Few-shot machine learning attempts to predict outputs given only a very small number of training examples. The key idea behind most few-shot learning approaches is to pre-train the model with a large number of instances from a different but related class of data, classes for which a large number of instances are available for training. Few-shot learning has been most successfully demonstrated for classification problems using Siamese deep learning neural networks. Few-shot learning is less extensively applied to time-series forecasting. Few-shot forecasting is the task of predicting future values of a time-series even when only a small set of historic time-series is available. Few-shot forecasting has applications in domains where a long history of data is not available. This work describes deep neural network architectures for few-shot forecasting. All the architectures use a Siamese twin network approach to learn a difference function between pairs of time-series, rather than directly forecasting based on historical data as seen in traditional forecasting models. The networks are built using Long short-term memory units (LSTM). During forecasting, a model is able to forecast time-series types that were never seen in the training data by using the few available instances of the new time-series type as reference inputs. The proposed architectures are evaluated on Vehicular traffic data collected in California from the Caltrans Performance Measurement System (PeMS). The models were trained with traffic flow data collected at specific locations and then are evaluated by predicting traffic at different locations at different time horizons (0 to 12 hours). The Mean Absolute Error (MAE) was used as the evaluation metric and also as the loss function for training. The proposed architectures show lower prediction error than a baseline nearest neighbor forecast model. The prediction error increases at longer time horizons.
more »
« less
Traffic Volume Prediction with Automated Signal Performance Measures (ATSPM) Data
Predicting short-term traffic volume is essential to improve transportation systems management and operations (TSMO) and the overall efficiency of traffic networks. The
real-time data, collected from Internet of Things (IoT) devices, can be used to predict traffic volume. More specifically, the Automated Traffic Signal Performance Measures (ATSPM) data
contain high-fidelity traffic data at multiple intersections and can reveal the spatio-temporal patterns of traffic volume for each signal. In this study, we have developed a machine learningbased approach using the data collected from ATSPM sensors of a corridor in Orlando, FL to predict future hourly traffic. The hourly predictions are calculated based on the previous six hours volume seen at the selected intersections. Additional factors that play an important role in traffic fluctuations include peak hours, day of the week, holidays, among others. Multiple machine learning models are applied to the dataset to determine the model with the best performance. Random Forest, XGBoost, and LSTM models show the best performance in predicting hourly traffic volumes.
more »
« less
- Award ID(s):
- 1852002
- PAR ID:
- 10410229
- Date Published:
- Journal Name:
- IEEE Symposium on Computers and Communications (ISCC)
- Page Range / eLocation ID:
- 1 to 6
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Traffic forecasting plays an important role in urban planning. Deep learning methods outperform traditional traffic flow forecasting models because of their ability to capture spatiotemporal characteristics of traffic conditions. However, these methods require high-quality historical traffic data, which can be both difficult to acquire and non-comprehensive, making it hard to predict traffic flows at the city scale. To resolve this problem, we implemented a deep learning method, SceneGCN, to forecast traffic speed at the city scale. The model involves two steps: firstly, scene features are extracted from Google Street View (GSV) images for each road segment using pretrained Resnet18 models. Then, the extracted features are entered into a graph convolutional neural network to predict traffic speed at different hours of the day. Our results show that the accuracy of the model can reach up to 86.5% and the Resnet18 model pretrained by Places365 is the best choice to extract scene features for traffic forecasting tasks. Finally, we conclude that the proposed model can predict traffic speed efficiently at the city scale and GSV images have the potential to capture information about human activities.more » « less
-
Microscopic simulation-based approaches are extensively used for determining good signal timing plans on traffic intersections. Measures of Effectiveness (MOEs) such as wait time, throughput, fuel consumption, emission, and delays can be derived for variable signal timing parameters, traffic flow patterns, etc. However, these techniques are computationally intensive, especially when the number of signal timing scenarios to be simulated are large. In this paper, we propose InterTwin, a Deep Neural Network architecture based on Spatial Graph Convolution and Encoder-Decoder Recurrent networks that can predict the MOEs efficiently and accurately for a wide variety of signal timing and traffic patterns. Our methods can generate probability distributions of MOEs and are not limited to mean and standard deviation. Additionally, GPU implementations using InterTwin can derive MOEs, at least four to five orders of magnitude faster than microscopic simulations on a conventional 32 core CPU machine.more » « less
-
null (Ed.)In this paper we describe how high performance computing in the Google Cloud Platform can be utilized in an urgent and emergency situation to process large amounts of traffic data efficiently and on demand. Our approach provides a solution to an urgent need for disaster management using massive data processing and high performance computing. The traffic data used in this demonstration is collected from the public camera systems on Interstate highways in the Southeast United States. Our solution launches a parallel processing system that is the size of a Top 5 supercomputer using the Google Cloud Platform. Results show that the parallel processing system can be launched in a few hours, that it is effective at fast processing of high volume data, and can be de-provisioned in a few hours. We processed 211TB of video utilizing 6,227,593 core hours over the span of about eight hours with an average cost of around $0.008 per vCPU hour, which is less than the cost of many on-premise HPC systems.more » « less
-
The problem of traffic prediction is paramount in a plethora of applications, ranging from individual trip planning to urban planning. Existing work mainly focuses on traffic prediction on road networks. Yet, public transportation contributes a significant portion to overall human mobility and passenger volume. For example, the Washington, DC metro has on average 600,000 passengers on a weekday. In this work, we address the problem of modeling, classifying and predicting such passenger volume in public transportation systems. We study the case of the Washington, DC metro exploring fare card data, and specifically passenger in- and outflow at stations. To reduce dimensionality of the data, we apply principal component analysis to extract latent features for different stations and for different calendar days. Our unsupervised clustering results demonstrate that these latent features are highly discriminative. They allow us to derive different station types (residential, commercial, and mixed) and to effectively classify and identify the passenger flow of “unknown” stations. Finally, we also show that this classification can be applied to predict the passenger volume at stations. By learning latent features of stations for some time, we are able to predict the flow for the following hours. Extensive experimentation using a baseline neural network and two naïve periodicity approaches shows the considerable accuracy improvement when using the latent feature based approach.more » « less