Short-term prediction of traffic dynamics with real-time recurrent learning algorithms

Abstract

Short-term prediction of dynamic traffic states remains critical in the field of advanced traffic management systems and related areas. In this article, a novel real-time recurrent learning (RTRL) algorithm is proposed to address the above issue. We dabble in comparing pair predictability of linear method versus RTRL algorithms and simple non-linear method versus RTRL algorithms individually using a first-order autoregressive time-series AR(1) and a deterministic function. A field study tested with flow, speed and occupancy series data collected directly from dual-loop detectors on a freeway is conducted. The numerical results reveal that the performance of RTRL algorithms in predicting short-term traffic dynamics is satisfactorily accepted. Furthermore, it is found that the dynamics of short-term traffic states characterised in different time intervals, collected in diverse time lags and times of day may have significant effects on the prediction accuracy of the proposed algorithms.

Trend analysis of traffic management based on literature data mining and graph analysis tools

“…Meanwhile, short‐term traffic flow prediction was proposed and studied [72]. After different algorithms used for short‐term traffic flow prediction, the prediction results need to be displayed and evaluated using a simulation model [73]. There are both macro traffic analysis and micro traffic analysis on traffic management.…”

Bayesian predictive travel time methodology for advanced traveller information system

“…a pattern recognition or another statistical model, a neural network model) in association with real time data can form the basis for estimating expected travel time. Recent literature sources suggest that significant methodological advances have taken place that can be of assistance in predicting traffic and network performance [20][21][22][23].…”

Developing Short-Term Traffic Forecasting Models Using Seasonal ARIMA Method for D-200 Highway

“…Sonuç olarak, bütün modellerin yaklaşık sonuçlar ürettiği, KF' nin en az işlem yükü ile sonuçlara ulaştığı, geliştirilen SARIMA modelinin sahada AUS uygulamaları için kolaylıklar sağlayacağı fikri ortaya atılmıştır. Sheu ve diğerleri, akım, hız ve yoğunluk tahminleri yapmak amacıyla gerçek zamanlı tekrarlı öğrenme algoritması geliştirmiştir [6]. Ardından araştırmacılar, bu trafik tahmin algoritmasını, makalede açıklanan lineer ve lineer olmayan yöntemler ile karşılaştırmıştır.…”