Neural Network-Based Prediction of Traffic Accidents and Congestion Levels Using Real-World Urban Road Data

The traffic congestion detection based on the internet of vehicles is gaining enormous research interest. A vehicle‐to‐vehicle (V2V)‐based method for the detection of road traffic congestion is proposed. Firstly, a fuzzy controller was constructed based on the vehicle speed, traffic density, and traffic congestion rating system, and the level of local traffic congestion was evaluated. Then, the level of local traffic congestion of neighbouring vehicles was queried based on V2V communication, and the level of regional traffic congestion was obtained based on a large sub‐sample hypothesis test. Finally, a simulation test platform was built based on vehicles in network simulation, and the back‐off time slots and received packets of vehicle nodes were calculated. The accuracy of the proposed method for detecting road traffic congestion was compared to the cooperative traffic congestion detection (CoTEC) method and the geomagnetic coil method. The results show that the detection accuracy of the proposed method increased by 5.5 and 7.5%, respectively, compared to the geomagnetic coil method and CoTEC method. The V2V communication network overhead of the proposed traffic congestion detection method is reduced by 90.8% compared to the adopted CoTEC method. The communication overhead of the vehicle node using the proposed method is significantly decreased when there is no traffic congestion.

In urban areas, traffic system is one of the significant indicators to show the growth and progress of a city and it also influences the quality of life of people living in metropolitan cities. In recent years, there is a significant increase in usage of road vehicles which is becoming challenge for existing transportation system. The currently deployed traffic system is not based on the traffic congestion level and a predefined time is allocated for traffic lights at every road crossing which results in traffic congestion and situation becomes worst in the peak traffic hours. This high level traffic congestion contributes in the pollution by the emission of CO 2 and several other pollutants in air. Moreover, it also causes tripling of the fuel consumption and consequently put adverse effects on the economy as well. To address the above problem, this paper presents the development of congestion level based dynamic traffic management system using IoT. It regulates the traffic lights duration based on the real-time congestion level of the traffic measured at the road crossings by using ultrasonic sensors. The development of this project is divided in three phases i-e simulation and logic development, development of IoT based system and finally hardware implementation. In first phase the simulations are done in Proteus and results are presented in four cases i-e normal routine, low level congestion, medium level congestion and high level congestion. In second phase the IoT based system is developed by making the communication link between the end nodes and the gateway over the internet. Finally, the real-time prototype is implemented.

There is an ongoing debate among transport planners and safety policy makers as to whether there is any association between the level of traffic congestion and road safety. One can expect that the increased level of traffic congestion aids road safety and this is because average traffic speed is relatively low in a congested condition relative to an uncongested condition, which may result in less severe crashes. The relationship between congestion and safety may not be so straightforward, however, as there are a number of other factors such as traffic flow, driver characteristics, road geometry, and vehicle design affecting crash severity. Previous studies have employed count data models (either Poisson or negative binomials and their extensions) while developing a relationship between the frequency of traffic crashes and traffic flow or density (as a proxy for traffic congestion). The use of aggregated crash counts at a road segment level or at an area level with the proxy for congestion may obscure the actual relationship. The objective of this study is to explore the relationship between the severity of road crashes and the level of traffic congestion using disaggregated crash records and a measure of traffic congestion while controlling for other contributory factors. Ordered response models such as ordered logit models, heterogeneous choice models, and generalized ordered logit (partially constrained) models suitable for both ordinal dependent variables and disaggregate crash data are used. Data on crashes, traffic characteristics (e.g., congestion, flow, and speed), and road geometry (e.g., curvature and gradient) were collected from the M25 London orbital motorway between 2003 and 2006. Our results suggest that the level of traffic congestion does not affect the severity of road crashes on the M25 motorway. The impact of traffic flow on the severity of crashes, however, shows an interesting result. All other factors included in the models also provide results consistent with existing studies.

Previous research has shown that current driving conditions and driving style have a strong influence over a vehicle's fuel consumption and emissions. This paper presents a methodology for inferring road type and traffic congestion (RT&TC) levels from available onboard vehicle data and then using this information for improved vehicle power management. A machine-learning algorithm has been developed to learn the critical knowledge about fuel efficiency on 11 facility-specific drive cycles representing different road types and traffic congestion levels, as well as a neural learning algorithm for the training of a neural network to predict the RT&TC level. An online University of Michigan-Dearborn intelligent power controller (UMD_IPC) applies this knowledge to real-time vehicle power control to achieve improved fuel efficiency. UMD_IPC has been fully implemented in a conventional (nonhybrid) vehicle model in the powertrain systems analysis toolkit (PSAT) environment. Simulations conducted on the standard drive cycles provided by the PSAT show that the performance of the UMD_IPC algorithm is very close to the offline controller that is generated using a dynamic programming optimization approach. Furthermore, UMD_IPC gives improved fuel consumption in a conventional vehicle, alternating neither the vehicle structure nor its components.

This study evaluates the efficacy of the Federal Road Safety Corps (FRSC) in mitigating traffic congestion along the Ilesa-Benin Highway in Akure, Ondo State, Nigeria. It scrutinizes the efficiency of FRSC's education, patrol, and enforcement strategies to discern their impact on traffic congestion levels. The data were collected from primary sources via questionnaires and activity-based trip surveys, supplemented by secondary sources including satellite imagery and literature. The results revealed high internal consistency in the respondents' perceptions and no evidence of multicollinearity in the dataset. It also indicates the significant contributions of education and enforcement in reducing traffic congestion, contrary to the limited effectiveness of patrols. Among the recommendations include prioritizing improvements in education and enforcement efforts through enhanced training programs and community collaborations as well as re-evaluating and potentially restructuring patrol activities. These insights offer valuable guidance for policymakers and stakeholders in devising strategies to alleviate traffic congestion and enhance road safety on critical highway routes like the Ilesa-Benin Highway.

In densely populated cities, severe vehicle congestion is an obstacle due to the massive increase of vehicles and bottleneck of roads. Traffic congestion brings severe economic and environmental hazards. Although it is hard to eliminate, it can be mitigated by finding the occurrence of the congestion well in advance to initiate appropriate control measures. Accurately estimating traffic speed, travel time, and other external factors is mandatory for forecasting congestion levels in a dynamic road network. In recent years, sensor-based traffic detection technologies have evolved to achieve accuracy and safety in intelligent transportation systems (ITS). However, some of them need more reliability, have weak usage in different visibility conditions, are hard to set, and have a high cost. Alternatively, modern navigation systems are developed to find the fastest, alternative routes, travel times, and congestion levels for a given source to destination. These systems use Global Positioning Systems (GPS) data to provide the routing information. Due to inaccurate/ inadequate GPS data samples, sometimes these systems may give misleading information. This work presents Long-Range Wide Area Network (LoRaWAN) architecture to overcome the abovementioned limitations. In our proposed case study, we have used the Dragino LoRaWAn Gps sensor (IN865), air quality sensor (MQ135), temperature and humidity sensor (SHT31), and other sensors to extract direct and indirect factors that influence traffic congestion. Haversine formula and hyper-heuristic-based Encoder-Decoder using Gated Recurrent Units with attention mechanism (ED-GRUAT) are used to estimate the segment-based distance, speed, vehicle travel times, and congestion levels. Our case study helps to calculate accurate traffic speed, travel times, and congestion levels.

In this study, we investigated an alternative technique to automatically classify road traffic congestion with travelers’ opinions. The method utilized an intelligent traffic camera system orchestrated with an interactive web survey system to collect the traffic conditions and travelers’ opinions. A large numbers of human perceptions were used to train the artificial neural network (ANN) model and the decision tree (J48) model that classify velocity and traffic flow into three congestion levels: light, heavy, and jam. The both model was then compared to the Occupancy Ratio (OR) technique, currently in service in the Bangkok Metropolitan Administration (BMA). The accuracy of ANN was more than accuracy of the J48. The evaluation indicated that our ANN model could determine the traffic congestion levels 12.15% more accurately than the existing system. The methodology, though conceived for use in Bangkok, is a general Intelligent Transportation System (ITS) practice that can be applied to any part of the world.KeywordsTraffic congestion level determinationintelligent transportation system (ITS)human judgmentartificial neural network (ANN)decision tree (J48)occupancy ratio (OR)

Since the weather condition can be a cause of serious traffic congestion, it is necessary to establish a methodology to forecast future traffic congestions caused by rainfall and snowfall. However, there are few studies with simple methods that are applicable for practitioners such as road administrators. Therefore, in this paper we challenged to construct a statistical model to predict locations and levels of traffic congestion in a city, using only existing data that is open to the public. We collected hourly precipitation amount, hourly snowfall amount and cumulative snowfall amount from the Japan Meteorological Agency as weather observation data and images of Google Maps as traffic congestion data. As a result of the correlation analysis, we found that the hourly precipitation amount and the hourly snowfall amount did not correlate much with the relative congestion level whereas the correlation between the cumulative snowfall amount and 18-hour snowfall amount was found to be high. Consequently, a logistic regression analysis was conducted to explain the relative congestion level at various points on the roads using the 18-h snowfall amount and the cumulative snowfall amount. As a result, the model demonstrated good performance to reproduce the occurrence of increase in traffic congestion levels with >80% hit rates. In future, we would like to improve the present model to forecast potential road congestion based on weather forecast by using highly accurate weather information and longer term data.

Traffic congestion poses a persistent and escalating problem for major cities in both developed and developing countries, exerting a direct impact on the economic growth and development of these urban areas. Quantifying the extent of traffic congestion is the fundamental initial step in comprehending the severity of traffic congestion in order to devise effective methods for alleviation. The city of Addis Ababa is currently experiencing significant traffic congestion at its main intersections. The primary aim of this research is to assess the current level of traffic congestion at specific intersections. The assessment of traffic congestion was conducted using the travel time method. Data on travel time, traffic volume, and travel speed were gathered at three blocks and two intersections using a combination of quantitative and qualitative data collection methods. The travel rate, delay rate, and total travel delay (in vehicle-minutes) were computed. The total vehicle-minute delay for the selected three segments is estimated to be approximately 12,708 vehicle-minutes (or 212 vehicle-hours). The text reveals the significance of measuring the various components of traffic congestion in order to ensure a sustainable traffic system. It also highlights the importance of maintaining a satisfactory level of service for the future sustainability of City.

COVID-19 has changed the landscape within which we travel. Working from Home (WFH) in many countries has increased significantly, and while it was often forced on a society it has delivered some unintended positive consequences associated in particular with the levels of congestion on the roads and crowding on public transport. With a likelihood of some amount of WFH continuing as we move out of the active COVID-19 period, the question being asked is whether the post-COVID-19 period will return the pre-COVID-19 levels of traffic congestion and crowding. In many jurisdictions, there is a desire to avoid this circumstance and to use WFH as a policy lever that has appeal to employees, employers and government planning agencies in order to find ways of better managing future levels of congestion and crowding. This chapter uses the ongoing research and surveys we have been undertaking in Australia since March 2020 to track behavioural responses that impact on commuting and non-commuting travel, and to examine what the evidence tells us about opportunities into the future in many geographical settings to better manage congestion and crowding. This is linked to a desire by employers to maintain WFH where it makes sense as a way of not only supporting sustainability charters but also the growing interest in a commitment to a broader social licence. We discuss ways in which WFH can contribute to flattening peaks in travel; but also the plans that some public transport authorities are putting in place to ensure that crowding on public transport is mitigated as people increasingly return to using public transport. Whereas we might have thought that we now have plenty of public transport capacity, this may not be the case if we want to control crowding, and more capacity may be needed which could be a challenge for trains more than buses given track capacity limits. We conclude the chapter by summarising some of the positive benefits associated with WFH, and the implications not just for transport but for society more widely.

Crowd-shipping, employing private drivers to partially replace company-owned trucks in distribution, has emerged as a prominent trend for its cost-effectiveness and sustainability. While crowd-shipping is known as a distribution pattern that combines economic efficiency and environmental benefits, however, the frequent occurrence of traffic congestion has made this pattern less effective than it should be. In this research, the problem of vehicle routing optimization under traffic congestion is investigated from the perspective of simultaneously reducing environmental pollution and costs. Considering private drivers picking up and delivering parcels on the way, this study incorporates the objective of minimizing transport as well as particulate matter (PM) and nitrogen oxides (NOx) emission costs into route optimization for crowd-shipping and proposes a Green Pickup and Delivery Problem with Private Drivers (GPDP-PD). To be more realistic, vehicle speeds depend on the level of traffic congestion, reflecting the time-dependent nature of the proposed model. An improved adaptive large neighborhood search (ALNS) algorithm is developed, and computational experiments are conducted to demonstrate the efficiency of the improved ALNS. Case studies show that there is uncertainty about the environmental benefits of crowd-shipping under traffic congestion. Our proposed model is capable of efficiently allocating private drivers and optimizing vehicle routes according to road conditions, thus identifying the crowd-shipping operational scheme with the lowest cost and emissions. Moreover, a time limit of 0.7-0.8 h and the low cost of private drivers can achieve environmental and economic benefits simultaneously. It provides useful insights into the sustainability of logistics and distribution.

Lane-changing behaviour is one of the complex driving behaviours. The lane-changing behaviour of drivers may exacerbate congestion, however driver behavioural characteristics are difficult to accurately acquire and quantify, and thus tend to be simplified or ignored in existing lane-changing models. In this paper, the Bik-means clustering algorithm is used to analyse the urban road congestion state discrimination method. Then, simulated driving tests were conducted for different traffic congestion conditions. Through the force feedback system and infrared camera, the data of driver lane-changing behaviours at different traffic congestion levels are obtained separately, and the definitions of the start and end points of a vehicle changing lanes are determined. Furthermore, statistical analysis and discussion of key feature parameters including driver lane-changing behaviour data and visual data under different levels of traffic congestion were conducted. It is found that the average lane-change intention times in each congestion state are 2 s, 4 s, 6 s and 7 s, while the turn-signal duration and the number of rear-view mirror observations have similar patterns of change to the data on lane-changing intention duration. Moreover, drivers’ pupil diameters become smaller during the lane-changing intention phase, and then relatively enlarge during lane-changing; the range of pupil variation is roughly 3.5 mm to 4 mm. The frequency of observing the vehicle in front of the target lane increased as the level of congestion increased, and the frequency of observation in the driver's mirrors while changing lanes approximately doubled compared to driving straight ahead, and this ratio increased as the level of congestion increased.

Urban traffic congestion has become an important issue not only affecting our daily lives, but also limiting economic development. The primary cause of urban traffic congestion is that the number of vehicles is higher than the permissible limit of the road. Previous studies have focused on dispersing traffic volume by detecting urban traffic congestion zones and predicting future trends. However, to solve the fundamental problem, it is necessary to discover the cause of traffic congestion. Nevertheless, it is difficult to find a research which presents an approach to identify the causes of traffic congestion. In this paper, we propose a technique to analyze the cause of traffic congestion based on the traffic flow theory. We extract vehicle flows from traffic data, such as GPS trajectory and Vehicle Detector data. We detect vehicle flow changes utilizing the entropy from the information theory. Then, we build cumulative vehicle count curves (N-curve) that can quantify the flow of the vehicles in the traffic congestion area. The N-curves are classified into four different traffic congestion patterns by a convolutional neural network. Analyzing the causes and influence of traffic congestion is difficult and requires considerable experience and knowledge. Therefore, we present a visual analytics system that can efficiently perform a series of processes to analyze the cause and influence of traffic congestion. Through case studies, we have evaluated that our system can classify the causes of traffic congestion and can be used efficiently in road planning.

Historical traffic sensor data–including speed, counts, and occupancy–was used to estimate levels of short-term traffic congestion in Las Vegas, Nevada. Various data-mining techniques were tested, including the J48 decision tree, artificial neural networks, the support vector machine, PART, and K-nearest-neighbors. Future time slots were categorized as congested or uncongested based on the corresponding historical data. In this study, one-minute traffic data was provided by the Freeway and Arterial System of Transportation in Las Vegas, Nevada. A 10-mile northbound segment of Interstate 15 from Interstate 215 to Desert Inn Road was evaluated. Among the algorithms tested, the J48 algorithm provided the best performance, and was able to estimate congestion up to nine minutes into the future with an accuracy of approximately 80%.

As the number of vehicles grows rapidly each year, more and more traffic congestion occurs, becoming a big issue for civil engineers in almost all metropolitan cities. In this paper, we propose a novel pheromone-based traffic management framework for reducing traffic congestion, which unifies the strategies of both dynamic vehicle rerouting and traffic light control. Specifically, each vehicle, represented as an agent, deposits digital pheromones over its route, while roadside infrastructure agents collect the pheromones and fuse them to evaluate real-time traffic conditions as well as to predict expected road congestion levels in near future. Once road congestion is predicted, a proactive vehicle rerouting strategy based on global distance and local pheromone is employed to assign alternative routes to selected vehicles before they enter congested roads. In the meanwhile, traffic light control agents take online strategies to further alleviate traffic congestion levels. We propose and evaluate two traffic light control strategies, depending on whether or not to consider downstream traffic conditions. The unified pheromone-based traffic management framework is compared with seven other approaches in simulation environments. Experimental results show that the proposed framework outperforms other approaches in terms of traffic congestion levels and several other transportation metrics, such as air pollution and fuel consumption. Moreover, experiments over various compliance and penetration rates show the robustness of the proposed framework.