Accident-Focused Travel Time Prediction on Expressway Using Multivariate LSTM

Previous work was done to compare the static resource allocation (StRA) scheme and scalable resource allocation (ScRA) scheme within an integrated GSM and GPRS network (see Tang, S. and Tafazolli, R., Vehicular Technology Conference - Spring, 2003). In order to maintain the QoS for both, the ScRA scheme assumed that the load history database was a credible estimate of the present state of traffic intensity in the cell area. In cases when there is a sudden change in traffic conditions, the ScRA scheme may not be able to perform well. Hence, in order to cope with sudden changes of traffic in the network, we propose to incorporate prediction schemes into the ScRA algorithm. For comparison purposes, we present and evaluate three prediction strategies, namely linear regression prediction, exponential regression prediction, and current value prediction, in six various traffic conditions. In addition, resource utilisation gains between the StRA scheme and the proposed ScRA schemes are also studied. The findings show that ScRA exponential regression prediction scheme can cope best in all the tested traffic scenarios and experiences an average of 0.5 percent of resource utilisation gain over the StRA scheme.

In the realm of intelligent transportation systems, accurately predicting vehicle trajectories is paramount for enhancing road safety and optimizing traffic flow management. Addressing the impacts of complex traffic environments and efficiently modeling the diverse behaviors of vehicles are the key challenges at present. To achieve precise prediction of vehicle trajectories, it is essential to fully consider the dynamic changes in traffic conditions and the long-term dependencies of time-series data. In response to these challenges, we propose the Memory-Enhanced Spatio-Temporal Graph Network (MESTGN), an innovative model that integrates a Spatio-Temporal Graph Convolutional Network (STGCN) with an attention-enhanced Long Short-Term Memory (LSTM)-based sequence to sequence (Seq2Seq) encoder–decoder structure. MESTGN utilizes STGCN to capture the complex spatial dependencies between vehicles and reflects the interactions within the traffic network through road traffic data and network topology, which significantly influences trajectory prediction. Additionally, the model focuses on historical vehicle trajectory data points using an attention-weighted mechanism under a traditional LSTM prediction architecture, calculating the importance of critical trajectory points. Finally, our experiments conducted on the urban traffic dataset ApolloSpace validate the effectiveness of our proposed model. We demonstrate that MESTGN shows a significant performance improvement in vehicle trajectory prediction compared with existing mainstream models, thereby confirming its increased prediction accuracy.

Предметом вивчення в статті є новий підхід визначення параметрів руху транспортних засобів і осості визначення пропускної здатності (ПС) нерегульованих перехресть різного типу. Метою є розробка способу визначення ПС міських нерегульованих перетинів, який дозволяє отримати і Дольний комплекс критеріальних оцінок якості функціонування таких перехресть. Завдання дослідження: аналіз су-суспільством і розробка нових підходів, методів і алгоритмів визначення ПС нерегульованих перехресть, які враховують у новій пропозиції позитивних якостей відомих рішень; розширення функціональних віз-можностей відомих рішень для отримання широкого комплексу критеріїв оцінки якості функціонування перехресть; пропозиція і розробка такого способу, який відповідав би вимогам універсальної адаптивної системи, що дозволяє ефективно в динаміці реагувати на всілякі зміни умов руху транс-кравців потоків (ТП) на перехресті. Отримані наступні результати. Розкрито деякі особливості сучас-менной організації моніторингу ТП на локальних об'єктах, а також проаналізовано їх основні переваги та недоліки. Представлений новий підхід визначення ПС нерегульованих перехресть в умовах динамічного изме-нения параметрів руху ТП і адаптивного реагування системи на дорожню ситуацію. Показано особливості реалізації алгоритмів визначення ПС для різного типу перетинів. Висновки. Запропонована технологія изме ренію комплексу параметрів руху ТП одночасно позволяє здійснювати оцінку якості функціонування нерегульованих перетинів за критерієм ПС і в динаміці контролювати цей параметр. Визначення ПС таким способом з високой частотою сканування і за реальними значеннями часу роз'їзду на перехресті різних типів транспортних засобів, дає можливість істотно підвищити точність вимірювання кінцевих параметрів. Пропозиція і розробка способу визначення ПС нерегульованих перехресть за такою технологією відповідає вимогам універсальних адаптивних систем, які еффективно в динаміці реагують на всілякі зміни умов руху ТП на перехресті і УДС міста

Short-term traffic volume prediction by ensemble learning in concept drifting environments

The metro station can form a new commercial district.The residential development in the outskirts has a major impact on the value of real estate.However, the area near the metro station can be wrongfully affected due to the structure and function of the railway station, noise or congestion, etc.This study aims to analyze how changes in traffic conditions affect the apartments.The opening of the urban road is considered as traffic change.Regarding it, the price trend will be investigated.There is an assumption that preference will grow when opened.Daegu Subway Line 3 in Korea, which opened in 2015, is selected.It is the relatively recent case.Sales price of apartment houses included in the station area and apartment houses outside the station area is obtained, from which we calculate the rate of change.The difference between station area and non-station area was analyzed under two hypotheses whether it existed or not.The 2015 case and after-one-year case were compared.The subway station proved to raise housing price.T-test was utilized to confirm the reliability of the results.They were significant at the 95% confidence level of the two-sided test.Consequently, the station apparently tends to increase public housing price.

This study investigates the factors affecting estimation accuracy of queue length at signalized intersections under low penetration of connected vehicles. A shockwave-based algorithm is proposed to estimate the maximum queue length and residual queue on a cycle-by-cycle basis. Simulation data collected from three consecutive signalized intersections were used to extract trajectories of connected vehicles under five different market penetration rates and two different traffic conditions (under-saturated and moderate). The results confirm that the queue length estimation process is probabilistic and affected by the stochastic changes in traffic conditions. This probabilistic nature is defined by a queue formation coverage index (QI) that proved to significantly affect the queue length estimation accuracy. Overall, the results show that the accuracy of the queue estimates is acceptable when a QI value of at least 50% is achieved. In such limited data environments, the QI showed potential to help as an assessment tool to evaluate the obtained queue estimates.

Increasing traffic volumes leads to changes in traffic conditions, especially at intersections. These changes affect the ability of Vulnerable Road Users (VRUs) such as pedestrians and bicyclists to cross the road safely, especially at un-signalized crossings (without traffic lights), where many road users are conflicting at the same point of the road and neglecting 'in some cases' the priority issue. Although the emphasis on pedestrian safety has recently increased, there is still a need to analyze the causal factors of VRUs accidents and define their relationship to road design characteristics. This paper presents a study bout vehicle – pedestrian and bicyclist conflicts analysis using Swedish conflict technique at three un-signalized pedestrian crossings in the city of Győr- Hungary, where some pedestrian accidents were happened between 2014 and 2018, and reported based on accidents database in Hungary. The aim is to analyse vehicle-pedestrian and bicyclist interactions, and evaluate the severity of conflicts with the help of the Swedish technique graph. The findings concluded that there is a compatibility between Swedish conflict technique and accident records with regard to conflicts severity and the probability of accidents occurrence.

An asynchronous transfer mode ATM network is a high-speed multimedia network that handles various kinds of traffic with different bit rates and different qualities of services QoS. To maintain QoS for each traffic source and to avoid a possible congestion problem, an ATM network requires highly sophisticated and flexible controllers to insure that the demanding performance can be achieved under unexpected changes in traffic conditions. In this article we propose an intelligent architecture using neural networks for traffic congestion control in an ATM network. The congestion control using neural networks is suitable for an ATM because neural networks can learn the offered traffic characteristics and the dynamic changes of the traffic. The proposed mechanism is based on the adaptive prediction of the future value of the offered traffic and the flow rate for each traffic source. At every given time slot, the controllers in the proposed architecture predict whether the congestion will happen or not and regulate the volume of input traffic for each traffic source before the congestion happens, maintaining the user-required QoS for each traffic source based on the predefined rules. Consequently, the mechanism guarantees the QoS for each traffic source and efficiently prevents congestion.

Traffic problems have seriously affected people's life quality and urban development, and forecasting the short-term traffic congestion is of great importance to both individuals and governments. However, understanding and modeling the traffic conditions can be extremely difficult, and our observations from real traffic data reveal that (1) similar traffic congestion patterns exist in the neighboring time slots and on consecutive workdays; (2) the levels of traffic congestion have clear multiscale properties. To capture these characteristics, we propose a novel method named PCNN based on deep Convolutional Neural Network, modeling Periodic traffic data for short-term traffic congestion prediction. PCNN has two pivotal procedures: time series folding and multi-grained learning. It first temporally folds the time series and constructs a two-dimensional matrix as the network input, such that both the real-time traffic conditions and past traffic patterns are well considered; then with a series of convolutions over the input matrix, it is able to model the local temporal dependency and multiscale traffic patterns. In particular, the global trend of congestion can be addressed at the macroscale; whereas more details and variations of the congestion can be captured at the microscale. Experimental results on a real-world urban traffic dataset confirm that folding time series data into a two-dimensional matrix is effective and PCNN outperforms the baselines significantly for the task of short-term congestion prediction.

Global traffic management encounters a significant challenge in traffic congestion. This paper presents a hybrid machine learning method for predicting traffic congestion. It leverages Convolutional Neural Network-Gated Recurrent Unit (CNN-GRU) network for parameter prediction and combines clustering and classification models, using the K-means algorithm to categorize traffic states into four levels and constructing a KNN classification model based on this segmentation. This results in the K-means-KNN model. Predicted parameters are inputted into the K-means-KNN model for congestion level prediction. Validation with real traffic flow data shows that the CNN-GRU network can capture spatiotemporal features more effectively. For instance, in traffic flow prediction, it reduces the MAPE by 7.39% and 51.14% compared to CNN and GRU, respectively. K-means-KNN excels in traffic state discrimination, achieving a congestion prediction accuracy of 91.8%. These results underscore the efficacy of the hybrid machine learning method in assessing and predicting urban traffic congestion.

The Izu Islands comprise nine inhabited islands ranging from Izuôshima to Aogashima. Geologically volcanic with the sea-eroded bluffs generally unfavorable for port facilities, each of these islands has long remained a typical specimen of isolated existence. With the recent consolidation of airport and port facilities, however, the features of these islands have been rapidly undergoing substantial changes, supported by their advantage of location not far distant from the metropolitan province.Regional transformation incidental to the change in traffic conditions is conceived to appear most conspicuously in hitherto forlorn isolated islands. On such islands, even without quantitative analysis, it is certainly of deep interest to trace how, as a way of changing the self-supporting nature of the traditional economy, specific production and tourist industries have been introduced and developed in harmony with their old industries. The present study was designed to scrutinize the changes, mainly in its industrial structure, of an isolated island with the peculiarity of a regionally segregated existence. For this purpose, out of the nine islands, Kôzushima Isle, which is regarded as typical of the modernizing Izu islands with its improved port facilities and markedly developed tourist industry was examined, and an analysis was attempted of this isle's industrial structure and its changes, with on-the-spot surveys conducted to integrate a series of investigations.1) Formerly, the industrial structure of Keszushima Isle was composed mainly of conventional marine products industry, boat-fishing by men and seaweed gathering by women, complemented by small-scale livestock raising and sweet potato cultivation, and also by the government's unemployment relief public works. This minor-scale industrial makeup, essentially on a self-supporting basis, has undergone a drastic change, with the features specifically showing a decreasing weight of dependence on marine products, fish and sea-weed, and compensated for by an increasing weight of sugar pea cultivation, specialized on a commerical basis, and a rapidly extended sight-seeing business.In these radical changes in the industrial structure, which forced female employment to be widely reshuffled, the transition is considered to have been made with an appropriate labor allocation between men and women accompanied by a definitve pattern of specialization. It is certainly worthy of note that in this isolated island, with a narrow and weak economic base, all units of industry exist with close interrelations, including those in seasonally fluctuating work fields.2) It has been generally believed that the segregated existence of an isolated island is attributable to its inferiority in transportation links. This preoccupation is now facing a doubt raised by the fact that the industrial structure alteration appeared in Kôzushima Isle prior to the actual improvement of transportation conditions. Improvement in transportation is important in the industrial structure and people's life on an isolated island, but is it always an indispensable factor? It seems more adequate to think that the requirements from outside the island, which have become strong enough to overcome the poor transportation conditions, have brought great pressure to bear on the isolated island in changing its narrow and feeble economy.

In the contemporary landscape of global logistics, the capacity to dynamically optimize delivery routes in real time is a critical determinant of operational efficiency, customer satisfaction, and environmental sustainability. Conventional routing algorithms, while effective in static or semi-dynamic contexts, often fail to adapt to rapid changes in traffic conditions, weather disruptions, road closures, and last-minute delivery requests. This paper proposes a novel Deep Learning-based architecture that leverages Long Short-Term Memory (LSTM) networks for real-time traffic forecasting and Deep Q-Networks (DQNs) for adaptive route decision-making. The system processes live inputs from GPS sensors, weather APIs, and traffic feeds to dynamically generate optimal delivery paths. Extensive simulations using synthetic and real-world datasets from urban logistics providers demonstrate substantial improvements in delivery time (up to 21%), fuel consumption (13%), and vehicle utilization rates (17%) compared to traditional heuristics-based methods. This research provides both theoretical contributions and practical implementation guidelines for logistics operators seeking intelligent, AI-driven solutions for route optimization.

Expected changes in the energy performance of traffic flows as penetrated by electrified vehicles

SummaryLow‐diameter networks require non‐minimal adaptive routing to deal with varying traffic characteristics and avoid pathological performance. Such routing is based on local estimations of network congestion, based on link‐level flow control credits. Dragonfly networks based on the extensions of commodity Ethernet networks using OpenFlow have been proposed for large HPC deployments with low power consumption. However, this network technology does not implement credit‐based flow control. This work explores a range of routing solutions based on exploiting explicit congestion notification messages (in particular, 802.1Qau) to adapt the number of packets using non‐minimal paths. The design (denoted QCN‐Switch) associates a probability value to each output port. This value is updated to reflect downstream congestion and used to statistically divert traffic away from congested areas when the load is uneven, as in the case of adversarial traffic. A feedback comparison variant is designed to separate the cases of uniform traffic at saturation and adversarial traffic at low loads. Evaluation results show that QCN‐Switch is a competitive design for both the uniform traffic and adversarial traffic. Furthermore, it is able to react to changes in traffic conditions in 0.4 ms or less. A sensitivity analysis identifies the best configuration and shows its performance trade‐offs.

An asynchronous transfer mode (ATM) network is a high-speed multimedia network which handles various kinds of traffic with different bit-rates and different qualities of service (QoS). In order to maintain the QoS for each traffic source and to avoid possible congestion problems, an ATM network requires highly sophisticated and flexible controllers to insure that this demanding performance can be achieved under unexpected changes in traffic conditions. In this paper, we propose an intelligent architecture using recurrent neural networks and an expert system for traffic control in ATM networks. This traffic control using neural networks is suitable for ATM because neural networks can learn the traffic characteristics and the dynamic changes in the traffic. The proposed mechanism is based on the adaptive prediction of the future values of the traffic and the flow rate for each traffic source. At every given time slot, the controllers in the proposed architecture predict whether the congestion will happen or not and regulate the volume of input traffic for each traffic source before the congestion happens, maintaining the user-required QoS for each traffic source based on the predefined rules. Consequently, the mechanism guarantees the QoS for each traffic source and efficiently prevents congestion.