Driver Response to Variable Message Signs-Based Traffic Information

Driver Response to Variable Message Signs-Based Traffic Information

Srinivas Peeta1 and Jorge L. Ramos2

1

Srinivas Peeta School of Civil Engineering Purdue University 550 Stadium Mall Drive West Lafayette, IN 47907-2051 Phone: (765) 494-2209 E-mail: [email protected]

2

Jorge L. Ramos, Jr. Vice-President of Operations Euro-American Steel Co., Inc. P.O. Box 950 Toa Baja, Puerto Rico 00951 Phone: (787)-784-2610 E-mail: [email protected]

Peeta and Ramos

1

ABSTRACT

This study investigates driver response attitudes to traffic information provided through variable message signs (VMS). It develops VMS driver response models using stated preference data collected through three different survey administration methods: an onsite survey, a mail-back survey, and an Internet-based survey. In the process, it highlights the strengths and limitations of each method in eliciting driver response attitudes to information provision. The use of different media for the survey administration provides insights for the design of travel surveys. A key study focus is to evaluate the effectiveness of Internet-based surveys for analyzing driver behavior under information provision. The results illustrate that a combination of survey administration methods may generate more representative data. They also indicate a high correlation between VMS message type and driver response. This suggests message content as a control variable for traffic system operators to trigger optimal routing policies under congested conditions to improve network performance. The paper highlights the benefits afforded by Internet-based surveys in the study context. They are cost-effective, amenable to automation, less laborintensive, can target certain market segments more effectively, and can enable greater clarity in the survey through better visual articulation. However, their widespread use requires greater market penetration in terms of Internet access.

Key Words: variable message signs, driver behavior, survey methods, Internet-based surveys.

Peeta and Ramos

2

1. INTRODUCTION

There is an increasing focus in recent years on the use of Advanced Traveler Information Systems (ATIS) to mitigate traffic congestion by enabling motorists to use the existing transportation systems more efficiently. ATIS combines the power of communications and information technologies to disseminate real-time traffic information so as to enable travelers to make more informed decisions regarding departure time, route choice and congestion avoidance. Variable Message Signs (VMS) are a cost-effective ATIS technology commonly used by traffic controllers to provide motorists with information on changing traffic conditions en-route. While VMS have been widely used for disseminating descriptive information on the traffic conditions ahead, work zones, and HOV lane access control, the extent of their capabilities for improving network performance under incident/congested situations has not been exploited sufficiently. Recent studies [1-2] have shown that displaying route guidance information on the relevant VMS has the potential to improve system performance by influencing driver route diversion decisions. However, the task of improving network performance using VMS is not trivial and involves two aspects. First, the information provided needs to be accurate and reliable. Unreliable information can reduce the credibility of VMS information amongst drivers, which could eventually degrade the system performance. Second, the messages need to be consistent with driver diversion behavior in the traffic region under consideration. The success of VMS-based control strategies depends largely on driver response to the conveyed information. Hence, it is essential to understand the decision-making process of drivers under real-time information provision.

Peeta and Ramos

3

There are two well-known methodologies for eliciting relationships between the supplied traffic information and the driver response: Revealed Preference (RP) and Stated Preference (SP). The RP approach analyzes driver behavior under real situations and is usually based upon driver reports (diaries of actual trips) and observations of actual driver behavior through field studies. While such studies have been conducted in the past [3-4], a common drawback is that their results are limited to the messages conveyed during the study period. The SP approach analyzes driver behavior by presenting individuals with a series of hypothetical travel alternatives. It has been commonly used to analyze transportation mode choice, and has increasingly become a common means of analyzing driver route choice behavior under real-time traffic information. Driver surveys are the main sources of SP data. In the VMS arena, numerous experiments [5-7] based on travel surveys in the form of questionnaire have been conducted to study driver compliance with information. There are several ways to administer these surveys. Until recently, travel surveys have relied extensively on the use of mail and on-site interviews. The increasing need for more detailed travel data warrants the exploration of more efficient survey administration methods. The spawning of the Internet, coupled with recent advances in information technology, offers potential mechanisms to enhance the effectiveness, efficiency and accuracy of surveys while enabling them to be more economical and less labor-intensive. This paper develops and compares VMS driver response models for the I-80/94 corridor in northwestern Indiana using SP data collected through three different administration methods: an on-site survey, a mail-back survey, and an Internet-based survey. In the process, it highlights the strengths and limitations of each method. The use

Peeta and Ramos

4

of different media for the survey administration provides insights for the design of travel surveys. The study illustrates that a combination of survey administration methods may generate more representative data for the problem being addressed. A key objective of the paper is to evaluate the effectiveness of Internet-based surveys for analyzing driver attitudes to information provision.

2. LITERATURE REVIEW

The most common methods to administer travel surveys are residence-based telephone, mail-back, and on-site surveys. Telephone interviews have become widespread for travel surveys because of the low cost of telephone services, combined with the increased efficiency afforded by modern computer-assisted telephone interviewing systems (CATIS). A telephone survey of households in northwest Indiana [8] sought to identify the factors that explain driver route diversion behavior to assist in the design of an ATIS for that region. The study results suggest that driver familiarity with the network and confidence in the information provided are key factors influencing motorists’ route diversion decisions.

Another study in Warrington, England [6] used a mail-back

questionnaire survey to explore commuter response to a wide variety of VMS messages. VMS pictures depicting different messages on traffic conditions were presented and the commuters were asked if they would divert to one of four available alternate routes. Multinomial logit models were estimated using the survey data to analyze how route diversion decisions are influenced by the different levels of delay and the incident type. They indicated that the impact of VMS-based information depends upon local circumstances and drivers’ socioeconomic characteristics. A common limitation of the

Peeta and Ramos

5

above studies is that only residents in the associated regions were surveyed, thereby excluding other relevant market segments such as non-resident occasional drivers and drivers who just pass through those areas. Hence, the sample may not be representative of the actual network users, especially if these segments constitute significant fractions. A well-known limitation of the SP approach is that SP responses may not satisfactorily reflect actual behavior. Recent developments in communications and information technology provide the potential to enhance the effectiveness and accuracy of SP surveys. In this context, two interactive survey approaches are Computer-Assisted Self Interviewing (CASI) and Internet administration. CASI has been proven to improve accuracy by making use of graphical user interfaces (GUI), images, and other related multimedia elements which holds the respondent’s interest. The increasing popularity of the Internet provides new technological opportunities and suggests its potential use as a survey administration method. Internet-based surveys are a viable option for reaching selected segments of the population, and a valuable option for multi-method surveys. They offer some advantages vis-à-vis travel surveys. First, graphically rich elements as in CASI surveys provide the ability to better articulate survey aspects, enabling the respondents to better understand the survey. Second, Internet-based surveys appear to be preferred by well-educated persons who usually have high refusal rates in mail-back surveys. Third, they can be highly cost-effective. Fourth, they can be automated, both for survey administration and post-survey processing. This reduces labor requirements substantially and simplifies the data processing tasks for the analyst. A current limitation of this method is the requirement that participants have access to the Internet. Since not everyone willing to participate in an Internet-based survey has such access, some bias is

Peeta and Ramos

6

likely in the sample. However, this bias is likely to decrease in the future with the continuing rapid penetration of Internet access into households, and the overall increasing popularity of Internet usage. A recent study [9] used an Internet-based survey to analyze trip-maker responses to real-time traffic information for shopping trips. The participation of faculty/staff members from the University of Texas at Austin and Austin residents was solicited through e-mail. A total of 3,490 requests for participation were sent out and 199 participated in the survey, which represents an overall response rate of 6%. In general, young males with high education levels were over-represented in the sample. This bias is due to the requirement that participants have access to the Internet. Modeling driver diversion behavior in the presence of traffic information is not trivial. In the context of VMS-based information, a compounding factor is the need for eliciting relationships between message content and driver diversion response under the multiple actual scenarios possible (e.g., varying content of messages, situational factors). This paper focuses on building VMS driver response models using SP data collected through a multi-pronged survey approach. It also seeks insights on the administration mechanisms of travel surveys vis-à-vis VMS, including those characteristics of Internetbased surveys that enhance the survey effectiveness.

3. STATED PREFERENCE SURVEY DESIGN

After identifying the measurable factors that potentially influence route diversion decision under VMS, a SP survey in the form of a questionnaire was developed. The first part addresses questions on socioeconomic characteristics such as gender, age, level of education, and household size. The second part elicits driver propensity to divert under

Peeta and Ramos

7

specific scenarios. In the last part of the questionnaire, the respondents were presented with eight generic descriptions of VMS message contents with varying amounts of information, and asked whether they would divert to an alternate route. The responses were recorded on a five point Likert scale (1-5), where 1 implies low willingness to divert and 5 implies high willingness to divert. Respondents answering 4 or 5 were assumed to divert under that VMS message. The objective of this part was to determine how different levels of information content provided by a VMS influence motorists’ inclination to divert. The generic messages used are discussed in the survey results section of the paper. Three survey administration methods were adopted in this study: (1) on-site survey, (2) mail-back survey, and (3) Internet-based survey. The multi-method approach was adopted to more effectively sample all segments of the target population. There are advantages and disadvantages of using one method over the other. The on-site method allows the researchers to personally interview the respondents, and hence the responses tend to be more reliable. However, the on-site and mail-back surveys are time-consuming as data needs to be manually archived in the computer for analysis.

3.1 Internet-based Survey

Internet-based surveys have become more widespread in recent years, providing substantial experience on their design and administration. Potentially, they allow broad access of participants to the survey since it only requires Internet access. The recruitment process entails sending a participation request to randomly selected subjects using e-mail or regular mail. The respondents’ choices are automatically recorded in a database

Peeta and Ramos

8

residing in the server containing the survey, and can be retrieved at a later time for data analysis. Hence, it is less time-consuming and less labor-intensive than the other two approaches. Potential pitfalls of this approach include the need for high levels of accessibility of Internet among potential respondents, and overcoming the inertia of having to access a specific Internet address. In addition, security issues are a generic concern with Internet-based applications. Access to the questionnaire in this study was provided through a respondentunique secure point of entry. It is critical to ensure that each respondent has the opportunity to complete only one questionnaire, and that those who are not recruited do not have access to the survey. This is done by providing a password to the targeted respondents along with the participation request. To ensure that respondents complete only one questionnaire, an HTTP cookie is sent by the server to store on the respondent web browser so it can later be read back from that browser. Cookies are a general mechanism which server side connections (such as CGI scripts) can use to both store and retrieve information on the client side of the connection. Therefore, if a respondent attempts to access the survey after submitting his/her responses, the server automatically identifies the cookie and informs the respondent that his/her participation has already been recorded. A Java script is used to ensure that the participant answers all questions before submitting the questionnaire.

4. SURVEY RESULTS

The data collection effort targeted travelers on the Borman Expressway (I-80/94) in northwestern Indiana where an Advanced Traffic Management System (ATMS) has been

Peeta and Ramos

9

deployed for incident management in the region. Among the information dissemination technologies being deployed there are highway advisory radio (HAR) and VMS. Both of them aim to disseminate traffic information to travelers to enable them to make more informed route switching decisions en-route when network conditions deteriorate. The on-site survey targets commuters as well as infrequent Borman travelers. Since truck traffic on the Borman Expressway comprises 30% of the total volume during the day and about 70% at night, truck drivers were also targeted through the on-site survey. The mail-back survey helps in sampling travelers that use the Borman on a daily basis (commuters). The Internet-based survey targets businesses and individuals who value time highly and who as a group entail high refusal rates in mail-back surveys [9]. The survey results from the different administration methods adopted in this study are presented hereafter.

4.1 On-site Survey Results

The on-site survey was conducted at two locations: a truck stop on the Borman Expressway and a rest area on I-65 few miles south of the Borman. The Borman Expressway represents part of the journey for most travelers who stop at the latter location. The survey conducted 248 interviews. The socioeconomic characteristics of the surveyed sample are summarized in Table 1. Of the 248 respondents, 79% were male. The distribution of motorists of different age groups within the sample was quite even except for the age groups of less than 20 and greater than 65. 59% of the interviewees had at least some college experience while 41% received at least one college degree. 61% had a household with 3 or more

Peeta and Ramos

10

members including themselves. The question on the household size was included in this survey because studies have shown that transportation-related decisions are a function of household size [8, 10-12]. Table 2 shows the diversion-related attitudes of the respondents. About half of them stated that they were regular drivers in the Borman Expressway region. However, this does not necessarily imply that such drivers are familiar with alternate routes other than their regular one. Hence, regular drivers were asked to state their familiarity level with alternate routes. Among this group, 65% were familiar with at least one alternate route besides the Borman Expressway. 70% stated that they would divert to an alternate route to avoid unexpected congestion under adverse weather conditions if a VMS message suggested it. This could be due to the effect of incident clearance time as bad weather conditions may hamper incident clearance operations, persuading drivers to avoid potential excessive delays by diverting to an alternate route. Also, 65% of the respondents stated that they would divert to an alternate route at night. These results are consistent with the results of previous studies (8) in this region. While the survey obtains responses on weather and time-of-day variables, these responses are not based on the consideration of other relevant factors, such as incident severity, that make driver responses to these variables more meaningful. Such a capability entails providing the respondents a large number of specific situations involving many factors through SP to elicit their response attitudes. Since portable VMS have been in place on the Borman Expressway for a few years, drivers’ trustworthiness on the information provided was sought in the survey. While 39% of the drivers stated that they would divert to an alternate route even if they believe that it would be longer than their current route, 29%

Peeta and Ramos

11

stated that they would not. 32% stated that they were undecided. 71% stated that they would divert to an alternate route under a work-related trip if that alternate route offered travel time savings ranging from 5 to 30 minutes. However, only 47% stated that they would divert on a personal trip for identical time savings. This reaffirms the notion of the higher value of time for work-related trips. Table 3 summarizes the driver willingness to divert to an alternate route when different types of VMS messages are displayed. The messages were specified in a random order in the questionnaire to avoid potential directional bias due to increasing information content. As discussed earlier, on the Likert scale used, ‘5’ represents a strong willingness to divert and ‘1’ indicates a strong unwillingness to divert. The results suggest no significant difference in diversion response to messages 1 and 2. That is, qualitative VMS information such as Occurrence of Accident and Location of the Accident have similar effect on driver propensity to divert. However, quantitative data and active messages conveying information on Expected Delay and/or Best Detour Strategy are considered valuable vis-à-vis influencing driver route diversion decisions. A shortcoming of using generic VMS messages in the survey is illustrated by the perceived relative importance of expected delay and location. While expected delays are perceivable in terms of magnitude, the value of location is perceivable only in actual situations or specifically constructed SP scenarios. In reality, the incident location can represent valuable information under many real situations.

4.2 Mail-back Survey Results

Peeta and Ramos

12

The mail-back survey questionnaire was sent to 3,600 randomly selected households in the Borman Expressway region. A total of 402 residents responded, which represents a response rate of about 11%. Table 1 reports the socioeconomic characteristics of the sample. Among the respondents, 59% were male. Compared to the on-site survey sample, there is a greater representation of women in the mail-back survey sample. The distribution of the respondents in terms of age group is skewed towards the older age groups; 73% were older than 40 years. This is different from the on-site survey where the distribution is more uniform. In terms of the level of education, there is no statistical difference between the on-site survey and the mail-back survey samples. Table 2 illustrates the diversion attitudes of the mail-back survey sample respondents. During the design of the on-site survey, it was assumed that Borman users were not familiar with VMS. However, several on-site survey respondents stated that they were familiar with VMS. Therefore, the mail-back questionnaire survey, which was conducted later, included a question on driver familiarity with VMS. It indicated that 84% of the respondents have some experience with VMS. This high percentage is because most respondents are daily commuters in the Borman Expressway region and, hence, are familiar with the portable VMS that has been in place on the Borman for a few years. However, the portable VMS are currently used in a rudimentary manner only, to inform users on roadwork and traffic conditions at the simplest information level. Almost 81% of the respondents stated that they were regular drivers on the Borman Expressway. This percentage is higher compared to that of the on-site survey because the mail-back survey sample is mostly composed of Borman commuters. Among regular

Peeta and Ramos

13

drivers, 90% indicated that they were familiar with alternate routes besides the Borman Expressway. Although the survey attempts to capture the effect of the magnitude of delay on the propensity to divert, the ratio of this delay to the expected trip time may be a more robust explanatory variable. Therefore, a question on the respondents’ average commute time was included in the mail-back questionnaire. 50% experience an average commute time less than 30 minutes, 26% experience between 30 to 60 minutes, 10% experience more than 60 minutes, and 14% are either unemployed or retired. Akin to the on-site survey, most respondents stated that they would divert to an alternate route to avoid unexpected traffic delay under adverse weather conditions or during the night. When asked if they would divert to an alternate route if a VMS suggested it, even if they believe that it would be longer than their current route, 39% stated that they would divert, 30% were undecided, and 31% stated that they would not divert. The distribution of the responses to this question is almost identical to that under the on-site survey. 84% of the participants stated that they would divert to an alternate route under a work-related trip if the alternate route offered time savings ranging from 5 to 30 minutes, while 75% stated that they would divert under a personal trip for identical time savings. The difference (9%) between drivers diverting under work-related trips and personal trips is lesser than the corresponding difference (24%) under the on-site survey. This is consistent with the responses to the question on the amount of delay that would convince drivers to divert. In the on-site survey, 53% stated that they would divert to an alternate route when the delay ranges from 5 to 30 minutes, while in the mail-back survey 82% would divert for the same range of delay. This can be partly attributed to the fact that Borman area residents, who constitute the mail-back survey sample, identify better with the ambient traffic

Peeta and Ramos

14

conditions in the region and are more familiar with its traffic network compared to nonregular drivers. Table 3 summarizes the willingness to divert to an alternate route when different VMS message contents are displayed. The results are consistent with those of the on-site survey, that is, driver propensity to divert increases as the level of detail in the information increases. The results suggest no significant difference in the diversion response to VMS message types 3 and 4, and between 6 and 7. The Best Detour Strategy and the Location of Accident have added value only in conjunction with information on Expected Delay and Best Detour Strategy, respectively.

4.3 Internet-based Survey Results

The Internet-based survey participants were recruited through e-mail. 880 e-mail addresses of residents in the Borman Expressway region and 125 e-mail addresses of businesses in the area were used to target potential survey participants. The sample consists of 29 residents and 5 businesses. This represents a response rate of 4.4% and 4.7% for the residents and businesses, respectively. The survey was implemented on the URL http://www.ecn.purdue.edu/Action/Survey/index.htm. The socioeconomic characteristics of the sample are illustrated in Table 1. About 74% of the respondents were male. This distribution compares quite well with that of the on-site survey. Most participants (82%) are less than 40 years old. In addition, the majority of the respondents (79%) received at least one college degree, and the rest (21%) had some college experience. Hence, the sample is biased towards young and well-educated persons.

Peeta and Ramos

15

The sample diversion characteristics are summarized in Table 3. Since the survey targeted residents and businesses in the Borman Expressway region, the results are similar to those obtained in the mail-back survey. Almost all respondents (97%) are regular drivers in the Borman Expressway region and are familiar with at least one alternate route besides their regular one. A notable difference between the Internet-based survey and the other two surveys is in the context of the driver’s trust in the traffic information provided. In the Internet-based survey, 47% of the respondents stated that they would divert to an alternate route even if they believed that it would be longer than their current route. Since the majority of the participants in the Internet-based survey are well-educated individuals, they are likely to be more at ease with technological innovations early on, and hence may exhibit lesser inertia and a greater level of compliance with VMS-based information. Table 3 summarizes the willingness to divert to an alternate route under different VMS message contents. The results are consistent with those obtained in the other two surveys.

5. MODELING DRIVER BEHAVIOR

5.1 Methodology In this section, VMS route diversion prediction models are developed using the SP data collected in the Borman region. Because the choice set of each respondent consists of only two alternatives, divert (i) or not divert (j), the modeling approach taken in this paper is the estimation of binary logit models. It expresses the probability (P) that an

Peeta and Ramos

16

individual n chooses alternative i (divert) as a function of the difference of the systematic components of the utilities (V) of the two alternatives in the choice set:

Pn (i ) =

1 1+ e

(

− Vin −V jn

)

(1)

The difference in the utility functions is related to relevant variables representing drivers’ socioeconomic characteristics (X), travel situations (Y) and the VMS message types provided (VMS):

V = f (αX , βY , γVMS )

(2)

A comprehensive list of the explanatory variables included in the utility function is presented in Table 4. For constructing the dependent variable, the Likert scale used in the survey is mapped to a binary choice by assuming that respondents answering 4 or 5 divert [5]. This is based on preliminary analyses that focus on the explanatory power of the various combinations of the Likert scale responses.

5.2 Model Estimation Results

To investigate the effect of different types of VMS messages on driver route diversion behavior, binary logit models were estimated using the aforementioned methodology. The data used for estimating the models came from the three SP surveys conducted. The

Peeta and Ramos

17

Maximum Likelihood Estimation (MLE) procedure was used to estimate the parameters of the models.

5.2.1 Model Estimation Using the On-site Survey Data

The model is illustrated in Table 5 and shows that motorists’ socioeconomic characteristics and perceptions are important factors influencing their decision to switch routes under VMS-based traffic information. The overall observations were used to estimate the model, which gave a total of 1,984 pooled observations (i.e., 248 respondents, each making diversion decisions under 8 different VMS messages). The negative sign of the alternative specific constant (ONE) implies that there is an inherent resistance to diverting from the current route. In the context of VMS, it suggests that users exhibit a tendency to remain on their current route when presented with little or no information on the incident ahead. This illustrates the potential of VMS-based traffic information to influence route diversion decisions. Gender and age also had a significant effect on route switching behavior. The positive sign of variable SEX indicates that males are more likely to divert to an alternate route under an incident situation. The negative sign of variable AGE suggests that younger drivers are more inclined to divert than older drivers when all other conditions are identical. These results are intuitive because males and younger drivers are more risk-willing in trying to minimize their travel time. Being a regular driver on the Borman Expressway is another significant factor affecting route diversion decisions. DRIV has a positive sign implying that regular drivers in the study region are more likely to divert. The model also suggests that the education level (EDU) of a driver may be an important factor influencing his/her diversion decisions. Well-

Peeta and Ramos

18

educated individuals exhibit greater compliance with VMS than their less-educated counterparts as reflected by the positive sign of the variable EDU. Education is a wellknown proxy for the income level of a person [13]. Therefore, well-educated people are likely to value time higher and may be more sensitive to delays on their planned route. The reliability of traffic information is another significant factor. If the traffic information is perceived to be less reliable, it is less likely to influence driver decisions as indicated by the positive sign of variable TRUST. The effect of VMS messages is illustrated through the increasing coefficient values corresponding to the VMS variables 2 to 8 (VMS1 is the base case). As stated previously, VMS messages 1 to 8 are in the order of increasing amount of information. Route diversion rates of drivers increase with the amount of information provided by the VMS, especially when estimated delay and alternative route information are provided. This can be inferred from the estimated coefficients of the VMSk variables, which increase from VMS2 through VMS8. All of these variables are very significant and provide the largest increase in log-likelihood among all variables in the model. Though VMS2 is not statistically significant, it was included in the final model because the incident location plays an important role in the route switching decision in real-world situations. This represents a limitation of the SP approach in capturing the situational behavior of users. The results show that traffic controllers could use message content as a control variable to influence network traffic conditions positively without compromising the integrity of information.

Peeta and Ramos

19

5.2.2 Model Estimation Using the Mail-back Survey Data

The logit model estimated using the mail-back survey data is presented in Table 5. All variables were included in the estimation procedure. Variables determined as being insignificant in intermediate models were omitted at the corresponding stages. The variable DRIV has a positive sign implying that regular drivers in the Borman Expressway region are more likely to divert. This is consistent with the survey since 84% of the survey participants use the Borman Expressway regularly. Since the survey targeted residents and businesses in the Borman Expressway region, respondents are familiar with at least one alternate route besides the Borman. This is reflected in the high explanatory power of the variable FAM. The model also suggests that individuals whose average work commute time is more than 30 minutes are more likely to divert under unexpected traffic congestion. This is because drivers with longer travel times generally have more perceptible travel time savings through diversion. Also, they may have greater opportunities to switch routes, especially if they have longer travel distances. As before, the variable TRUST is an important explanatory variable. Delay thresholds were investigated in this study by incorporating different expected delay ranges in the survey. They were represented by a dummy variable (DELAY) at the 10-minute threshold which was the only statistically significant one. The model indicates that drivers are more likely to divert if the expected delay is at least 10 minutes. This is important in the context of designing VMS information strategies, implying that it would be better to incorporate a delay threshold and display diversion advice only if some threshold delay is exceeded.

Peeta and Ramos

20

The trends in the VMS variables are similar to those in the on-site survey based model. There is a small reduction in the coefficient values from VMS3 to VMS4, but this difference is not statistically significant as determined by a paired-t-test. This suggests that explicit information on delays is valuable to drivers.

5.2.3 Model Estimation Using the Internet Survey Data The logit model for driver response under VMS using the Internet-based survey data is presented in Table 5. It should be noted here that the number of sample observations for this survey is small compared to those for the other two surveys. In the context of the socioeconomic variables, the results are consistent with those obtained in the previous models. A major difference is that TRUST is not a significant variable and hence was not included in the model. The low statistical significance for this variable might be due to the high level of education of the survey respondents (79% received at least one college degree). Well-educated individuals are likely to be more at ease with technological innovations and hence may exhibit greater compliance with VMS-based information. The trends in the VMSk variables are similar to those observed in the on-site and mail-back sample models.

5.2.4 Model Estimation Using the Combined Survey Data

As stated earlier, the multiple administration approach allows sampling the various segments of the target population more effectively. Therefore, a model combining the data collected from the on-site, mail-back, and Internet-based surveys was estimated. It is

Peeta and Ramos

21

illustrated in Table 5. The results lead to similar conclusions on VMS message content as in the previous models.

6. CONCLUDING COMMENTS

This study develops driver behavior models for on-line operations under VMS-based information provision. It provides insights on driver diversion attitudes under different VMS message contents, thereby providing guidelines for the design of VMS-based information supply strategies to enhance system performance under congested conditions. The strong correlation between the VMS message type and driver response suggests message content as an important control variable for improving system performance without compromising the integrity of the information provided. This led to a framework for optimizing system performance under incidents using the VMS message content [14]. It aims to provide the traffic controller the time-dependent information to be displayed on the relevant VMS. The models developed in this paper are essential inputs to such frameworks, so as to assure that the displayed messages factor in the driver response behavior, thereby enhancing system performance. A primary focus of the study was to analyze the strengths and limitations of various SP survey administration methods vis-à-vis effective survey design for real-time VMS-based information provision. A key objective was to obtain insights on the benefits afforded by Internet-based surveys in this context. Three different methods were adopted for collecting the data used to develop the behavioral models: (i) on-site survey, (ii) mailback survey, and (iii) Internet-based survey. The multiple administration method was adopted to more effectively sample the target population. The on-site survey targeted

Peeta and Ramos

22

commuters, infrequent Borman travelers, and truck drivers on the Borman Expressway. The mail-back and Internet-based surveys were used to target households and businesses in the Borman region. Hence, the combined survey sample is representative of the different groups of individuals using the Borman Expressway. Although the Internetbased survey offered some advantages, the response rate was relatively low. This is because, currently, the market penetration of Internet access is not widespread. However, as Internet usage becomes more commonplace, the multiple advantages of Internet-based surveys such as greater target audience access, greater data automation, and lesser costs, will increase their attractiveness vis-à-vis driver behavior sampling. Since the on-line VMS route advisory and guidance framework [14] can be ported to any traffic network, Internet-based surveys represent a cost-effective approach for driver behavior sampling in other traffic regions without significant additional effort.

ACKNOWLEDGEMENTS

This research was based on funding from the National Science Foundation under Grant No. CMS-9702612.

REFERENCES 1. Mammar, S., A. Messmer, P. Jensen, M. Papageorgiou, H. Haj-Salem, and L. Jensen. Automatic Control of Variable Message Signs in Aalborg. Transportation Research, Vol. 4C, No. 3, 1996, pp. 131-150. 2. Yim, Y. and J. L. Ygnace. Link Flow Evaluation Using Loop Detector Data: Traveler Response to Variable Message Signs. In Transportation Research Record 1550, Washington, D.C., 1996, pp. 58-64.

Peeta and Ramos

23

3. Durand-Raucher, Y., Y. Yim, and J. L. Ygnace. Traffic Information and Driver Behavior in Paris Region. Presented at the Pacific Rim Transtech Conference, Seattle, Vol. 1, 1993, pp. 167-169. 4. Kawashima, H. Informatics in Road Transport: An Overview of Examples in Japan and Future Issues. Proceedings of the 4th German-Japanese Seminar on Transportation Systems, 1991. 5. Peeta, S., J. L. Ramos, and R. Pasupathy. Content of Variable Message Signs and On-line Driver Behavior. In Transportation Research Record 1725, Washington, D.C., 2000, pp. 102-108. 6. Wardman, M., P. W. Bonsall, and J. D. Shires. Driver Responses to Variable Message Signs: A Stated Preference Investigation. Transportation Research, Part C, Vol. 5, No. 6, 1997, pp. 389-405. 7. Ullman, G. L., C. L. Dudek, and K. N. Balke. Effect of Freeway Corridor Attributes on Motorist Diversion Responses to Travel Time Information. In Transportation Research Record 1464, TRB, National Research Council, Washington, D.C., 1994, pp. 19-27. 8. Madanat, S. M., C. Y. Yang, and Y. M. Yen. Analysis of Stated Route Diversion Intentions Under Advanced Traveler Information Systems Using Latent Variable Modeling. In Transportation Research Record 1485, TRB, National Research Council, Washington, D.C., 1995, pp. 10-17. 9. Kraan, M., H. S. Mahmassani, and N. Huynh. Traveler Responses to Advanced Traveler Information Systems for Shopping Trips. In Transportation Research Record 1725, TRB, National Research Council, Washington, D.C., 2000, pp. 116123. 10. Polydoropoulou, A., M. Ben-Akiva, and I. Kaysi. Influence of Traffic Information on Drivers’ Route Choice Behavior. In Transportation Research Record 1453, TRB, National Research Council, Washington, D.C., 1993, pp. 56-65. 11. Pal, R. Investigation on Latent Factors Affecting Route Diversion Intentions. Journal of Transportation Engineering, Vol. 124, No. 4, American Society of Civil Engineers, 1998, pp. 362-367. 12. Jain, N. K. Modeling Drivers’ Route Choice Behavior Under the Influence of Advanced Traveler Information Systems. M.S. Thesis, School of Civil Engineering, Purdue University, Indiana, 1996. 13. Khattak, A. J., and A. J. Khattak. Comparative Analysis of Spatial Knowledge and En-route Diversion Behavior in Chicago and San Francisco: Implications for

Peeta and Ramos

24

Advanced Traveler Information Systems. In Transportation Research Record 1621, TRB, National Research Council, Washington, D.C., 1998, pp. 27-35. 14. Peeta, S. and S. Gedela. Real-Time Variable Message Signs Based Route Guidance Consistent with Driver Behavior. In Transportation Research Record 1752, TRB, National Research Council, Washington, D.C., 2001, pp. 117-125.

Peeta and Ramos

25 TABLE 1. Sample Socioeconomic Characteristics On-site Survey Sample (%) 78.6 21.4

Mail-back Survey Sample (%) 58.5 41.5

Internet Survey Sample (%) 73.5 26.5

< 20 20-29 30-39 40-49 50-64 ≥ 65

3.6 23.8 27.8 23.4 14.1 7.3

0.0 8.5 18.4 27.3 29.9 15.9

14.7 38.2 29.4 11.8 5.9 0.0

Education Level

High School or less Some College College Graduate Post Graduate

26.2 32.7 27.0 14.1

24.7 30.5 26.4 18.4

0.0 20.6 55.9 23.5

Household Size

1 2 3 ≥4

14.5 24.6 23.8 37.1

17.0 40.1 14.6 28.3

5.9 55.9 11.8 26.4

Attribute

Range

Gender

Male Female

Age Group

Peeta and Ramos

26 TABLE 2. Sample Diversion Characteristics

Sample Attributes Familiar with VMS Yes No

On-Site (%)

Survey Media Mail-back (%)

Internet (%)

---

84.1 15.9

79.4 20.6

Regular driver on the Borman Expressway Yes No

50.4 49.6

80.8 19.2

97.1 2.9

Familiarity with alternate routes Very familiar Familiar Undecided Not familiar Not familiar at all

32.8 32.0 13.6 17.6 4.0

59.7 29.9 4.2 6.2 0.0

72.7 24.2 0.0 3.1 0.0

Diverting under adverse weather conditions Yes No

73.8 26.2

89.5 10.5

82.4 17.6

Diverting at night Yes No

65.3 34.7

66.9 33.1

67.6 32.4

Diverting even when believing that alternate route would be longer Strongly agree Agree Undecided Disagree Strongly disagree

8.5 30.2 32.7 16.5 12.1

6.3 32.5 30.4 18.6 12.1

14.7 32.4 29.4 17.6 5.9

Travel time savings under work-related trip 5-10 min. 10-30 min. 30-60 min. More than 60 min. None

22.6 48.0 20.6 3.6 5.2

35.7 48.3 9.4 0.5 6.0

35.3 52.9 11.7 0.0 0.0

Travel time savings under personal trip 5-10 min. 10-30 min. 30-60 min. More than 60 min. None

9.7 37.5 38.3 10.1 4.4

21.5 53.3 18.9 4.7 1.6

35.3 41.2 17.6 5.9 0.0

Delay before diverting 5-10 min. 10-30 min. 30-60 min. More than 60 min. None

12.9 39.9 35.9 10.1 1.2

24.7 57.7 14.7 2.4 0.5

32.3 47.1 20.6 0.0 0.0

Peeta and Ramos

27

TABLE 3. Diversion Propensity under Different VMS Messages VMS Message Type 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8

Message Content

Relative Willingness to Divert 1 2 3 4 5 % % % % %

On-site Survey Sample Occurrence of accident only 13.7 Location of the accident only 20.2 Expected delay only 9.3 The best detour strategy only 7.7 Location of the accident and the best 2.0 detour strategy Location of the accident and the expected 0.8 delay Expected delay and the best detour 2.0 strategy Location of the accident, expected delay 1.2 and the best detour strategy Mail-back Survey Sample Occurrence of accident only 20.7 Location of the accident only 8.7 Expected delay only 5.0 The best detour strategy only 5.5 Location of the accident and the best 1.3 detour strategy Location of the accident and the expected 1.6 delay Expected delay and the best detour 1.6 strategy Location of the accident, expected delay 0.5 and the best detour strategy Internet-based Survey Sample Occurrence of accident only 26.5 Location of the accident only 5.9 Expected delay only 2.9 The best detour strategy only 2.9 Location of the accident and the best 0.0 detour strategy Location of the accident and the expected 2.9 delay Expected delay and the best detour 2.9 strategy Location of the accident, expected delay 0.0 and the best detour strategy

33.9 33.1 12.9 18.5

26.6 22.6 39.5 30.2

13.3 11.3 23.8 25.0

12.5 12.9 14.5 18.5

4.0

22.6

35.1

36.3

0.8

19.8

38.3

40.3

2.0

13.7

33.5

48.8

2.0

5.6

19.8

71.4

17.1 24.1 8.4 21.0

33.1 34.1 32.3 21.3

18.6 21.0 36.5 31.8

10.5 12.1 17.8 20.5

3.7

21.0

32.0

42.0

3.1

14.4

35.7

45.1

3.4

13.1

28.9

53.0

1.8

4.5

19.4

73.8

17.6 35.3 5.9 20.6

29.4 29.4 38.2 29.4

17.6 23.5 44.1 41.2

8.8 5.9 8.8 5.9

2.9

35.3

32.4

29.4

2.9

14.7

41.2

38.2

5.9

8.8

38.2

44.1

0.0

5.9

29.4

64.7

Peeta and Ramos

28 TABLE 4. List of Explanatory Variables

Explanatory Variable Mnemonics Alternative specific constant ONE Sex = 1, if male SEX = 0, if female Age group AGE = 1, if age ≥ 40 years = 0, if age < 40 years Level of education EDU = 1, if education ≤ some college = 0, if education > college Regular driver in the Borman Expressway region = 1, if Yes DRIV = 0, if No Familiarity with alternate routes = 1, if familiar FAM = 0, if not familiar Average commuting travel time to work COMM = 1, if travel time ≥ 30 minutes = 0, if travel time < 30 minutes Trust in information provided = 1, if high TRUST = 0, otherwise Amount of delay convincing a driver to divert DELAY = 1, if delay ≥ 10 minutes = 0, if delay < 10 minutes Dummy variables corresponding to each VMS VMSk message type, k = 2 to 8

Peeta and Ramos

29

TABLE 5. Binary Logit Models for Driver Response Under VMS

Variable ONE SEX AGE EDU DRIV FAM COMM TRUST DELAY VMS2 VMS3 VMS4 VMS5 VMS6 VMS7 VMS8 Sample Size L(0 ) L βˆ

()

ρ2

On-site Survey Coeff. (t-ratio) -1.942 (-10.45) 0.433 (3.26) -0.458 (-4.17) -0.308 (-2.74) 0.207 (1.87)

0.666 (5.84)

0.656 (3.83) 0.886 (5.22) 2.128 (11.78) 2.535 (13.16) 2.775 (13.73) 3.593 (14.27) 1984 -1375.20 -1037.75

0.245

Data Source Mail-back Survey Internet Survey Coeff. (t-ratio) Coeff. (t-ratio) -1.085 (-6.01) -2.033 (-3.51) 1.023 (2.73) -1.608 (-3.66) -0.148 (-1.69) -1.029 (-2.55) 0.446 (2.62) 0.834 (5.35) 1.509 (3.81) 0.298 (3.34) 1.048 (2.94) 0.190 (1.39) 1.100 (7.08) 1.386 (2.74) 1.013 (6.53) 1.075 (2.14) 2.004 (12.11) 1.860 (3.584) 2.408 (13.73) 2.970 (5.02) 2.479 (13.96) 3.205 (5.21) 3.604 (15.32) 4.638 (5.36) 3048 272 -2018.98 -122.81 -1583.66 -183.91

0.216

0.332

Combined Coeff. (t-ratio) -1.623 (-14.04) 0.123 (1.76)

0.169 (1.54) 0.540 (5.25) 0.435 (6.29) 0.311 (3.86) 0.848 (8.25) 0.888 (8.64) 1.938 (17.47) 2.364 (19.75) 2.480 (20.81) 3.472 (21.32) 5032 -3382.58 -2619.04

0.218