Effective long-term travel time prediction with fuzzy rules for tollway

Abstract

Abstract:Advanced traveller information system is an important Intelligent Transportation Systems (ITS) application area, which provides information to transport users and managers in order to improve the efficiency and effectiveness of the transportation system, in the face of increasing congestion in urban cities around the world. So far very limited research attention has been focused on long term travel time prediction (i.e. predicting greater than 60 minutes ahead). Long term travel time forecasts can play a critical role in journey planning decisions for both private road users and logistics operators. In this paper, we have considered a fuzzy neural network incorporated with both imprecise and numerical information and developed a hybrid long term travel time prediction model, which shows the better prediction capability than naive methods and highlights the importance of different data variables. The model combines the learning ability of neural networks and the knowledge extraction ability of fuzzy inference systems. The model was validated by using travel time data compiled from electronic toll tags on a 14 km length section of the CityLink tollway in Melbourne, Australia. The validation results highlight the ability of the fuzzy neural network model to accommodate imprecise and linguistic input information, while providing reliable predictions of travel times up to a few days ahead.

AdaBoost algoritmasını kullanarak demiryolu trafik yönetimi için seyir süresinin tahmini

“…In [10], the mean error values of the prediction results are 3% for training data and 7% for test data. In [11], an error range of 5-22% was obtained for predictions produced regarding 90% travel time data.…”

“…In [10], a work was conducted on the prediction of the travel time using the advanced artificial neural networks method. In [11], the prediction of the travel time was carried out in a long-term and efficient manner with the help of fuzzy logic. In the study [12], the prediction of the travel time was carried out using speed predictions.…”

A novel Monte Carlo-based neural network model for electricity load forecasting

“…Although this paper significantly extended the work that we reported in (Yong et al, 2017), especially by providing a discussion on the performance of the GVM, the combination of GVM model with other model is still not discussed, which needs to be undertaken in the future. For example, applications of fuzzy theories to incorporate linguistic values in ANN are widely used in time series prediction, such as the long-term travel time prediction model with a fuzzy neural network (Li et al, 2017). The fluctuations of electricity supplies and demands could be affected by many indeterministic factors such as extreme weather and human errors or even terrorist attacks.…”

Investigation and Analysis of Travel Time Variability of Selected Road Intersection

“…A high degree of variability indicates that the travel time would be unpredictable making the traffic service less reliable (Turochy and Smith, 2002). Accurate and reliable prediction of travel time helps drivers to make an appropriate decision on the selection of optimal routes and traffic management authorities to efficiently control and manage traffic congestion on a route (Anwar, 2010;Shi et al, 2017;Li et al, 2018).…”