Multilevel data and Bayesian analysis in traffic safety

Background

Traditional crash prediction models, such as generalized linear regression model, are incapable of taking into account multilevel data structure. Therefore they suffer from a common underlying limitation that each observation (e.g. a crash or a vehicle involvement) in the estimation procedure corresponds to an individual situation in which the residuals exhibit independence.

Problem

However, this “independence” assumption may often not hold true since multilevel data structures exist extensively because of the traffic data collection and clustering process. Disregarding the possible within-group correlations may lead to production of models with unreliable parameter estimates and statistical inferences.

Proposed theory

In this paper, a 5 × ST-level hierarchy is proposed to represent the general framework of multilevel data structures in traffic safety, i.e. [Geographic region level − Traffic site level − Traffic crash level − Driver-vehicle unit level − Occupant level] × Spatiotemporal level. The involvement and emphasis for different sub-groups of these levels depend on different research purposes and also rely on the heterogeneity examination on crash data employed. To properly accommodate the potential cross-group heterogeneity and spatiotemporal correlation due to the multilevel data structure, a Bayesian hierarchical approach that explicitly specifies multilevel structure and reliably yields parameter estimates is introduced and recommended.

Case studies

Using Bayesian hierarchical models, the results from several case studies are highlighted to show the improvements on model fitting and predictive performance over traditional models by appropriately accounting for the multilevel data structure.     

A method of human reliability analysis and quantification for space missions based on a Bayesian network and the cognitive reliability and error analysis method

Analyses of human reliability during manned spaceflight are crucial because human error can easily arise in the extreme environment of space and may pose a great potential risk to the mission. Although various approaches exist for human reliability analysis (HRA), all these approaches are based on human behavior on the ground. Thus, to appropriately analyze human reliability during spaceflight, this paper proposes a space‐based HRA method of quantifying the human error probability (HEP) for space missions. Instead of ground‐based performance shaping factors (PSFs), this study addresses PSFs specific to the space environment, and a corresponding evaluation system is integrated into the proposed approach to fully consider space mission characteristics. A Bayesian network is constructed based on the cognitive reliability and error analysis method (CREAM) to model these space‐based PSFs and their dependencies. By incorporating the Bayesian network, the proposed approach transforms the HEP estimation procedure into a probabilistic calculation, thereby overcoming the shortcomings of traditional HRA methods in addressing the uncertainty of the complex space environment. More importantly, by acquiring more information, the HEP estimates can be dynamically updated by means of this probabilistic calculation. By studying 2 examples and evaluating the HEPs for an International Space Station ingress procedure, the feasibility and superiority of the developed approach are validated both mathematically and in a practical scenario.     

Risk-based reconfiguration of safety monitoring system using dynamic Bayesian network

To prevent an abnormal event from leading to an accident, the role of its safety monitoring system is very important. The safety monitoring system detects symptoms of an abnormal event to mitigate its effect at its early stage. As the operation time passes by, the sensor reliability decreases, which implies that the decision criteria of the safety monitoring system should be modified depending on the sensor reliability as well as the system reliability. This paper presents a framework for the decision criteria (or diagnosis logic) of the safety monitoring system. The logic can be dynamically modified based on sensor output data monitored at regular intervals to minimize the expected loss caused by two types of safety monitoring system failure events: failed-dangerous (FD) and failed-safe (FS). The former corresponds to no response under an abnormal system condition, while the latter implies a spurious activation under a normal system condition. Dynamic Bayesian network theory can be applied to modeling the entire system behavior composed of the system and its safety monitoring system. Using the estimated state probabilities, the optimal decision criterion is given to obtain the optimal diagnosis logic. An illustrative example of a three-sensor system shows the merits and characteristics of the proposed method, where the reasonable interpretation of sensor data can be obtained.     

Analysis of traffic accident size for Korean highway using structural equation models

Accident size can be expressed as the number of involved vehicles, the number of damaged vehicles, the number of deaths and/or the number of injured. Accident size is the one of the important indices to measure the level of safety of transportation facilities. Factors such as road geometric condition, driver characteristic and vehicle type may be related to traffic accident size. However, all these factors interact in complicate ways so that the interrelationships among the variables are not easily identified. A structural equation model is adopted to capture the complex relationships among variables because the model can handle complex relationships among endogenous and exogenous variables simultaneously and furthermore it can include latent variables in the model. In this study, we use 2649 accident data occurred on highways in Korea and estimate relationship among exogenous factors and traffic accident size. The model suggests that road factors, driver factors and environment factors are strongly related to the accident size.     

Analysis of traffic accidents on rural highways using Latent Class Clustering and Bayesian Networks

One of the principal objectives of traffic accident analyses is to identify key factors that affect the severity of an accident. However, with the presence of heterogeneity in the raw data used, the analysis of traffic accidents becomes difficult. In this paper, Latent Class Cluster (LCC) is used as a preliminary tool for segmentation of 3229 accidents on rural highways in Granada (Spain) between 2005 and 2008. Next, Bayesian Networks (BNs) are used to identify the main factors involved in accident severity for both, the entire database (EDB) and the clusters previously obtained by LCC. The results of these cluster-based analyses are compared with the results of a full-data analysis. The results show that the combined use of both techniques is very interesting as it reveals further information that would not have been obtained without prior segmentation of the data. BN inference is used to obtain the variables that best identify accidents with killed or seriously injured. Accident type and sight distance have been identify in all the cases analysed; other variables such as time, occupant involved or age are identified in EDB and only in one cluster; whereas variables vehicles involved, number of injuries, atmospheric factors, pavement markings and pavement width are identified only in one cluster.     

A multinomial logit analysis of risk factors influencing road traffic injury severities in the Erzurum and Kars Provinces of Turkey

A retrospective cross-sectional study is conducted analysing 11,771 traffic accidents reported by the police between January 2008 and December 2013 which are classified into three injury severity categories: fatal, injury, and no injury. Based on this classification, a multinomial logit analysis is performed to determine the risk factors affecting the severity of traffic injuries. The estimation results reveal that the following factors increase the probability of fatal injuries: drivers over the age of 65; primary-educated drivers; single-vehicle accidents; accidents occurring on state routes, highways or provincial roads; and the presence of pedestrian crosswalks. The results also indicate that accidents involving cars or private vehicles or those occurring during the evening peak, under clear weather conditions, on local city streets or in the presence of traffic lights decrease the probability of fatal injuries. This study comprises the most comprehensive database ever created for a Turkish sample. This study is also the first attempt to use an unordered response model to determine risk factors influencing the severity of traffic injuries in Turkey.     

Quantification method analysis of the relationship between occupant injury and environmental factors in traffic accidents

Injury analysis following a vehicle crash is one of the most important research areas. However, most injury analyses have focused on one-dimensional injury variables, such as the AIS (Abbreviated Injury Scale) or the IIS (Injury Impairment Scale), at a time in relation to various traffic accident factors. However, these studies cannot reflect the various injury phenomena that appear simultaneously. In this paper, we apply quantification method II to the NASS (National Automotive Sampling System) CDS (Crashworthiness Data System) to find the relationship between the categorical injury phenomena, such as the injury scale, injury position, and injury type, and the various traffic accident condition factors, such as speed, collision direction, vehicle type, and seat position. Our empirical analysis indicated the importance of safety devices, such as restraint equipment and airbags. In addition, we found that narrow impact, ejection, air bag deployment, and higher speed are associated with more severe than minor injury to the thigh, ankle, and leg in terms of dislocation, abrasion, or laceration.     

Application of propensity scores and potential outcomes to estimate effectiveness of traffic safety countermeasures: Exploratory analysis using intersection lighting data

More than 5.5 million police-reported traffic crashes occurred in the United States in 2009, resulting in 33,808 fatalities and more than 2.2 million injuries. Significant funds are expended annually by federal, state, and local transportation agencies in an effort to reduce traffic crashes. Effective safety management involves selecting highway and street locations with potential for safety improvements; correctly diagnosing safety problems; identifying appropriate countermeasures; prioritizing countermeasure implementation at selected sites; and, evaluating the effectiveness of implemented countermeasures. Accurate estimation of countermeasure effectiveness is a critical component of the safety management process. In this study, a statistical modeling framework, based on propensity scores and potential outcomes, is described to estimate countermeasure effectiveness from non-randomized observational data. Average treatment effects are estimated using semi-parametric estimation methods. To demonstrate the framework, the average treatment effect of fixed roadway lighting at intersections in Minnesota is estimated. The results indicate that fixed roadway lighting reduces expected nighttime crashes by approximately 6%, which compares favorably to other, recent lighting-safety research findings.     

Investigation of road network features and safety performance

The analysis of road network designs can provide useful information to transportation planners as they seek to improve the safety of road networks. The objectives of this study were to compare and define the effective road network indices and to analyze the relationship between road network structure and traffic safety at the level of the Traffic Analysis Zone (TAZ). One problem in comparing different road networks is establishing criteria that can be used to scale networks in terms of their structures. Based on data from Orange and Hillsborough Counties in Florida, road network structural properties within TAZs were scaled using 3 indices: Closeness Centrality, Betweenness Centrality, and Meshedness Coefficient. The Meshedness Coefficient performed best in capturing the structural features of the road network. Bayesian Conditional Autoregressive (CAR) models were developed to assess the safety of various network configurations as measured by total crashes, crashes on state roads, and crashes on local roads. The models’ results showed that crash frequencies on local roads were closely related to factors within the TAZs (e.g., zonal network structure, TAZ population), while crash frequencies on state roads were closely related to the road and traffic features of state roads. For the safety effects of different networks, the Grid type was associated with the highest frequency of crashes, followed by the Mixed type, the Loops & Lollipops type, and the Sparse type. This study shows that it is possible to develop a quantitative scale for structural properties of a road network, and to use that scale to calculate the relationships between network structural properties and safety.     

Human reliability analysis for manned submersible diving process based on CREAM and Bayesian network

The diving mission of manned submersibles is a long‐term, high‐intensity work that is affected by many factors and is in a narrow confined space. In order to improve the reliability of oceanauts' safe operations, this paper is based on the cognitive reliability and error analysis method (CREAM) and the Bayesian network method to study the human errors of the diving mission. First, we construct a Bayesian network framework of the diving process by analyzing the diving steps. Second, the CREAM is applied to calculate the prior probability of each root node's error. Then, the backward reasoning ability of the Bayesian network is used to calculate the posterior probabilities and identify the top few risk nodes. Finally, we obtained the top few risk factors. Among them, we find that the light distribution design in the risk nodes is the more influential risk factor, so a brief design is made on them.     

Reliability assessment for failure-dependent and uncertain systems: A Bayesian network based on copula method and probability-box

Managing failure dependence of complex systems with hybrid uncertainty is one of the hot problems in reliability assessment. Epistemic uncertainty is attributed to complex working environment, system structure, human factors, imperfect knowledge, etc. Probability-box has powerful characteristics for uncertainty analysis and can be effectively adopted to represent epistemic uncertainty. However, arithmetic rules on probability-box structures are mostly used among structures representing independent random variables. In most practical engineering applications, failure dependence is always introduced in system reliability analysis. Therefore, this paper proposes a developed Bayesian network combining copula method with probability-box for system reliability assessment. There are four main steps involved in the reliability computation process: marginal distribution identification and estimation, copula function selection and parameter estimation, reliability analysis of components with correlations and Bayesian forward analysis. The benefits derived from the proposed approach are used to overcome the computational limitations of n-dimensional integral operation, and the advantages of useful properties of copula function in reliability analysis of systems with correlations are adopted. To demonstrate the effectiveness of the developed Bayesian network, the proposed method is applied to a real large piston compressor.     

Safety analysis in process facilities: Comparison of fault tree and Bayesian network approaches

Safety analysis in gas process facilities is necessary to prevent unwanted events that may cause catastrophic accidents. Accident scenario analysis with probability updating is the key to dynamic safety analysis. Although conventional failure assessment techniques such as fault tree (FT) have been used effectively for this purpose, they suffer severe limitations of static structure and uncertainty handling, which are of great significance in process safety analysis. Bayesian network (BN) is an alternative technique with ample potential for application in safety analysis. BNs have a strong similarity to FTs in many respects; however, the distinct advantages making them more suitable than FTs are their ability in explicitly representing the dependencies of events, updating probabilities, and coping with uncertainties. The objective of this paper is to demonstrate the application of BNs in safety analysis of process systems. The first part of the paper shows those modeling aspects that are common between FT and BN, giving preference to BN due to its ability to update probabilities. The second part is devoted to various modeling features of BN, helping to incorporate multi-state variables, dependent failures, functional uncertainty, and expert opinion which are frequently encountered in safety analysis, but cannot be considered by FT. The paper concludes that BN is a superior technique in safety analysis because of its flexible structure, allowing it to fit a wide variety of accident scenarios.     

An algorithm of discrete-time Bayesian network for reliability analysis of multilevel system with warm spare gate

Discrete-time Bayesian network (DTBN) is a popular tool for the reliability analysis of dynamic systems, which, however, is insufficient in analyzing the reliability of multilevel system (MLS) with warm spare (WSP) gates. Additionally, conventional DTBNs are not able to consider the situation that dormant components and primary components may fail during the same time interval. To this end, this paper analyzes the dynamic reliability characteristics of dormant systems with WSP gates by utilizing DTBNs. Moreover, an algorithm of modeling the conditional probability table (CPT) for WSP gates together with a new schedule of constructing dynamic Bayesian networks is put forward. The validation of the proposed techniques is implemented by Monte Carlo simulation (MCS) and reliability analysis of an actual communication station system.     

A method for risk-informed safety significance categorization using the analytic hierarchy process and bayesian belief networks

A risk-informed safety significance categorization (RISSC) is to categorize structures, systems, or components (SSCs) of a nuclear power plant (NPP) into two or more groups, according to their safety significance using both probabilistic and deterministic insights. In the conventional methods for the RISSC, the SSCs are quantitatively categorized according to their importance measures for the initial categorization. The final decisions (categorizations) of SSCs, however, are qualitatively made by an expert panel through discussions and adjustments of opinions by using the probabilistic insights compiled in the initial categorization process and combining the probabilistic insights with the deterministic insights. Therefore, owing to the qualitative and linear decision-making process, the conventional methods have the demerits as follows: (1) they are very costly in terms of time and labor, (2) it is not easy to reach the final decision, when the opinions of the experts are in conflict and (3) they have an overlapping process due to the linear paradigm (the categorization is performed twice—first, by the engineers who propose the method, and second, by the expert panel). In this work, a method for RISSC using the analytic hierarchy process (AHP) and bayesian belief networks (BBN) is proposed to overcome the demerits of the conventional methods and to effectively arrive at a final decision (or categorization). By using the AHP and BBN, the expert panel takes part in the early stage of the categorization (that is, the quantification process) and the safety significance based on both probabilistic and deterministic insights is quantified. According to that safety significance, SSCs are quantitatively categorized into three categories such as high safety significant category (Hi), potentially safety significant category (Po), or low safety significant category (Lo). The proposed method was applied to the components such as CC-V073, CV-V530, and SI-V644 in Ulchin Unit 3 NPP in South Korea. The expert panel consisted of two probabilistic safety assessments (PSA) experts and one system design expert. Before categorizing the components, the design basis functions, simplified P and IDs, and the Fussell-Vesely (FV) importance and the Risk Achievement Worth (RAW) of the PSA were prepared for the experts' evaluations. By using this method, we could categorize the components quantitatively on the basis of experts' knowledge and experience in an early stage.     

Prediction of hopper discharge rate using combined discrete element method and artificial neural network

An Artificial Neural Network (ANN) was developed to predict the mass discharge rate from conical hoppers. By employing Discrete Element Method (DEM), numerically simulated flow rate data from different internal angles (20°–80°) hoppers were used to train the model. Multi-component particle systems (binary and ternary) were simulated and mass discharge rate was estimated by varying different parameters such as hopper internal angle, bulk density, mean diameter, coefficient of friction (particle-particle and particle-wall) and coefficient of restitution (particle-particle and particle-wall). The training of ANN was accomplished by feed forward back propagation algorithm. For validation of ANN model, the authors carried out 22 experimental tests on different mixtures (having different mean diameter) of spherical glass beads from different angle conical hoppers (60° and 80°). It was found that mass discharge rate predicted by the developed neural network model is in a good agreement with the experimental discharge rate. Percentage error predicted by ANN model was less than ±13%. Furthermore, the developed ANN model was also compared with existing correlations and showed a good agreement.     

Failure mode and effects analysis by integrating Bayesian fuzzy assessment number and extended gray relational analysis‐technique for order preference by similarity to ideal solution method

Failure mode and effects analysis (FMEA) is a prospective risk assessment tool used to identify, assess, and eliminate potential failure modes (FMs) in various industries to improve security and reliability. However, the traditional FMEA method has been criticized for several shortcomings and even the improved FMEA methods based on predefined linguistic terms cannot meet the needs of FMEA team members' diversified opinion expressions. To solve these problems, a novel FMEA method is proposed by integrating Bayesian fuzzy assessment number (BFAN) and extended gray relational analysis‐technique for order preference by similarity to ideal solution (GRA‐TOPSIS) method. First, the BFANs are used to flexibly describe the risk evaluation results of the identified failure modes. Second, the Hausdorff distance between BFANs is calculated by using the probability density function (PDF). Finally, on the basis of the distance, the extended GRA‐TOPSIS method is applied to prioritize failure modes. A simulation study is presented to verify the effectiveness of the proposed approach in dealing with vague concepts and show its advantages over existing FMEA methods. Furthermore, a real case concerning the risk evaluation of aero‐engine turbine and compressor blades is provided to illustrate the practical application of the proposed method and particularly show the potential of using the BFANs in capturing FMEA team members' diverse opinions.     

Bayesian ranking of sites for engineering safety improvements: decision parameter, treatability concept, statistical criterion, and spatial dependence

In recent years, there has been a renewed interest in applying statistical ranking criteria to identify sites on a road network, which potentially present high traffic crash risks or are over-represented in certain type of crashes, for further engineering evaluation and safety improvement. This requires that good estimates of ranks of crash risks be obtained at individual intersections or road segments, or some analysis zones. The nature of this site ranking problem in roadway safety is related to two well-established statistical problems known as the small area (or domain) estimation problem and the disease mapping problem. The former arises in the context of providing estimates using sample survey data for a small geographical area or a small socio-demographic group in a large area, while the latter stems from estimating rare disease incidences for typically small geographical areas. The statistical problem is such that direct estimates of certain parameters associated with a site (or a group of sites) with adequate precision cannot be produced, due to a small available sample size, the rareness of the event of interest, and/or a small exposed population or sub-population in question. Model based approaches have offered several advantages to these estimation problems, including increased precision by "borrowing strengths" across the various sites based on available auxiliary variables, including their relative locations in space. Within the model based approach, generalized linear mixed models (GLMM) have played key roles in addressing these problems for many years. The objective of the study, on which this paper is based, was to explore some of the issues raised in recent roadway safety studies regarding ranking methodologies in light of the recent statistical development in space-time GLMM. First, general ranking approaches are reviewed, which include na?ve or raw crash-risk ranking, scan based ranking, and model based ranking. Through simulations, the limitation of using the na?ve approach in ranking is illustrated. Second, following the model based approach, the choice of decision parameters and consideration of treatability are discussed. Third, several statistical ranking criteria that have been used in biomedical, health, and other scientific studies are presented from a Bayesian perspective. Their applications in roadway safety are then demonstrated using two data sets: one for individual urban intersections and one for rural two-lane roads at the county level. As part of the demonstration, it is shown how multivariate spatial GLMM can be used to model traffic crashes of several injury severity types simultaneously and how the model can be used within a Bayesian framework to rank sites by crash cost per vehicle-mile traveled (instead of by crash frequency rate). Finally, the significant impact of spatial effects on the overall model goodness-of-fit and site ranking performances are discussed for the two data sets examined. The paper is concluded with a discussion on possible directions in which the study can be extended.     

Development of a simple and practical method for estimating the liquid absorption of pharmaceutical porous materials using a capillary rise technique

Porous materials are attractive substances for designing pharmaceutical particulates. However, understanding the behavior of liquid absorption into the intra-pores and interstices of porous carrier particles is important to effectively manufacture active pharmaceutical ingredients (APIs) using these carriers. In this study, we established a simple and practical method for evaluating the liquid absorption behavior of porous carriers using force tensiometry and a capillary rise technique. Different-sized tablets of porous materials were prepared and evaluated by this method using various solvents to estimate liquid absorption into the intra-particle pores and interstices of the particles. The amount of liquid trapped in the interstices of the particles decreased with decreasing tablet volume, after which the amount of liquid in the intra-particle pores could be estimated. Finally, API-loaded particles were prepared by absorbing the API solution into porous carriers based on the intra-capacity revealed above. No free API was found on the surface of the prepared particles, as it was well absorbed into the intra-particle pores. Collectively, this tensiometer method using different-sized tablets of porous materials appears to be a promising technique for evaluating the liquid absorption characteristics of porous pharmaceutical materials.     

Reliability analysis of chatter stability for milling process system with uncertainties based on neural network and fourth moment method

A reliability analysis of the milling system with uncertainties is developed in this paper to predict reliable chatter-free machining parameters. The chatter reliability refers to the probability of no chatter for the dynamic milling system. Then a reliability model is established to predict the chatter vibration, in which the dominant modal parameters of the dynamic milling system are defined as random variables. To solve the reliable model with the second-order fourth-moment (SOFM) method, the limiting axial cutting depth is substituted by an explicit expression obtained using a neural network. Therefore, after distributions of the random parameters are experimentally determined, the reliability of the given machining parameters can be computed with the SOFM method. Furthermore, a reliable stability lobe diagram (RSLD) can be plotted to obtain more reliable and accurate stable region instead of the conventional SLD. A case study is performed to validate the feasibility of the proposed method. The reliability of the milling system was calculated with the SOFM method, the first-order second-moment (FOSM) method and the Monte Carlo simulation (MCS) method. The results from the SOFM method and MCS method were found to be more consistent. Moreover, a RSLD with the reliability level 0.99 was compared with a conventional SLD plotted using the mean values of the random parameters. Chatter tests shown that the RSLD with the higher reliability level was more accurate for predicting the chatter-free machining parameters.