoo, Finland, 2005, pp. 166-172.
[14] Y.S. Jeong, M. Castro-Neto, M. Jeong, L. Han, A wavelet-based freeway incident detection algorithm with adapting threshold parameters, Transportation Research-Part C: Emerging Technologies 19 (2011) 1-19.
[15] B. Vidakovic, Statistical Modeling by Wavelets, Wiley, Hoboken, New Jersey, 1999.
[16] D. Montgomery, Introduction to Statistical Quality Control, New York, Wiley, 2005.