Maximum likelihood method of estimating the conflict-crash relationship.

The fundamental matters of how traffic conflicts are connected to crashes and how to estimate this connection with traffic conflict data is an active subject of research and refinements. There are still open questions about traffic events that can be analytically extrapolated to related crashes, and how to efficiently estimate the probability of crash associated with such events to enable conversion of observed events to the corresponding expected number of crashes. There are two important uses of a working estimation method: (1) rapid assessment of safety at specific roads locations and evaluation of countermeasures by safety engineers, (2) modeling of safety effects by analysts based on relatively short observations at multiple locations or at limited number of locations but during extended periods. This paper focuses on the application of traffic conflicts by safety engineers where the method practicality is important. The paper first recalls the OLS method of estimating the shape parameter of the underlying Lomax distribution proposed in (Tarko, 2018). Then, the ML method is introduced and the Lomax-based crash estimates obtained with the two methods are compared. Both the methods assume the scale parameter to estimate the shape parameter. The effect of assuming the scale parameter on estimates of the expected number of crashes is evaluated. To bring the scale parameter's effect into a meaningful perspective, it is compared to two other effects: (1) type of driver, and (2) limited number of observations. Finally, re-parametrized Lomax distribution is pointed out as a potential way to address the difficulties with estimating the two distribution parameters simultaneously. The summary of the results closes the paper.

Feasibility study of highway alignment design controls for autonomous vehicles.

In recent years, the development and testing of autonomous driving technology have become widespread around the world. However, due to differences in perception abilities between autonomous vehicles and human drivers, the current geometric design controls for highway alignments, designed for the human driver, may not be applicable to the autonomous vehicle (AV). Few studies, however, have systematically investigated the design controls for autonomous vehicles, though we face full driving automation in the next few decades. Because the range of modern AV sensors reaches 250 m, with expected further improvements in the near future, there is a need to determine how the sensors' perception field and perception-reaction time may affect the current road design standards developed for human drivers. This study therefore tested the feasibility of the current design controls for fully-autonomous vehicles by separately computing controls for vertical alignments and combined horizontal and vertical alignments, considering the AV's perception abilities of perception-reaction time (PRT), sensor height, and upward angle from the horizontal. The required stopping sight distance (SSD) and minimum length of sag and crest vertical curves were derived and compared with those for human drivers. Computations for combined alignments were based on Green Book coordination guidelines: as the minimum length of horizontal curve can be used for alignments adhering to guidelines, preview sight distance (PVSD) was computed for alignments that do not. Results showed that 1) AV-based design controls on vertical curves were more tolerant than those based on human drivers; and 2) the dominating criterion of sag vertical curve design control was comfort for autonomous vehicles, versus required SSD for human drivers.

A unifying view on traffic conflicts and their connection with crashes.

After decades of research on traffic conflicts and other crash surrogate events, defining these events and conclusively connecting them with crashes continue to be the most important tasks. This paper aims to help establish a consensus on these two fundamental matters by discussing the underlying concepts by which they can be connected in a consistent construct justified with theory and empirical evidence. The importance of insight into a safety-relevant event beyond what is externally observable is emphasized by considering two distributions of crash nearness: (1) values observed by external observers and (2) driver-preferred values that are usually unobservable. Traffic encounters and traffic conflicts are discussed here in the context of crash possibility illustrated with these two distributions. The difference between the preferred and observed crash nearness values is introduced as the delay of response to an error that violates the crash nearness preference. Traffic conflicts caused by driver errors that violate the driver populace's minimum crash nearness are recommended for safety analysis if only external observations are available. The conditions of properly detecting such traffic conflicts and estimating the probability of crash are identified and their validity is emphasized based on the past SHRP2 study. The mentioned study identified two additional conditions for proper identification of traffic conflicts: (1) speeds sufficiently high to induce driver responses consistent with the theory and with Lomax distribution and (2) elimination of self-clearing encounters such as a preceding vehicle exiting a lane in rear-end interactions. The most encouraging finding of this study is the mentioned sufficiently high speeds that tend to coincide with collision outcomes sufficiently serious to be reportable to the authorities. Another encouraging element is the insight about preferred crash nearness values that may be brought by autonomous vehicles. The biggest challenge in applying EV modeling today is using proper safety-relevant events to ensure that the tail of a distribution estimated based on observed events is consistent with the distribution tail that represents a crash. Autonomous vehicles may help eliminate this challenge since their preferred crash nearness values should be known.

Validity of failure-caused traffic conflicts as surrogates of rear-end collisions in naturalistic driving studies.

Although the frequency and severity of crashes are direct measures of road safety, crash data are typically of limited quality and they require long data collection periods to produce conclusive results. Surrogates of crashes that would allow a quick and accurate estimation of safety have been an active topic for years. Among multiple alternatives, traffic conflicts have been established as a promising surrogate measure. This paper is aimed to demonstrate the validity of failure-caused traffic conflicts by applying a recently proposed Lomax-based method to estimate the expected number of crashes from observed traffic conflicts. The data collected in the naturalistic driving program, the Second Strategic Highway Research Program (SHRP2), were used in the validation task. The rear-end crashes recorded during the SHRP2 program and the corresponding rear-end traffic conflicts were analyzed for three categories of drivers: young male, mature male, and mature female. Past research has indicated that these three categories have a distinctively different proneness to involvement in crashes. Out of all rear-end traffic conflicts included in the SHRP2 database, 1.4 % were used to estimate the crash frequencies and rates for each studied type of driver. The Lomax distribution was applied within the counterfactual framework. Then, the conflict-based crash rate estimates were compared to the crash rates of the studied types of drivers calculated from all the rear-end crashes observed in the SHRP2 study period. The conflict-based rate estimates followed well the crash-based rates and the existing knowledge about the safety performance of the studied drivers. The conflict-based results confirmed the over-representation of young male drivers in crashes. It was also confirmed that mature male drivers are involved in rear-end crashes more frequently than mature female drivers. The results demonstrate both the validity of the Lomax-based analysis of failure-caused traffic conflicts and the benefit of traffic conflicts analysis that considerably reduces a period of data collection from years for crashes to days or weeks for traffic conflicts.

Analyzing road near departures as failure-caused events.

Surrogate measures of safety attract revived interest thanks to the advancements in traffic observations techniques and the growing need for rapid safety evaluation. A new method of safety analysis based on failure-caused traffic conflicts and the Lomax distribution was recently proposed to estimate crash frequency more efficiently than with crash data. This paper has two objectives: (1) demonstrate the method applicability to near-departure data collected in a driving simulator, and (2) provide initial evidence of the method validity. Traffic failures and road users' delayed responses to these failures is considered as the primary cause of both conflicts and crashes. Unlike early postulated exceedance distributions the proposed Lomax distribution of response delays was derived from the causal mechanism. From this perspective, the proposed method may use the entire range of the underlying distribution as long as the observed conflicts are failure-related. The fundamentals of the method are briefly explained with the emphasis on certain behavior of crash frequency estimates implied by the proposed theory. Then, an example application of the method to analyze the risk of road departures in a driving simulator is presented. The results are then inspected and the trend in the estimates derived from the theory is confirmed. This finding points to the method validity. Additional applications of the method are expected to further increase the confidence towards the method and to encourage its introduction to the safety engineering practice.

Adequacy of negative binomial models for managing safety on rural local roads.

Count models, such as negative binomial regression, are well-established statistical methods for analyzing road safety. Although count models are widely used for arterial roads, their application to rural local roads is sparse, partly due to the concern of possible estimation bias caused by low crash counts. This paper revisits the matter to further evaluate the suitability of negative binomial models for rural local roads with low crash frequencies, comparing the performance of the model to probabilistic regression (ordered probit) proposed in the past. The negative binomial model was estimated to predict crashes for rural local intersections and compared to predictions obtained from the ordered probit model. Bivariate versions of both models were applied to improve model efficiency by incorporating correlation between two severity outcomes, fatal/injury (FI) and property damage only (PDO) crashes. The estimated models included several significant variables with intuitive signs. These results are discussed in the paper to support the claim that both models are adequate. Furthermore, the cumulative sums of the model-predicted and observed crashes conditioned on the estimated effects were compared to detect any systematic bias in the results. Although both models showed similar performance and no obvious biases could be detected, the negative binomial model seemed to behave slightly better than the ordered probit model, demonstrating the model's suitability in the analyzed case. The results point to the possibility of applying the Highway Safety Manual methodology to lower-volume county roads with focus shifted from individual high-crash locations to safety-deficient road features present at multiple locations.

Investigating factors of crash frequency with random effects and random parameters models: New insights from Chinese freeway study.

In response to the rapid economic growth in China, its freeway system has become the longest in the world and likely will continue to expand. Unfortunately, the safety issues on freeways in China have grown as well and are of great concern to Chinese transportation authorities and drivers. While many proven safety countermeasures developed and implemented by other countries are available for reference, they may be not fully transferrable to China due to the differences in driving cultures and conditions. As a result, an investigation of China's unique safety factors and effective relevant countermeasures are urgently needed. The study presented in this paper thoroughly investigated the factors contributing to freeway crashes in China based on detailed crash data, traffic characteristics, freeway geometry, pavement conditions, and weather conditions. To properly account for the over-dispersion of data and unobserved heterogeneity, a random effects negative binomial (RENB) model and a random parameters negative binomial (RPNB) model were applied, along with a negative binomial (NB) model. The analysis revealed a large number of crash frequency factors, including several interesting and important factors rarely studied in the past, such as the safety effects of climbing lanes. Moreover, the RENB and RPNB models were found to considerably outperform the NB model; however, although the RPNB exhibited better goodness-of-fit than the RENB model, the difference was rather small. The findings of this study shed more light on the factors influencing freeway crashes in China. The results will be useful to highway designers and engineers for creating, building, and operating safe freeways as well as to safety management departments for developing effective safety countermeasures. The study presented in this paper also provides additional guidance for choosing relevant methods to analyze safety and to identify safety factors.

Estimating the expected number of crashes with traffic conflicts and the Lomax Distribution - A theoretical and numerical exploration.

This paper justifies the Lomax distribution for counterfactual modeling of the probability of crash given a traffic conflict. The pre-crash process leading to a conflict or a crash as the result of a failure is discussed as this conceptualization is the basis for proposing a simple model of the probability of a crash at the moment when a conflict is still progressing. Then, a model applicable to heterogeneous conditions is derived; and the model's relevance, useful properties, and limitations are discussed. The published concepts and study results that support the derived model are provided in the paper. The existing Maximum Likelihood Estimate (MLE) method and the Probability-Weighted Moments (PWM) method of estimating the probability of crash and the expected number of crashes based on the proposed theory are presented. Then, a new Single Parameter Estimation (SPE) method is proposed and evaluated with extensive Monte Carlo experiments. The performance of the MLE, PWM, and SPE methods are compared. The SPE method is found more accurate and efficient than the other two methods. Unlike the benchmark methods, the proposed method produces real estimates in each case. The most important outcome of the presented study is confirmation that traffic conflicts claimed based on sufficiently small threshold separation (such as Time to Collision) allow unbiased estimation of the expected number of crashes during the conflicts observation period. A practical procedure of estimating safety is proposed that identifies the longest suitable threshold separation for each case based on the trends in the estimation results.

Analyzing crash frequency in freeway tunnels: A correlated random parameters approach.

The majority of past road safety studies focused on open road segments while only a few focused on tunnels. Moreover, the past tunnel studies produced some inconsistent results about the safety effects of the traffic patterns, the tunnel design, and the pavement conditions. The effects of these conditions therefore remain unknown, especially for freeway tunnels in China. The study presented in this paper investigated the safety effects of these various factors utilizing a four-year period (2009-2012) of data as well as three models: 1) a random effects negative binomial model (RENB), 2) an uncorrelated random parameters negative binomial model (URPNB), and 3) a correlated random parameters negative binomial model (CRPNB). Of these three, the results showed that the CRPNB model provided better goodness-of-fit and offered more insights into the factors that contribute to tunnel safety. The CRPNB was not only able to allocate the part of the otherwise unobserved heterogeneity to the individual model parameters but also was able to estimate the cross-correlations between these parameters. Furthermore, the study results showed that traffic volume, tunnel length, proportion of heavy trucks, curvature, and pavement rutting were associated with higher frequencies of traffic crashes, while the distance to the tunnel wall, distance to the adjacent tunnel, distress ratio, International Roughness Index (IRI), and friction coefficient were associated with lower crash frequencies. In addition, the effects of the heterogeneity of the proportion of heavy trucks, the curvature, the rutting depth, and the friction coefficient were identified and their inter-correlations were analyzed.

Barrier-relevant crash modification factors and average costs of crashes on arterial roads in Indiana.

The objective of this study was to develop crash modification factors (CMFs) and estimate the average crash costs applicable to a wide range of road-barrier scenarios that involved three types of road barriers (concrete barriers, W-beam guardrails, and high-tension cable barriers) to produce a suitable basis for comparing barrier-oriented design alternatives and road improvements. The intention was to perform the most comprehensive and in-depth analysis allowed by the cross-sectional method and the crash data available in Indiana. To accomplish this objective and to use the available data efficiently, the effects of barrier were estimated on the frequency of barrier-relevant (BR) crashes, the types of harmful events and their occurrence during a BR crash, and the severity of BR crash outcomes. The harmful events component added depth to the analysis by connecting the crash onset with its outcome. Further improvement of the analysis was accomplished by considering the crash outcome severity of all the individuals involved in a crash and not just drivers, utilizing hospital data, and pairing the observations with and without road barriers along same or similar road segments to better control the unobserved heterogeneity. This study confirmed that the total number of BR crashes tended to be higher where medians had installed barriers, mainly due to collisions with barriers and, in some cases, with other vehicles after redirecting vehicles back to traffic. These undesirable effects of barriers were surpassed by the positive results of reducing cross-median crashes, rollover events, and collisions with roadside hazards. The average cost of a crash (unit cost) was reduced by 50% with cable barriers installed in medians wider than 50ft. A similar effect was concluded for concrete barriers and guardrails installed in medians narrower than 50ft. The studied roadside guardrails also reduced the unit cost by 20%-30%. Median cable barriers were found to be the most effective among all the studied barriers due to the smaller increase in the crash frequency caused by these barriers and the less severe injury outcomes. More specifically, the occupants of vehicles colliding with near-side cable barriers tended to have less severe injuries than occupants of vehicles entering the median from median's farther side. The near-side cable barriers provided protection against rollover inside the median and against a potentially dangerous collision with or running over the median drain; therefore, the greatest safety benefit can be expected where cable barriers are installed at both edges of the median. The CMFs and unit crash costs for 48 road-barrier scenarios produced in this study are included in this paper.

An insight into the performance of road barriers - redistribution of barrier-relevant crashes.

Unlike most of traffic safety treatments that prevent crashes, road barriers reduce the severity of crash outcomes by replacing crashes with a high risk of severe injury and fatality (such as median crossover head-on collisions or collisions with high-hazard objects) with less risky events (such as collisions with barriers). This "crash conversion" is actually more complex than one-to-one replacement and it has not been studied yet. The published work estimated the reduction of selected types of crashes (typically, median crossover collisions) or the overall effect of barriers on crash severity. The objective of this study was to study the probabilities of various types of crash events possible under various road and barrier scenarios. The estimated probabilities are conditional given that at least one vehicle left the travelled way and the resulted crash had been recorded. The results are meant to deliver a useful insight onto the conversion of crashes by barriers from more to less risky to help better understand the mechanism of crash severity reduction. Such knowledge should allow engineers more accurate estimation of barriers' benefits and help researchers evaluate barriers' performance to improve the barrier's design. Seven barrier-relevant crash events possible after a vehicle departs the road could be identified based on the existing crash data and their probabilities estimated given the presence and location of three types of barriers: median concrete barriers, median and roadside W-beam steel guardrails, and high-tension median cable barriers. A multinomial logit model with variable outcomes was estimated based on 2049 barrier-relevant crashes occurred between 2003 and 2012 on 1258 unidirectional travelled ways in Indiana. The developed model allows calculating the changes in the probabilities of the barrier-relevant crash events. The results of this study indicated that road departures lead to less frequent crossings of unprotected (no barriers) medians 50-80ft. wide than for narrower medians 30-50ft wide. This benefit decreased with an increase in rollovers inside the median. Although our data indicated no median crossover events when a median barrier was present, the risk of crossovers, although low, is still present and could manifest itself if the sample were larger. The presence of barriers near a travelled way was associated with a higher risk of redirecting errant vehicles back to the roadway where they could collide with other vehicles continuing on the road. As expected, cable barriers installed on the far-side edge of a median were associated with a lower probability of being hit by errant vehicles and of redirecting vehicles into traffic than the nearside cable barriers. On the other hand, the probability of off-road non-barrier crashes was higher because vehicles penetrating the median from the unprotected side were exposed to median ditches and similar obstacles. The roadside guardrails were confirmed to reduce the percentage of hazardous off-road crashes. The results of this study facilitate a more transparent evaluation of the safety effect of road barriers.

Effectiveness of cable barriers, guardrails, and concrete barrier walls in reducing the risk of injury.

Roadway departure crashes tend to be severe, especially when the roadside exposes the occupants of errant vehicles to excessive injury hazards. As a cost-effective method when the clear zone width is insufficient, road barriers are often installed to prevent errant vehicles from colliding with dangerous obstacles or traversing steep slopes. This paper focuses on the safety performance of road barriers in Indiana in reducing the risk of injury. The objective of the study presented here is to compare the risk of injury among different hazardous events faced by an occupant in a single-vehicle crash. The studied hazardous events include rolling over, striking three types of barriers (guardrails, concrete barrier walls, and cable barriers) with different barrier offsets to the edge of the travelled way, and striking various roadside objects. A total of 2124 single-vehicle crashes (3257 occupants) that occurred between 2008 and 2012 on 517 pair-matched homogeneous barrier and non-barrier segments were analyzed. A binary logistic regression model with mixed effects was estimated for vehicle occupants. The segment pairing process and the use of random effects were able to handle the commonality within the same segment pair as well as the heterogeneity across segment pairs. The modeling results revealed that hitting a barrier is associated with lower risk of injury than a high-hazard event (hitting a pole, rollover, etc.). The odds of injury are reduced by 39% for median concrete barrier walls offset 15-18ft from the travelled way, reduced by 65% for a guardrail face offset 5-55ft, reduced by 85% for near-side median cable barriers (offset between 10ft and 29ft), and reduced by 78% with far-side median cable barriers (offset at least 30ft). Comparing different types of barriers is useful where some types of barriers can be used alternatively. This study found that the odds of injury are 43% lower when striking a guardrail instead of a median concrete barrier offset 15-18ft and 65% lower when striking a median concrete barrier offset 7-14ft. The odds of injury when striking a near-side median cable barrier is 57% lower than the odds for a guardrail face. This reduction for a far side median cable barrier is 37%. Thus, a guardrail should be preferred over a concrete wall and a cable barrier should be preferred over a guardrail where the road and traffic conditions allow. In the light of the results, installing median cable barriers on both sides of the median to reduce their lateral offset is beneficial for safety. The study also found that the unexplained heterogeneity across vehicles is much larger than it was across matched segment pairs.

Predicting crash likelihood and severity on freeways with real-time loop detector data.

Real-time crash risk prediction using traffic data collected from loop detector stations is useful in dynamic safety management systems aimed at improving traffic safety through application of proactive safety countermeasures. The major drawback of most of the existing studies is that they focus on the crash risk without consideration of crash severity. This paper presents an effort to develop a model that predicts the crash likelihood at different levels of severity with a particular focus on severe crashes. The crash data and traffic data used in this study were collected on the I-880 freeway in California, United States. This study considers three levels of crash severity: fatal/incapacitating injury crashes (KA), non-incapacitating/possible injury crashes (BC), and property-damage-only crashes (PDO). The sequential logit model was used to link the likelihood of crash occurrences at different severity levels to various traffic flow characteristics derived from detector data. The elasticity analysis was conducted to evaluate the effect of the traffic flow variables on the likelihood of crash and its severity.The results show that the traffic flow characteristics contributing to crash likelihood were quite different at different levels of severity. The PDO crashes were more likely to occur under congested traffic flow conditions with highly variable speed and frequent lane changes, while the KA and BC crashes were more likely to occur under less congested traffic flow conditions. High speed, coupled with a large speed difference between adjacent lanes under uncongested traffic conditions, was found to increase the likelihood of severe crashes (KA). This study applied the 20-fold cross-validation method to estimate the prediction performance of the developed models. The validation results show that the model's crash prediction performance at each severity level was satisfactory. The findings of this study can be used to predict the probabilities of crash at different severity levels, which is valuable knowledge in the pursuit of reducing the risk of severe crashes through the use of dynamic safety management systems on freeways.

Use of crash surrogates and exceedance statistics to estimate road safety.

The limited ability of existing safety models to properly reflect crash causality has its source in cross-sectional analysis applied to the estimation of the intrinsically complex safety factors with highly aggregated and frequently poor quality of data. The adequacy of the data may be improved thanks to the unprecedented progress in sensing technologies and the invention of the naturalistic driving method of data collection. Proposed in this paper is a new modeling paradigm that integrates several types of safety models. The primary improvement results from a more adequate representation of the crash occurrence process by incorporating crash precursor events into the modeling framework. A Pareto-based estimating method for the likelihood of a collision occurrence, given a precursor event, is explained and illustrated with the simple example of road departures.

Modeling drivers' speed selection as a trade-off behavior.

This paper proposes a new model of driver-preferred speeds derived from the assumption that drivers trade-off a portion of their safety for a time gain. The risk of receiving a ticket for speeding is also considered. A trip disutility concept is selected to combine the three components of speed choice (safety, time, and enforcement). The perceived crash risk and speed enforcement are considered as speed deterrents while the perceived value of a time gain is considered as a speed enticement. According to this concept, speeds that minimize the perceived trip disutility are preferred by drivers. The modeled trade-off behavior does not have to be fully rational since it is affected by drivers' preferences and their ability to perceive the risk. As such, the proposed framework follows the concept of bound rationality. The attractiveness of the model lies in its parameters being estimable with the observed preferred speeds and then interpretable as the factors of risk perception, the subjective value of time, and the perceived risk of speed enforcement. The proposed method may successfully supplement behavioral studies based on a driver survey. The study focuses on four-lane rural and suburban roads in Indiana, USA. The behavior of two types of drivers (trucks and cars) is modeled. The selection of test sites was such that the roads and other local characteristics varied across the studied sites while the population of drivers could be assumed as the same. The density of intersections, land development along the road, and the presence of sidewalks were the identified prominent risk perception factors. Another interesting finding is that the speed limit seems to encourage slow drivers to drive faster and fast drivers to drive slower.

Markov switching negative binomial models: an application to vehicle accident frequencies.

In this paper, two-state Markov switching models are proposed to study accident frequencies. These models assume that there are two unobserved states of roadway safety, and that roadway entities (roadway segments) can switch between these states over time. The states are distinct, in the sense that in the different states accident frequencies are generated by separate counting processes (by separate Poisson or negative binomial processes). To demonstrate the applicability of the approach presented herein, two-state Markov switching negative binomial models are estimated using five-year accident frequencies on Indiana interstate highway segments. Bayesian inference methods and Markov Chain Monte Carlo (MCMC) simulations are used for model estimation. The estimated Markov switching models result in a superior statistical fit relative to the standard (single-state) negative binomial model. It is found that the more frequent state is safer and it is correlated with better weather conditions. The less frequent state is found to be less safe and to be correlated with adverse weather conditions.

Tobit analysis of vehicle accident rates on interstate highways.

There has been an abundance of research that has used Poisson models and its variants (negative binomial and zero-inflated models) to improve our understanding of the factors that affect accident frequencies on roadway segments. This study explores the application of an alternate method, tobit regression, by viewing vehicle accident rates directly (instead of frequencies) as a continuous variable that is left-censored at zero. Using data from vehicle accidents on Indiana interstates, the estimation results show that many factors relating to pavement condition, roadway geometrics and traffic characteristics significantly affect vehicle accident rates.

The extreme value theory approach to safety estimation.

Crash-based safety analysis is hampered by several shortcomings, such as randomness and rarity of crash occurrences, lack of timeliness, and inconsistency in crash reporting. Safety analysis based on observable traffic characteristics more frequent than crashes is one promising alternative. In this research, we proposed a novel application of the extreme value theory to estimate safety. The method is considered proactive in that it no longer requires historical crash data for the model calibration. We evaluated the proposed method by applying it to right-angle collisions at signalized intersections. Evaluation results indicated a promising relationship between safety estimates and historical crash data. Crash estimates at seven out of twelve sites remained within the range of Poisson-based confidence intervals established using historical crash data. The test has yielded large-variance safety estimates due to the short 8-h observation period. A simulation experiment conducted in this study revealed that 3-6 weeks of observation are needed to obtain safety estimates with confidence intervals comparable to those being obtained from 4-year observed crash counts. The proposed method can be applied to other types of locations and collisions as well.

Perception-based road hazard identification with Internet support.

One of the most important tasks faced by highway agencies is identifying road hazards. Agencies use crash statistics to detect road intersections and segments where the frequency of crashes is excessive. With the crash-based method, a dangerous intersection or segment can be pointed out only after a sufficient number of crashes occur. A more proactive method is needed, and motorist complaints may be able to assist agencies in detecting road hazards before crashes occur. This paper investigates the quality of safety information reported by motorists and the effectiveness of hazard identification based on motorist reports, which were collected with an experimental Internet website. It demonstrates that the intersections pointed out by motorists tended to have more crashes than other intersections. The safety information collected through the website was comparable to 2-3 months of crash data. It was concluded that although the Internet-based method could not substitute for the traditional crash-based methods, its joint use with crash statistics might be useful in detecting new hazards where crash data had been collected for a short time.