Secondary Crashes Research Articles

Multi-level time-varying causality analysis of secondary conflicts based on hazard-based duration models

Secondary conflicts occur frequently and would cause multi-vehicle collisions. In order to prevent multi-vehicle collisions, a better understanding of the factors that affect secondary conflict propagation is crucial. Previous studies have identified the influencing factors of primary conflicts’ occurrence, but have not explored the time-varying factors that affect secondary conflicts’ propagation. In addressing this gap, about 20,000 secondary conflicts are extracted from real trajectory dataset, and a multi-level variable system is established, including segment types, traffic status, front chain conflict status, and direct interaction behaviors. Further, a Kaplan-Meyer model and a random parameters hazard-based duration model are constructed to explore the single-factor and multiple-factor influence on the propagation of secondary conflicts, respectively. The results suggest that the first 2.6 s after a conflict is a critical post-monitoring period to prevent the secondary conflicts propagation. In addition, diverging and merging segments shorten the survival time of secondary conflicts by about 12%, indicating a higher occurrence probability of secondary conflicts near the ramps of expressways. More importantly, the front chain conflict status and the front direct conflict status reveal a different effect on the secondary conflicts. The high risk of chain conflict ahead would increase the occurrence probability of secondary conflicts, while the high risk of front conflict would decrease the probability. Overall, this research is of great significance to understand the influencing factors of secondary conflict and avoid secondary crashes.

Severity Analysis of Secondary Crashes on High-Speed Roadways: Pattern Recognition Using Association Rule Mining

Secondary crashes (SCs) are a major concern, posing additional safety threats to both non-involved vehicles and incident responders. The objective of this study was to identify the affiliated factors contributing to SCs on roadways with a speed limit of 55 mph or above. Traditional police-investigated crash dataset was analyzed, spanning more than four years (January 2016–February 2020) for the entire state of Alabama. As the crash database did not directly include information on SCs and did not allow for linking a primary crash with a subsequent SC, a data extraction process was developed to identify SCs and understand their characteristics. Association rule mining (ARM) was applied to identify crash patterns based on maximum injury severity levels. The generated rules were filtered based on support, confidence, and lift, and then validated by the lift increase criterion. The results revealed complex relationships between risk factors and severity of SCs. In relation to SCs with injuries, single-vehicle crashes were frequently observed during peak hours and when drivers swerved to avoid objects/persons/vehicles. In contrast, concerning SCs with possible/no injuries, single-vehicle collisions were more likely to occur when drivers failed to notice objects/persons/vehicles and were involved in speeding. On urban interstates, single-vehicle SCs were frequently associated with injuries, while rear-end SCs were often linked to possible/no injuries. The findings of this study can be helpful in enhancing existing traffic incident management programs to mitigate the occurrence of SCs.

Point-process modeling of secondary crashes

Secondary crashes or crashes that occur in the wake of a preceding or primary crash are among the most critical incidents occurring on highways, due to the exceptional danger they present to the first responders and victims of the primary crash. In this work, we developed a self-exciting temporal point process to analyze crash events data and classify it into primary and secondary crashes. Our model uses a self-exciting function to describe secondary crashes while primary crashes are modeled using a background rate function. We fit the model to crash incidents data from the Florida Department of Transportation, on Interstate-4 (I-4) highway for the years 2015–2017, to determine the model parameters. These are used to estimate the probability that a given crash is secondary crash and to find queue times. To represent the periodically varying traffic levels and crash incidents, we model the background rate, as a stationary function, a sinusoidal non-stationary function, and a piecewise non-stationary function. We show that the sinusoidal non-stationary background rate fits the traffic data better and replicates the daily and weekly peaks in crash events due to traffic rush hours. Secondary crashes are found to account for up to 15.09% of traffic incidents, depending on the city on the I-4 Highway.

Application of an Effective Hierarchical Deep-Learning-Based Object Detection Model Integrated with Image-Processing Techniques for Detecting Speed Limit Signs, Rockfalls, Potholes, and Car Crashes

Against the backdrop of rising road traffic accident rates, measures to prevent road traffic accidents have always been a pressing issue in Taiwan. Road traffic accidents are mostly caused by speeding and roadway obstacles, especially in the form of rockfalls, potholes, and car crashes (involving damaged cars and overturned cars). To address this, it was necessary to design a real-time detection system that could detect speed limit signs, rockfalls, potholes, and car crashes, which would alert drivers to make timely decisions in the event of an emergency, thereby preventing secondary car crashes. This system would also be useful for alerting the relevant authorities, enabling a rapid response to the situation. In this study, a hierarchical deep-learning-based object detection model is proposed based on You Only Look Once v7 (YOLOv7) and mask region-based convolutional neural network (Mask R-CNN) algorithms. In the first level, YOLOv7 identifies speed limit signs and rockfalls, potholes, and car crashes. In the second level, Mask R-CNN subdivides the speed limit signs into nine categories (30, 40, 50, 60, 70, 80, 90, 100, and 110 km/h). The images used in this study consisted of screen captures of dashcam footage as well as images obtained from the Tsinghua-Tencent 100K dataset, Google Street View, and Google Images searches. During model training, we employed Gaussian noise and image rotation to simulate poor weather conditions as well as obscured, slanted, or twisted objects. Canny edge detection was used to enhance the contours of the detected objects and accentuate their features. The combined use of these image-processing techniques effectively increased the quantity and variety of images in the training set. During model testing, we evaluated the model’s performance based on its mean average precision (mAP). The experimental results showed that the mAP of our proposed model was 8.6 percentage points higher than that of the YOLOv7 model—a significant improvement in the overall accuracy of the model. In addition, we tested the model using videos showing different scenarios that had not been used in the training process, finding the model to have a rapid response time and a lower overall mean error rate. To summarize, the proposed model is a good candidate for road safety detection.

Secondary crash mitigation controller after rear-end collisions using reinforcement learning

Rear-end collisions result in a large number of casualties and property losses, and the serious injury risk in multiple impact accidents is much higher than that in single impact accidents. In this paper, we propose a novel controller to facilitate the prevention of secondary crashes after an initial rear-end collision, which expands the operational horizon of conventional vehicle active safety systems from preventive measures to post-event mitigation measures. Considering the complexity of the problem with multi-object synthesis optimization and vehicle nonlinear dynamics, this study combines the pre-collision control and post-collision control to reduce the initial crash loss and the subsequent control difficulty. The rule-based switching control and drift manipulation are embedded into the reinforcement learning algorithm to improve the training efficiency and control performance. The bench test results validate the superiority of the proposed controller over other strategies and algorithms in different rear-end collision scenarios.

Spatial and temporal prediction of secondary crashes combining stacked sparse auto-encoder and long short-term memory

Secondary crashes occur within the spatial and temporal impact area of primary crashes, resulting in traffic delays and safety problems. While most existing studies focus on the likelihood of secondary crashes, predicting the spatio-temporal location of secondary crashes could offer valuable insights for implementing prevention strategies. This includes guiding the deployment of emergency response measures and determining appropriate speed limits. The main objective of this study is to develop a prediction method for the spatial and temporal locations of secondary crashes. A hybrid deep learning model SSAE-LSTM is proposed by combining stacked sparse auto-encoder (SSAE) and long short-term memory network (LSTM). Traffic and crash data on the California I-880 highway covering the period of 2017–2021 are collected. The identification of secondary crashes is performed by the speed contour map method. The time and distance gaps between primary and secondary crashes are modeled using multiple 5-minute interval traffic variables as inputs. Multiple models are developed for benchmarking purposes, including PCA-LSTM, which incorporates principal component analysis (PCA) and LSTM, SSAE-SVM, which incorporates SSAE and support vector machine (SVM), and back propagation neural network (BPNN). The performance comparison indicates that the hybrid SSAE-LSTM model outperforms the other models in terms of both spatial and temporal prediction. In particular, SSAE4-LSTM1 (with 4 SSAE layers and 1 LSTM layer) demonstrates superior spatial prediction performance, while SSAE4-LSTM2 (with 4 SSAE layers and 2 LSTM layers) excels in temporal prediction. A joint spatio-temporal evaluation is also conducted to measure the overall accuracy of the optimal models over different permitted spatio-temporal ranges. Finally, practical suggestions are provided for secondary crash prevention.

Exploring the spatiotemporal heterogeneity of freeway secondary crashes using GTWR model

Understanding the effect of relevant variables on secondary crashes could provide reliable traffic intervention evidence for decision makers, which ultimately would contribute to the prevention of secondary crashes. Road crashes present obvious spatiotemporal heterogeneity, and previous studies (e.g., geographically weighted regression [GWR] based method) only analyzed the spatial heterogeneity without considering the timescale of related variables. This study proposes a geographically and temporally weighted regression model (GTWR), which adds a time variation to the conventional GWR model. The GTWR model is used to explore the spatiotemporal effects of relevant variables on the secondary crashes on expressways. The empirical study using expressways in Anhui, China, illustrates that the density of rainy days, snow and ice, and occupied middle lanes are negatively associated with secondary crashes. Then the temporal variation characteristics of each influencing factor at different timescales (e.g., weekdays, weekends, and time of day such as night, morning, afternoon, and evening) are analyzed, which reveals the spatial distribution characteristics of the influencing factors in each time period. The finding indicates that the explanatory variables have heterogeneous effects on the frequency of secondary crashes. The model comparison further demonstrates that the proposed GTWR model outperforms the ordinary least squares (OLS) and GWR methods in data fitting and spatiotemporal modeling. The research findings can aid in prevention of secondary crashes on freeways.

Time and Distance Gaps of Primary-Secondary Crashes Prediction and Analysis Using Random Forests and SHAP Model

Secondary crashes (SCs) are typically defined as the crash that occurs within the spatiotemporal boundaries of the impact area of the primary crashes (PCs), which will intensify traffic congestion and induce a series of road safety issues. Predicting and analyzing the time and distance gaps between the SCs and PCs will help to prevent the occurrence of SCs. In this paper, a combined data-driven method of static and dynamic approaches is applied to identify SCs. Then, the random forests (RF) method is implemented to predict the two gaps using temporal, primary crash, roadway, and real-time traffic characteristics data collected from 2016 to 2019 at California interstate freeways. Subsequently, the SHapley Additive explanation (SHAP) approach is employed to interpret the RF outputs. The results show that the traffic volume, speed, lighting, and population are considered the most significant factors in both gaps. Furthermore, the main and interaction effects of factors are also quantified. High volume possibly promotes the time and distance gaps, while low volume inhibits them. And volume affects the distance gap inconsiderably when it falls between 300 and 400 veh/5 min. Traffic conditions with high speed and low volume are strongly associated with short-time and short-distance gaps. Darker surroundings probably accelerate the occurrence of SCs. Moreover, crashes involving the violation categories of improper turns or unsafe lane changes likely result in long time and distance gaps. These results have important implications for proposing traffic management and improving road safety.

Examining the characteristics between time and distance gaps of secondary crashes

Abstract Understanding the characteristics of time and distance gaps between the primary (PC) and secondary crashes (SC) is crucial for preventing SC ccurrences and improving road safety. Although previous studies have tried to analyse the variation of gaps, there is limited evidence in quantifying the relationships between different gaps and various influential factors. This study proposed a two-layer stacking framework to discuss the time and distance gaps. Specifically, the framework took random forests (RF), gradient boosting decision tree (GBDT) and eXtreme gradient boosting as the base classifiers in the first layer and applied logistic regression (LR) as a combiner in the second layer. On this basis, the local interpretable model-agnostic explanations (LIME) technology was used to interpret the output of the stacking model from both local and global perspectives. Through SC dentification and feature selection, 346 SCs and 22 crash-related factors were collected from California interstate freeways. The results showed that the stacking model outperformed base models evaluated by accuracy, precision, and recall indicators. The explanations based on LIME suggest that collision type, distance, speed and volume are the critical features that affect the time and distance gaps. Higher volume can prolong queue length and increase the distance gap from the SCs to PCs. And collision types, peak periods, workday, truck involved and tow away likely induce a long-distance gap. Conversely, there is a shorter distance gap when secondary roads run in the same direction and are close to the primary roads. Lower speed is a significant factor resulting in a long-time gap, while the higher speed is correlated with a short-time gap. These results are expected to provide insights into how contributory features affect the time and distance gaps and help decision-makers develop accurate decisions to prevent SCs.

Modelling and Mitigating Secondary Crash Risk for Serial Tunnels on Freeway via Lighting-Related Microscopic Traffic Model with Inter-Lane Dependency.

This paper models and mitigates the secondary crash (SC) risk for serial tunnels on the freeway which is incurred by traffic turbulence after primary crash (PC) occurrence and location-heterogeneous lighting conditions along serial tunnels. A traffic conflict approach is developed where SC risk is quantified using a surrogate safety measure based on the simulated vehicle trajectories after PC occurs from a lighting-related microscopic traffic model with inter-lane dependency. Numerical examples are presented to validate the model, illustrate SC risk pattern over time, and evaluate the countermeasures for SC, including adaptive tunnel lighting control (ATLC) and advanced speed and lane-changing guidance (ASLG) for connected vehicles (CVs). The results demonstrate that the tail of the stretching queue on the PC occurrence lane, the adjacent lane of the PC-incurred queue, and areas near tunnel portals are high-risk locations. In serial tunnels, creating a good lighting condition for drivers is more effective than advanced warnings in CVs to mitigate SC risk. Combined ATLC and ASLG is promising since ASLG informs CVs of an immediate response to traffic turbulence on the lane where PC occurs and ATLC alleviates SC risks on adjacent lanes via smoothing the lighting condition variations and reducing inter-lane dependency.

Spatial and Temporal Prediction of Secondary Crashes Combining Stacked Sparse Auto-Encoder and Long Short-Term Memory

Spatial and Temporal Prediction of Secondary Crashes Combining Stacked Sparse Auto-Encoder and Long Short-Term Memory

Jointly analyzing freeway primary and secondary crash severity using a copula-based approach

A copula-based model is developed in this study to jointly model the severity of freeway primary crashes and secondary crashes. The copula-based model can concurrently account for the severity levels in the crash and the correlation among primary-secondary crash pairs’ severity. The model comprehensively considers a series of explanation variables, including temporal characteristics, crash characteristics, roadway characteristics and real-traffic conditions, and is estimated using traffic crash data from 2016 through 2019 for Los Angeles County, California. The proposed copula model is then contrasted with the traditional binary probit model and the results show a remarkable advantage of the copula model, which is evidenced by better fitting performance. It is found that weather, whether towed away, unsafe speed, collision type, road condition, terrain, road weaving and truck involvement have significant impact on primary crash severity propensity and collision type, road width, road condition, traffic volume and vehicle speed have significant impact on secondary crash severity propensity. In light of the findings, a number of countermeasures are proposed to mitigate freeway crashes, including emergency services, vehicle and roadway engineering, traffic law enforcement and driver education.

Assessment of Indiana Unmanned Aerial System Crash Scene Mapping Program

Many public safety agencies in the US have initiated a UAS-based procedure to document and map crash scenes. In addition to significantly reducing the time taken to document evidence as well as ensuring first responder safety, UAS-based mapping reduces incident clearance time and thus the likelihood of a secondary crash occurrence. There is a wide range of cameras used on these missions, but they are predominantly captured by mid-priced drones that cost in the range of $2000 to $4000. Indiana has developed a centralized processing center at Purdue University that has processed 252 crash scenes, mapped using 29 unique cameras, from 35 public agencies over the past three years. This paper includes a detailed case study that compares measurements obtained from a traditional ground-based real-time kinematic positioning base station and UAS-based photogrammetric mapping. The case study showed that UAS derived scale errors were within 0.1 ft (3 cm) of field measurements, a generally accepted threshold for public safety use cases. Further assessment was done on the 252 scenes using ground control scale error as the evaluation metric. To date, over 85% of the measurement errors were found to be within 0.1 ft (3 cm). When substantial errors are identified by the Purdue processing center, they are flagged for further dialog with the agency. In most of the cases with larger errors, the ground control distance was incorrectly measured, which is easily correctable by returning to the scene and performing new distance control measurements.

Examining the Influence of Work Zones on the Propensity of Secondary Crashes

Examining the Influence of Work Zones on the Propensity of Secondary Crashes

Characterizing Incident Responder Crashes Involving Move Over Law Violations

This research sought to understand secondary crashes involving incident responders in which a citation was issued for violation of Florida’s Move Over Law. From 2011 to 2020, there were 519 crashes involving a Move Over Law citation, for which the circumstances of the crashes had required a driver to slow or change lanes because of a stopped responder vehicle. The majority occurred during times other than daylight, and driver distraction and alcohol were notably present in the data. Alcohol increased crash and injury severity and the likelihood of injury. Law enforcement traffic stops were the most common precipitating activity, followed by previous crashes and vehicle disablements. Over two-thirds of crashes involved a law enforcement vehicle, the remainder comprised other responder vehicle types. A distribution of posted speed limits showed that more than half of move over violation crashes occurred on roadways with a posted speed of 45 mph or less. In 41 crashes, a pedestrian was struck, including 32 involving incident responders. In responder struck-by crashes, law enforcement officers, lack of high-visibility safety apparel, and operating on the traffic side of incident scenes were prominent. Local roadways comprised a significant number of struck-by crashes, but higher-speed roadways were more dangerous for towing operators and had higher incidence of serious injuries. The move over violation approach to filtering crash data for responder-involved incidents was valid, however, not every crash involving a responder resulted in a move over citation, so the approach was more representative of the responder-involved crash type than comprehensive.

Understanding drivers’ awareness, habits and intentions inside road tunnels for effective safety policies

Tunnels have a unique driving environment; thus, a small incident in a tunnel may result in severe consequences and a high probability of secondary crashes. Fortunately, studies have found that adopting safe driving behavior in a tunnel minimizes the severe outcomes of an incident. Therefore, implementing driver-oriented safety policies and conducting public awareness campaigns that emphasize safe behavior when driving through tunnels are essential. However, before devising policies and campaigns on the right issues, it is necessary to understand drivers' current level of knowledge regarding tunnel safety, their habits, behavioral intentions, and psychological condition while driving through tunnels. To achieve this objective, a sample of 841 responses was collected from China using a questionnaire survey consisting of fifty-two items. The results showed that several gaps exist in drivers' knowledge regarding tunnel safety and equipment. Drivers often adopt inappropriate habits and behaviors while driving through tunnels. Also, the tunnel environment has a significant influence on the psychological condition of the drivers. Moreover, drivers' demographic characteristics significantly affect their knowledge, reported habits and behavioral intentions, and psychological condition. The authorities and safety analysts could employ the suggestions highlighted in the present study for improving tunnel safety.

Exploring the impacts of traffic flow states on freeway normal crashes, primary crashes, and secondary crashes

AbstractThis study aims to explore the relationship between traffic flow states and crash type/severity in the scenarios of normal crashes, primary crashes, and secondary crashes using the association rules mining approach. The crash data and real‐time traffic data were collected from the I‐880 freeway for five years in California, USA. The secondary crashes were identified using a speed contour plot approach. Traffic flow states were identified by the three‐phase flow theory. The results showed that the free flow is associated with the proportion of the sideswipe normal crash, the hit object primary crash, and the injury primary crash. The synchronized flow, the wide moving jams, and the transitional state from synchronized flow to wide moving jams are associated with the proportion of the rear‐end secondary crash. The transitional state from synchronized flow to free flow is associated with the proportion of the rear‐end primary crash and the property damage only primary crash. In addition, the unsafe speed behaviour can increase the proportion of the rear‐end normal, primary, and secondary crashes. The unsafe lane change behaviour can increase the proportion of the sideswipe normal, primary, and secondary crashes. These results have the potential to reduce the secondary crash probability.

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

Secondary crashes usually occur within the spatio-temporal impact ranges of primary crashes, which could cause traffic disturbance and increase traffic safety problems. However, existing studies only focused on predicting the likelihood of crashes leading to secondary crashes without considering the likelihood of the occurrence of secondary crashes. In addition, previous studies did not consider the real-time implementation of secondary crash likelihood prediction models and included too many features that were not available in real-time. A real-time secondary crash likelihood prediction model aims to predict the likelihood of secondary crashes in a short period (e.g., 5–10 min) and update every minute. The main objective of this paper is to develop a machine learning model to predict secondary crash likelihood in real-time. Two XGBoost models were developed for predicting the likelihood of crashes leading to secondary crashes and the likelihood of the occurrence of secondary crashes, respectively. A hybrid model was proposed to combine the results from the two developed models. Results indicated that the proposed hybrid model significantly improved the accuracy of secondary crash likelihood prediction. The proposed model has the potential to be applied in proactive traffic safety management system and prevent the occurrence of secondary crashes. Moreover, experimental results suggested that several features related to real-time traffic flow conditions were crucial for predicting secondary crash likelihood, such as the average traffic volume and average occupancy.

Safety Performance Assessment of Connected Vehicles in Mitigating the Risk of Secondary Crashes: A Driving Simulator Study

Traffic crashes can be divided into primary and secondary crashes. Secondary crashes occur as a consequence of primary crashes within their spatiotemporal distances. Secondary crashes comprise nearly 20% of all crashes and 18% of fatal crashes, in which they can possibly have a higher crash severity than the primary crash. Interstate-80 in Wyoming is a major rural corridor with a high freight traffic volumes. The Federal Highway Administration selected Wyoming in which to deploy connected vehicle (CV) technology with a focus on commercial truck safety. Distress and rerouting applications were among the suite of CV pilot applications. Very few studies have investigated the safety performance of CVs in mitigating the risk of secondary crashes on heavy trucks, more specifically under adverse weather conditions. This study filled this gap by conducting a driving simulator experiment to assess the effectiveness of CV distress and rerouting applications in mitigating the effects of secondary crashes. A total of 23 truck drivers were recruited to this study. The analysis was conducted on the vehicle kinematics obtained from the driving simulator. A CV and a nonCV scenario were designed to compare the participants’ driving behavior under adverse weather conditions. The results showed that the tested CV applications succeeded in enhancing driving behaviors by reducing the operating speed as well as the speed variation, and all the participants avoided a secondary crash in the CV environment. In addition, the distress notification coupled with the road closure reduced the average operating speed by 26% from the provided speed limit.

Secondary Crash Identification using Crowdsourced Waze User Reports

Secondary crashes are crashes that occur as a result of the nonrecurrent congestion originating from primary crashes, and always have a greater impact on safety and traffic than a single crash. A better understanding of secondary crashes would benefit traffic incident management, and this requires accurate identification of secondary crashes. This study explores using crowdsourced Waze user reports to identify secondary crashes. A network-based clustering algorithm is proposed to extract the primary crash cluster, including all user reports originating from the primary crash, and any crash that occurred within the cluster would be a secondary crash. This method works as a filter to select accurate primary–secondary relationships, thus precisely identifying secondary crashes. A case study is performed with crashes occurring from June to December 2019 on a 30-mi stretch of I-40 in Knoxville, TN. A static threshold method (crash duration and 10 mi) was used to preselect the potential primary–secondary crash pairs, and 75 out of 708 crashes were identified as potential secondary crashes. Based on the preselected primary–secondary crash pairs, 17 secondary crashes were obtained with the proposed method and the results were compared with one of the commonly used methods, the speed contour plot method. Though the proposed method captured fewer secondary crashes, it did identify several secondary crashes that could not be observed with the speed contour plot method. The results showed the applicability of the method and the potential of crowdsourced Waze user reports in secondary crash identification.