Traffic Prediction using a Deep Learning Paradigm

For many years intelligent transportation systems (ITS) have been collecting and processing huge amounts of data from numerous sensors to generate a ground truth of urban traffic. Such data has set the foundation of traffic theory, planning and simulation to create rule-based systems. It has also been used in many different studies in data-driven short-term traffic flow forecasting with promising results. Still, the acceptance for data-driven predictions is quiet low in productive systems of the public sector. Without enough probe data from floating cars (FCD) ITS owners feel unable to reach accuracy like private telecommunication or car manufacturing companies. On the other hand, investigating into FCD requires a thoughtful treatment of user privacy and a close look on data quality which can also be very time consuming. Recent progress in hardware and deep learning software has lowered the bar to handle machine learning algorithms what urges the field of traffic forecasting to continue exploring the predictive power of artificial intelligence. With this paper we present our first approach of feeding sensor data to an Artificial Neural Network (ANN). We train the ANN with different spatial and temporal lags to find an optimal setup for an entire city.

[1]  Jiaqiu Wang,et al.  A Dynamic Spatial Weight Matrix and Localized Space–Time Autoregressive Integrated Moving Average for Network Modeling , 2014 .

[2]  Yu Zheng,et al.  Deep Spatio-Temporal Residual Networks for Citywide Crowd Flows Prediction , 2016, AAAI.

[3]  Shiliang Sun,et al.  A bayesian network approach to traffic flow forecasting , 2006, IEEE Transactions on Intelligent Transportation Systems.

[4]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[5]  Nicholas G. Polson,et al.  Deep Learning Predictors for Traffic Flows , 2016 .

[6]  Billy M. Williams,et al.  Modeling and Forecasting Vehicular Traffic Flow as a Seasonal ARIMA Process: Theoretical Basis and Empirical Results , 2003, Journal of Transportation Engineering.

[7]  Axel Leonhardt,et al.  Instance-Based Learning for Estimating and Predicting Traffic State Variables Using Spatiotemporal Traffic Patterns , 2012 .

[8]  H. V. van Zuylen,et al.  Predicting Urban Arterial Travel Time with State-Space Neural Networks and Kalman Filters , 2006 .

[9]  Paolo Frasconi,et al.  Short-Term Traffic Flow Forecasting: An Experimental Comparison of Time-Series Analysis and Supervised Learning , 2013, IEEE Transactions on Intelligent Transportation Systems.

[10]  Geoffrey E. Hinton,et al.  Learning internal representations by error propagation , 1986 .

[11]  Yiannis Kamarianakis,et al.  Space-time modeling of traffic flow , 2002, Comput. Geosci..

[12]  References , 1971 .

[13]  Zachary Chase Lipton.  A Critical Review of Recurrent Neural Networks for Sequence Learning , 2015, ArXiv.

[14]  Eleni I. Vlahogianni,et al.  Short-term traffic forecasting: Where we are and where we’re going , 2014 .

[15]  Yunpeng Wang,et al.  Long short-term memory neural network for traffic speed prediction using remote microwave sensor data , 2015 .

[16]  Xiaosi Zeng,et al.  Development of Recurrent Neural Network Considering Temporal‐Spatial Input Dynamics for Freeway Travel Time Modeling , 2013, Comput. Aided Civ. Infrastructure Eng..