Recurrent Autoencoder with Skip Connections and Exogenous Variables for Traffic Forecasting

TL;DR

This paper presents a novel traffic forecasting model that uses recurrent autoencoders with skip connections and exogenous variables, effectively predicting traffic dynamics from video-like aggregated data while ensuring diverse sequence sampling.

Contribution

It introduces a new approach combining recurrent autoencoders, skip connections, and exogenous variables for traffic prediction from video data, with a novel sequence sampling method.

Findings

Effective prediction of traffic speed, volume, and direction.

Improved modeling of traffic dynamics through low-dimensional representations.

Enhanced training with diverse sequence sampling.

Abstract

The increasing complexity of mobility plus the growing population in cities, together with the importance of privacy when sharing data from vehicles or any device, makes traffic forecasting that uses data from infrastructure and citizens an open and challenging task. In this paper, we introduce a novel approach to deal with predictions of speed, volume, and main traffic direction, in a new aggregated way of traffic data presented as videos. The approach leverages the continuity in a sequence of frames and its dynamics, learning to predict changing areas in a low dimensional space and then, recovering static features when reconstructing the original space. Exogenous variables like weather, time and calendar are also added in the model. Furthermore, we introduce a novel sampling approach for sequences that ensures diversity when creating batches, running in parallel to the optimization process.