Multi-output bus travel time prediction with convolutional LSTM neural network

Abstract Accurate and reliable travel time predictions in public transport networks are essential for delivering an attractive service that is able to compete with other modes of transport in urban areas. The traditional application of this information, where arrival and departure predictions are displayed on digital boards, is highly visible in the city landscape of most modern metropolises. More recently, the same information has become critical as input for smart-phone trip planners in order to alert passengers about unreachable connections, alternative route choices and prolonged travel times. More sophisticated Intelligent Transport Systems (ITS) include the predictions of connection assurance, i.e. an expert system that will decide to hold services to enable passenger exchange, in case one of the services is delayed up to a certain level. In order to operate such systems, and to ensure the confidence of passengers in the systems, the information provided must be accurate and reliable. Traditional methods have trouble with this as congestion, and thus travel time variability, increases in cities, consequently making travel time predictions in urban areas a non-trivial task. This paper presents a system for bus travel time prediction that leverages the non-static spatio-temporal correlations present in urban bus networks, allowing the discovery of complex patterns not captured by traditional methods. The underlying model is a multi-output, multi-time-step, deep neural network that uses a combination of convolutional and long short-term memory (LSTM) layers. The method is empirically evaluated and compared to other popular approaches for link travel time prediction and currently available services, including the currently deployed model at Movia, the regional public transport authority in Greater Copenhagen. We find that the proposed model significantly outperforms all the other methods we compare with, and is able to detect small irregular peaks in bus travel times very quickly.

[1]  Lelitha Vanajakshi,et al.  Comparison of Model Based and Machine Learning Approaches for , 2014 .

[2]  Lin Lu,et al.  Predicting short-term stock prices using ensemble methods and online data sources , 2018, Expert Syst. Appl..

[3]  D. Dailey,et al.  Preprint Title: an Algorithm for Predicting the Arrrival Time of Mass Transit Vehicles Using Automatic Vehicle Location Data an Algorithm for Predicting the Arrival Time of Mass Transit Vehicles Using Automatic Vehicle Location Data , 2022 .

[4]  Amer Shalaby,et al.  PREDICTION MODEL OF BUS ARRIVAL AND DEPARTURE TIMES USING AVL AND APC DATA , 2004 .

[5]  Mehmet Altinkaya,et al.  Urban Bus Arrival Time Prediction : A Review of Computational Models , 2013 .

[6]  N. P. Olewuezi,et al.  Note on the Comparison of Some Outlier Labeling Techniques , 2011 .

[7]  Shih-Che Lo,et al.  Design of real-time fuzzy bus holding system for the mass rapid transit transfer system , 2012, Expert Syst. Appl..

[8]  Lei Jia,et al.  Real-Time Bus Arrival Time Prediction: Case Study for Jinan, China , 2013 .

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

[10]  Laurence R. Rilett,et al.  Prediction Model of Bus Arrival Time for Real-Time Applications , 2005 .

[11]  Dit-Yan Yeung,et al.  Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting , 2015, NIPS.

[12]  Matthias Weidlich,et al.  Traveling time prediction in scheduled transportation with journey segments , 2017, Inf. Syst..

[13]  A. Transitway.  PERCEPTION OF WAITING TIME AT TRANSIT STOPS AND STATIONS , 2015 .

[14]  Jürgen Schmidhuber,et al.  Learning to Forget: Continual Prediction with LSTM , 2000, Neural Computation.

[15]  Jian Sun,et al.  Dynamic Bus Travel Time Prediction Models on Road with Multiple Bus Routes , 2015, Comput. Intell. Neurosci..

[16]  Haris N. Koutsopoulos,et al.  Effect of Real-Time Transit Information on Dynamic Path Choice of Passengers , 2011 .

[17]  Kun Tang,et al.  Personalized travel time estimation for urban road networks: A tensor-based context-aware approach , 2018, Expert Syst. Appl..

[18]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[19]  Martin J. Wainwright,et al.  Noisy matrix decomposition via convex relaxation: Optimal rates in high dimensions , 2011, ICML.

[20]  Xiaobo Liu,et al.  A Dynamic Bus‐Arrival Time Prediction Model Based on APC Data , 2004 .

[21]  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.

[22]  Cyrus Shahabi,et al.  Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting , 2017, ICLR.

[23]  Jürgen Schmidhuber,et al.  Applying LSTM to Time Series Predictable through Time-Window Approaches , 2000, ICANN.

[24]  Bo Yu,et al.  Hybrid Model for Prediction of Bus Arrival Times at Next Station , 2010 .

[25]  Steven I-Jy Chien,et al.  ESTIMATION OF BUS ARRIVAL TIMES USING APC DATA , 2004 .

[26]  Jürgen Schmidhuber,et al.  Recurrent nets that time and count , 2000, Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium.

[27]  Liping Fu,et al.  Predicting Bus Arrival Time on the Basis of Global Positioning System Data , 2007 .

[28]  Bin Yu,et al.  Bus arrival time prediction at bus stop with multiple routes , 2011 .

[29]  Fei-Yue Wang,et al.  Travel time prediction with LSTM neural network , 2016, 2016 IEEE 19th International Conference on Intelligent Transportation Systems (ITSC).

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

[31]  Carol L Schweiger.  Real-Time Bus Arrival Information Systems , 2003 .

[32]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[33]  Steven I-Jy Chien,et al.  Dynamic Freeway Travel-Time Prediction with Probe Vehicle Data: Link Based Versus Path Based , 2001 .