Real-time estimation of secondary crash likelihood on freeways using high-resolution loop detector data

A hybrid machine learning model for predicting Real-Time secondary crash likelihood

Nonrecurring traffic incidents, such as motor vehicle crashes, increase not only travel delays but also the risk of secondary crashes. Secondary crashes can cause additional traffic delays and reduce safety. Implementation of effective countermeasures to prevent or reduce secondary crashes requires that their characteristics be investigated. However, the related research has been limited, largely because of the lack of detailed incident and traffic data necessary to identify secondary crashes. Existing approaches, such as static methods employed to identify secondary crashes, cannot fully capture potential secondary crashes because of fixed spatiotemporal identification criteria. Improved approaches are needed to categorize secondary crashes accurately for further analysis. This paper develops an enhanced approach for identifying secondary crashes that uses the existing crash database and archived traffic data from highway sensors. The proposed method is threefold: (a) defining secondary crashes, (b) examining the impact range of primary crashes that possibly relate to secondary crashes, and (c) identifying secondary crashes. The proposed methodology establishes a practical framework for mining secondary crashes from existing sensor data and crash records. A case study was performed on a 27-mi segment of a major highway in New Jersey to illustrate the performance of the proposed approach. The results show that the proposed method provides a more reliable and efficient categorization of secondary crashes than commonly used approaches.

A two-step identification method of secondary crashes (SCs) is developed in this paper, and the effects of traffic variables on the SC risks on freeways are captured. Crash-related data were obtained from the I-880 freeway in California, USA. Combined with the speed contour map, the SCs were identified by the real shock wave speed index and the SC identification index. The random effect logit regression was applied to analyze the significantly contributing factors and their mechanism on the SCs. The results showed that the significantly contributing factors are different between the initial SC model and SC models (Threshold = 20mile/h, 15mile/h, 10mile/h, 5mile/h), except for the variable of the standard deviation of 30-s vehicle speed during 5-10 min. In the SC models, the number of significant contributing factors increases with the increase in the threshold value. The results of the elasticity analysis showed that the elasticity values of the hit object crash, wet surface road, the average 30-s vehicle speed during 5-10 min, and the average 30-s vehicle count during 5-10 min are greater than 10%.

A hybrid short-term traffic flow forecasting method based on spectral analysis and statistical volatility model

Secondary crashes are generally understood as crashes that occur as a result of primary incidents. Research is needed to determine the nature and cause of these crashes. However, to date, this research has been limited for two reasons. First, there is no uniform definition of secondary crashes. Second, the usually poor quality of incident data and the unavailability of related traffic data make any analysis of secondary crashes difficult. This paper describes a study that uses a comprehensive incident database from District 4 of the Florida Department of Transportation to identify freeway secondary crashes and their contributing factors. A method based on a cumulative arrival and departure traffic delay model was developed to estimate the maximum queue length and the associated queue recovery time for incidents with lane blockages. Descriptive statistics and logistic regression analysis were applied to understand the factors contributing to secondary crashes. The logistic regression analysis indicates that the following four factors have significant effects on the likelihood of secondary crashes: primary incident type, primary incident lane-blockage duration, time of day, and whether the incident occurred on northbound I-95.

Urban road congestion is increasingly serious, in order to optimize the traffic situation, and improve the efficiency of vehicles, so as to provide service for traveling user. Then analysis four major systems, namely the bus electronic station board system, evaluation system of urban road traffic state, urban traffic guidance system and website demonstration system. Regarding average travel speed, congestion coefficient and mean velocity gradient as evaluation index of evaluation system, to simulate the bus electronic station board to obtain corresponding real-time traffic flow data. A website design platform is used for dynamic demonstration of urban road traffic state, and the evaluation result is determined. Finally, the valid data will be released to the scene of traffic for reasonable traffic grooming.

Past studies on automatic traffic incident detection have mainly focused on the incidents on freeways, which are controlled-access roads. There are not many works on urban traffic incident detection. In addition, the traffic data used in detecting an incident are still mostly collected from fixed sensors such as loop detectors. With the advances in mobile sensing and vehicular technology, it is foreseeable that mobile sensors will be used increasingly in the near future. In fact, traffic data will be collected directly by vehicles. Detecting traffic incidents in an urban road network with the traffic data collected by mobile sensors poses several challenges. First, the urban roads are uncontrolled-access roads, which are typically full of flow-disruptive entities such as traffic signals, intersections, crossings, bus stops, etc. These entities can disrupt the traffic flow in a similar way that an incident does, making it more difficult to detect an incident. Second, it is still not clear which traffic variables, collected by mobile sensors, can be used in detecting an incident in an urban environment. In this paper, we investigate and identify the traffic variables that are effective in detecting an incident in an urban road network. Particularly, speed, acceleration, lane-change ratio and travel time are studied. The results show that these four traffic variables are generally effective for traffic incident detection. However, among the four variables, the least effective one is the travel time.

**Traffic increases have made road infrastructure construction difficult for transport planners and operators, risking pavement collapse or resource waste. Egyptian roads have deteriorated due to traffic loads, requiring periodic maintenance and reconstruction. Traffic data is scarce in developing countries like Egypt, making these analyses difficult. This study estimates traffic growth rates across the country using 2008–2018 Egyptian data to address traffic forecasting difficulties. Annual average daily traffic (AADT), population density, and other socioeconomic variables are used in regression analysis to estimate traffic increases. The AADT socio-economic relationship was substantial, with adjusted R2 values from 0.74 to 0.99. Comparing predicted traffic to last year's traffic data at each station in the research showed appropriate results for the proposed models. AADT error rates for proposed models and true values were 0.00–0.025. Thus, traffic growth rate data can improve and develop new roads in such regions. These findings enhance planning accuracy for future projects.

Traffic data are the fundamental inputs to traffic flow analysis and simulation studies, which facilitate decision making in the fi eld of traffic engineering. Hence, the accuracy of traffic data is of paramount importance. This study compares new technologies available for traffic data collection considering their accuracy and applicability in the Sri Lankan context. Traffic in Sri Lanka is of heterogeneous nature, as opposed to the homogeneous nature observed in most developed countries. Hence, collection of traffic data poses several challenges that affects its accuracy. Three techniques, the infrared driven TIRTL instrument, the video image processing-based TRAZER application and the traffic data collection method using the Google distance matrix application programming interface (API), with respect to their data collection accuracy are reviewed in this study. The fundamental macroscopic traffic data variables (fl ow and speed) were evaluated against control surveys. It was found that each technology has its strengths and weaknesses and needs to be used appropriately. The TIRTL instrument fared better on road sections on level terrain when the crossfall did not obstruct the infrared beams. Such occasions provided a rich set of microscopic traffic data. The TRAZER software delivered data up to a 100 % accuracy. However, this required the user to go through a lengthy postprocessing routine to extract the final set of traffic data. Google traffic data collection provides highly accurate results when estimating link speeds. This method is ideal for collection of bulk data with spatio-temporal variations and the process can be fully automated to reduce the human resource requirement.

Real-time and accurate prediction about current and future traffic conditions is one of the effective ways to alleviate traffic problems. However, missing data problem is inevitable for various reasons when obtaining real-time traffic flow information. Incomplete traffic information may seriously affect the prediction accuracy. To address this problem, in this paper, we propose a method that predicts the traffic flow in real time under missing data. Considering the spatio-temporal characteristics of traffic flows and the spatial location of road segments, we first evaluate the importance of traffic flows using a spatio-temporal correlation function to analyze the correlations of traffic flows. Then, we present a PPCA-based minimum data imputation optimization (P-MDIO) algorithm to reduce computation time of data imputation. Finally, we utilize the complete traffic data and relevant road segments sequences to predict real-time traffic flows. The experimental data are obtained from the real-time traffic data collected by loop detectors in Taipei, Taiwan. Our experimental results show the performance and validity of the proposed approach, particularly in large-scale prediction.

In recent years, tensor has been used in traffic flow data processing. However, the influx of the large number of real-time traffic flow data will cause dimension disaster and computational explosion. In order to deal the problem, this paper presents a Cyber-Physical-Social tensor framework (CPST) and an incremental SVD-based tensor train decomposition algorithm (ITTD), which effectively solves the problem of integrating incremental traffic flow data with original traffic flow data. Firstly, the singular value decomposition (SVD) is used to extract the core tensor of the new traffic flow data, and then the zero-padding operation is performed on the decomposed tensor-train (TT) to make it in the same dimension with the original tensor. After that the new data tensor is fused with the original data tensor. Finally, the fused data tensor train is obtained by orthogonal compression, and the superiority of the algorithm is verified on the UK Highway data set. Our work provides a new data model for traffic flow processing.

This paper presents an adaptive traffic signal controller algorithm that allows real time traffic flow data to control traffic signal timing for an isolated intersection. The proposed design allows greater flexibility and control of signal timings while keeping in view multiple criterions. It is based on simple rules that allow traffic flows on all intersection branches to control the timing of current signal phase without causing unrealistic delays on low volume side streets. The algorithm is also adjusted for capacity of different links. Detector data has been used to assess congestion situation and respond accordingly. A priority scheme for priority vehicles is also proposed. As a proof of concept, the proposed algorithm has been implemented on FPGA as a parameterized VERILOG model allowing adjustment to any intersection size and topology. An evaluator circuit was designed to assess the performance as compared to actuated signal controller. Improvement was observed in total stopped delay, number of stops and queue lengths.

Real time traffic flow data in terms of road-segments traffic densities are relevant information at the basis of several smart services, such as smart routing, smart planning for evacuations, planning of civil works on the city, etc. The traditional methods for traffic flow measured via sensors using data from navigator Apps (e.g., TomTom, Google map, Waze) could be very expensive to be acquired. For this reason, there is the space for low cost and fast solutions for dense traffic flow reconstruction from scattered data coming from a limited number of street sensors spread on the city. The proposed method is based on differential equations and physical constraints applied to a detailed street graph which is enriched of several features. A stochastic learning approach has been adopted to estimate the weights representing in certain sense the road-segments capacity at each time slot of the day. The proposed solution allows computing in real-time the traffic density reconstruction in unmeasured road-segments. The solution has been validated estimating the error in the places where the sensors are positioned, excluding each of them iteratively and reconstructing the flow without it. Then, it has been possible to estimate the error between the reconstructed traffic density and the measured values. This approach allowed setting up a converging algorithm for estimating the traffic density in the whole city graph from detailed parameters. The proposed reconstruction model has been created by exploiting open and real-time data in the context of Sii-Mobility research project by using Km4City infrastructure in the area of Florence, Italy, for its corresponding Smart City solution.

Traffic congestion on major roads consequent upon existing bottlenecks at intersections is a major problem in Akure Metropolis. To change this trend, this research was carried out in order to design effective traffic control mechanisms at the “troubled spots” in the metropolis. Odole intersection was identified as one of the critical intersections during a reconnaissance survey and as such, selected for study. Data on geometric features were collected using Oedometer and Google Earth software. Peak and off-peak traffic volume data were collected between 7:30 and 8:30 am and between 12.00 noon and 1.00 pm respectively every other day using Cine camera. Furthermore, discharge headway and delay data were collected using stop watch. The geometric and traffic data collected were analysed using Microsoft Excel. An appraisal of Odole Intersection indicated that the major contributors to traffic are Motorcycles (70.88%) and Passenger Cars (28.72%). Other modes of transportation account for about 0.4% of vehicles traversing the intersection. The critical traffic volume at the intersection was over 4000 veh/hr and the average delay was 22 seconds. An Average delay of 22 seconds at the intersection was an indication that the operating level of service was C (i.e. fairly stable traffic condition with average delays attributable to traffic control by personnel). By juxtaposing the results of geometric and traffic data analyses with the pros and cons of various traffic control mechanisms, traffic control by signalisation was selected and designed to suit Odole intersection. In tandem with the results of this research, appropriate measures have been recommended to ameliorate traffic and commuting problems in the metropolis.

There is a need to better understand and quantify the safety implications of geometric improvements and countermeasures. A commonly accepted measure of safety effectiveness is the crash modification (CMF) defined as the expected number of crashes with a countermeasure divided by the number expected without the countermeasure. This paper evaluates the use of a case-control design to directly estimate the safety effectiveness of lane and shoulder width using the odds ratio: the probability of a crash with the countermeasure divided by the probability of a crash without a countermeasure. The case-control design is well established in epidemiology where it is used to relate factors within a study population to a particular or disease. In the highway safety context, the outcome is defined as a crash, the risk factor is a particular geometric feature or countermeasure and the subjects are roadway segments. This paper describes the of an experiment in which commonly available crash, roadway and traffic data are used to estimate odds ratios which are compared to CMFs from the literature. The comparison of these slightly different effectiveness measures are intended as a test of the efficacy of the case-control method in safety effectiveness evaluation. Geometric, traffic and crash data were obtained for more than 28,000 rural two-lane undivided highway segments in Pennsylvania for years 1997 to 2001 inclusive. A matched case-control design is developed to estimate the safety effectiveness of different lane and shoulder widths while adjusting for the confounding variables: speed limit, average daily traffic and segment length. Conditional logistic regression models are developed to estimate the odds ratio for lane and shoulder width separately. Lane and shoulder width odds ratios compare favorably to CMFs from the literature and have the additional advantage of including confidence intervals on the estimates. The paper concludes with a discussion of strengths and weaknesses of the method and recommendations for future research.