Implementation and evaluation of the Pedestrian Safety Engineering and ITS-Based Countermeasures Program was challenging. The major steps in the project included:

• Establishing and maintaining a multi-agency pedestrian safety team to oversee and guide the project
• Identifying pedestrian safety and mobility problems, including potential contributing factors to crashes
• Selecting pedestrian safety countermeasures corresponding to the problems identified
• Obtaining funding and support for pedestrian safety improvements
• Procuring, deploying, and maintaining the countermeasures
• Evaluating the effectiveness of the countermeasures

Each step of the project offered new challenges to the project partners that are presented here as lessons learned. The lessons learned presented and discussed in this document include two primary types of lessons:

• General lessons learned
• Countermeasure-specific lessons learned

The lessons learned are detailed in the following sections.

GENERAL LESSONS LEARNED

General lessons learned include those relating to the major project steps, as listed above. Nine specific lessons learned are discussed in detail in this section of the report.

Lesson #1-Assemble a Diverse Set of Project Partners to Address the Range of Issues That Might Arise During the Study

As part of their contractual agreements, each of the field teams was to assemble a diverse group, consisting of traditional and non-traditional partners, which would support the project by providing financial support and/or "in-kind" support through staff time and unique support. This task was deemed an important one for the projects' success.

Each of the field teams did initially establish a diverse set of project partners. Traditional partners included representatives from metropolitan planning organizations (MPOs), city and county departments of public works, State departments of transportation (DOTs), State departments of public safety, university research centers, and consultants. The teams were also able to include a wide range of non-traditional partners, including representatives from local police departments, local health districts, local school districts, community outreach programs, local advocacy groups, medical trauma centers, and local food and drug stores.

While there were a range of partners included on the project teams, a number of issues arose during the course of the project that necessitated specific personnel. In some cases, the teams were prepared to address the issues, and in some cases, the needed personnel were not always available.

Below are specific lessons regarding project partners:

• Recruit and actively engage a partner with experience facilitating public participation to ensure that the team is responsive to citizens' needs. The Miami team included the region's MPO, specifically the MPO's bicycle/pedestrian coordinator as an active partner. One of the primary benefits of this inclusion was the MPO's ability to handle public participation through its existing well-developed public involvement procedure. In this regard, the MPO was a natural fit and helped other team members focus on other project tasks. The MPO staff person also possessed a well-developed network of other individuals who could assist with public aspects of the project. The team cautions that this may not have been the case with a newer MPO staff member. A larger overarching agency, such as a MPO, is a useful partner in a pedestrian safety program in order to match needs with resources, gather input from citizens, and delegate concerns to proper personnel.
• The Las Vegas team included Clark County Safe Communities, an ongoing outreach program of the University of Nevada Las Vegas (UNLV) Transportation Research Center. The mission of Clark County Safe Communities is to reduce traffic crashes, and they regularly bring together a variety of agencies and other groups to address traffic safety needs. By involving this group, the Las Vegas team was able to reach the elderly and disadvantaged populations and to get public feedback on their countermeasures.
• Maintain as an active member of the pedestrian safety project team a representative from the agency responsible for installing roadway countermeasures to facilitate procurement and deployment. The San Francisco team found it helpful to have a representative from the City of San Francisco traffic engineering department as a primary team member for communication with engineers who were responsible for implementation. In Las Vegas, regular communication with project partners enabled the City and Clark County to assist UNLV in acquiring supplies through vendors. At times, UNLV would receive little response from countermeasure vendors. By going through the City or County, the requests to vendors received greater attention perhaps because the City and County were larger customers. This helped the project to stay more on schedule and allowed the team to acquire the needed countermeasures.
• Coordinate improvements with other agencies. The San Francisco team found that it was critical to coordinate improvements with agencies involved in roadway construction.
• Be sure to include public safety officials, as they can be effective advocates for pedestrian safety. The Las Vegas team found the local sheriff to be a key supporter for the pedestrian safety program, and over time, this has become a mutually beneficial relationship. The sheriff enabled the Las Vegas team to obtain detailed crash data from a local police department to supplement data provided by the Nevada Department of Transportation (NDOT). With the support of the sheriff, UNLV has joined monthly meetings of public works, law enforcement, and departments of transportation to gain acceptance and support for countermeasure deployment. During the meetings, UNLV can give the region's public safety agencies advance notice when and where collecting data will be taking place. This helps alleviate any suspicion when police spot the data collectors. In addition, the increased communication with public safety provides the team with external feedback on the effectiveness of the countermeasures. As an example, two to three weeks after the Danish offset was implemented, local police reported during a meeting that the countermeasure was having a noticeable improvement on safety.
• Do not hesitate to include multiple members from the same organization who work in different divisions. The range of pedestrian safety countermeasures being deployed necessitated a range of expertise not usually housed within a single division at an agency. Required expertise included traffic engineering, roadway design, electrical power supply, lighting, and communications. In the City of San Francisco, street lighting is under the control of the Public Works Department, while Traffic Engineering is in the Municipal Transportation Agency. Including staff from various divisions within the same agency will help with countermeasure approval, procurement, and deployment.
• Dedicate safety staff to the pedestrian program/project. Dedicated staff members are helpful in designing and implementing the program. The City of San Francisco Department of Parking and Traffic employs two full-time transportation engineers and one full-time planner in their program, and these staff members spent a considerable amount of time on the project. The Las Vegas team did not dedicate full-time staff to the program because they did not anticipate needing that level of effort. Looking back, they frequently wished that they had more staff help, particularly during the deployment and evaluation phases. They primarily needed more staff assistance with administering and coordinating tasks. Much time and attention was required to manage vendors and sequence deployments in multiple jurisdictions.

Lesson #2-Implement Regular Communication and Participation Mechanisms for Project Partners from Project Kick-Off

While there were a range of partners included on the project teams, in some cases there was active participation by particular partners, while in other cases there was lack of participation by the partners throughout the course of the project.

Below are specific lessons regarding contact and communication with project partners:

• Provide all partners with regular status reports and conduct regular team meetings and teleconferences to keep partners informed and engaged and to facilitate an efficient flow of communication. In Miami, agencies were busy and as the project moved into its later stages, did not want to schedule calls each week. Upon reflection, a web site or some other cost-effective means for keeping everyone up-to-date would have been helpful. This would have been perceived as a less intrusive way of keeping people involved and would have allowed for partner feedback to be elicited only on an as-needed basis. While this would be effective, some project budget would need to be reserved for initial setup of this forum.
• Keep a written record of key inter-agency agreements and events in order to preserve progress made in the case of personnel changes. Support from partnering agencies significantly waned during staff turnover in Las Vegas. Often, there was no written record of the promises or intentions of the previous staff members who were supportive of pedestrian safety. Within one partner agency, the participating staff member changed three times. Bringing the new staff members up to speed and re-establishing agreement on details of the program caused a project delay of a month or two.
• Strive to keep key players actively involved throughout the project. The City of San Francisco formed a pedestrian safety interdepartmental working group, but in hindsight, this was insufficient. Even though there was participation through this working group, it was not extensive enough. The Department of Parking and Traffic needed better coordination with public works and transit. They needed key players from public works and transit to be more intensively involved in the project, as opposed to just participating in meetings every few months. The consequences were that the team got the Phase I report approved by the working group but ran into opposition during implementation. In the end, they were not able to implement smart lighting or the pedestrian scramble due to resistance from public works and transit, respectively.
  
  In Las Vegas, an executive advisory committee was formed of the primary project partners, the Regional Transportation Commission, the Nevada Office of Traffic Safety, the City of Las Vegas, Clark County, the Nevada Department of Transportation, and the City of Henderson. The coordination of these six invested parties was made possible through the monthly executive advisory committee meetings. The team was described as "incredibly useful," particularly in the initial phase of the program. Because there were multiple partners representing multiple jurisdictions, the project schedule needed to accommodate extra time to gain agreement with each agency.

Lesson #3-Use a Variety of Methods/Sources to Understand Problems and to Determine Causes of Crashes at Prominent Pedestrian Crash Locations

As part of the project, teams were to identify and characterize pedestrian safety problem zones or areas through comprehensive data collection in the local jurisdiction, including traffic and pedestrian volumes, gap selection, vehicular speed, pedestrian behavior and crash type. Pedestrian and Bicycle Crash Analysis Tool (PBCAT) software was used to analyze crash patterns.

The teams found the best way to identify and characterize pedestrian safety problems was to use a variety of methods/sources:

• Review police crash reports and crash diagrams. The Miami team reported that without reading the police reports, and specifically the crash diagrams, it was not clear whether the pedestrian was in the crosswalk, some distance from the crosswalk, and/or from what direction the pedestrian was coming, and specific information of this type can have a large bearing on the selection of countermeasures. The crash reports helped the team identify true causation of particular crashes.
  The Las Vegas team suggested using data from police to get pedestrian safety information that traffic engineers may not have. Because the data from NDOT were not available for a certain time period, the Las Vegas team turned to alternative sources for crash data. The team found much more detailed crash data and crash diagrams from a local police department. The DOT crash data had around 100 different attributes for each crash whereas the police data had approximately 750 attributes per crash. While, most of the attributes were not used in this study, Las Vegas did benefit from the crash diagrams and the elaborate road characteristics data.
• Review crash records from at least the previous five years. Looking at too few crashes might be misleading. The San Francisco team found a great deal of variability from year to year in the types of pedestrian collisions that occurred at the sites. For example, during a five-year period at one site a large majority of the pedestrian collisions occurred late at night. Comparing that five-year period to the next five years (or even a rolling five-year period), there was a dramatic change in the number of pedestrian injuries late at night that could not be otherwise explained.
• Visit high crash locations with crash data in hand. The Miami team visited each of the corridors to observe drivers, pedestrians, and pedestrian facilities. The members used a book with the specifics of each crash at each location including information on demographics, temporal variables, crash severity, and crash type. Simultaneously looking at the data and the crash location was helpful in determining the causes of the crashes. Additionally, the Miami team found site visits helpful because they were able to roughly assess driver and pedestrian behavior and examine existing engineering devices.
  The Miami team reported that the visits allowed them to confirm or disprove hypotheses that had been generated about the cause of crashes in each corridor based on their initial use of the PBCAT. PBCAT is a crash typing software that can assist state and local pedestrian/bicycle coordinators, planners and engineers improve walking and bicycling safety through the development and analysis of a database containing details of crashes between motor vehicles and pedestrians or bicyclists.
• Use surrogate measures when pedestrian volumes are not available. Surrogate indicators of pedestrian volume may include tourism data, bar or restaurant service volume, and others and can provide a more accurate picture of the number of pedestrians and related safety problems.
• Seek input from the project's advisory committee. UNLV presented their project's executive advisory committee with data to review and the partners provided a valuable non-technical perspective regarding the placement of countermeasures.

Lesson #4-Begin the Program by Implementing Low-cost Countermeasures for the Greatest Potential of Widespread Use

As with many programs, starting off with "quick wins" builds momentum and support for the program overall. In the case of these pedestrian safety programs, the countermeasures varied greatly in terms of price, procurement difficulty, ability to receive approval, and ease of deployment. The Las Vegas team found it beneficial to implement several low-cost, easy-to-deploy countermeasures early in their program, which allowed them to show results to their partners during the monthly executive advisory meetings. Countermeasures such as high-visibility crosswalk pavement treatments, in-street pedestrian signs, and pedestrian push buttons that confirm press were relatively easy to deploy because the local agencies could perform the labor without having to use a vendor.

For the most potential success:

• Promote wide-scale dissemination and adoption of the countermeasures. The strategy of initially deploying low-cost countermeasures has the greatest likelihood of wide scale dissemination and adoption, because countermeasures tend to be most effective when applied in multiple locations and in combination with other treatments.
• Do not deploy initial countermeasures sparingly. Because most treatments work best when applied at multiple locations and in combination with other treatments, it is important that the initial treatments are not so expensive that they can only be used sparingly.
• Help the program conserve its momentum. The high cost/effectiveness ratio of the low-cost initial applications will help ensure a series of "quick wins." The San Francisco team found that low-cost but effective measures have the advantages of quick implementation and the potential to draw support and funding for further improvements.
• Introduce more expensive interventions over time using a phased approach. Starting off with low-cost countermeasures allows the time required to budget for and procure more costly interventions, should they be required.

Lesson #5-Pursue a Variety of Funding Sources for the Pedestrian Safety Program

The local deployment teams were tasked with obtaining matching funds for their pedestrian safety projects. As pedestrian issues are not always the top priority of state and local transportation agencies, often falling behind roadway construction and maintenance projects, obtaining funding for pedestrian safety countermeasures can be challenging. Therefore, it is advantageous to pursue a variety of funding sources for pedestrian safety countermeasures:

• Gain political support and funding through high-profile demonstrations. The Miami team gained the support of the mayor by involving him in a public demonstration of a rectangular rapid flashing beacon at a pedestrian crossing, which has led to him trying to get more of them installed in the city. In Las Vegas, the MPO was impressed by the RRFD and supports additional installations.
• Bring together funding from several agencies. The Las Vegas team pieced together funding from five different partners, including NDOT, the Nevada Office of Traffic Safety, the City of Las Vegas, the local MPO, and Clark County Department of Public Works. In addition to funding, many of these agencies provided the labor necessary to install countermeasures, which conserved funds for other project needs.
• Apply for grants. The San Francisco team received a grant for pedestrian countdown signals on two corridors through the Safe Routes to School program.
• Consider funding from taxes. In San Francisco, there is funding available annually for "Pedestrian Circulation and Safety" projects from sales tax funds, and the San Francisco team was able to secure some of these funds for installing countermeasures.
• Assure sufficient staff labor. For a major metropolitan city, the staff labor to implement innovative pedestrian safety countermeasures can be far greater than the cost of the equipment and materials. This time includes:
  • Planning and designing installations
  • Coordinating with other departments/agencies
  • Obtaining legal approvals
  • Coordinating shop work or contracts
  • Managing interdepartmental financial arrangements
  • Purchasing equipment
  • Installing countermeasures

Lesson #6-Do Not Underestimate the Complexity of Procurement

The procurement of innovative countermeasures presented several obstacles for the pedestrian safety teams. Lessons regarding procurement include:

• Be aware of which vendors are approved by your department. For one of their relatively inexpensive countermeasures, the San Francisco team had to go through a distributor because the manufacturer was not an approved vendor. The use of a distributor added complications as well as costs to the project.
• Establish policies and procedures for procurement as early as possible in the program. Because the Las Vegas team was working with multiple jurisdictions, the approval process was complex. It often took much more time than anticipated because the steps were not always clear from the onset. With regard to the rapid flash beacons, Las Vegas had to obtain a permit from NDOT to install the beacons on an experimental basis. This was necessary because the beacons were not yet in the Manual on Uniform Traffic Control Devices (MUTCD) and the partners did not want to be held liable if the countermeasure showed adverse effects.
• Be prepared that vendors may have challenges producing new countermeasures. In Las Vegas, they were unable to find a vendor that was willing to produce enlarged pedestrian signal heads because of the expense and risk involved in creating a new design. If there does not appear to be a high demand for a product, a vendor may not be willing to risk a new design. In the case of pedestrian countdown signals with "animated eyes," Las Vegas experienced substantial procurement delays. Although the vendor was very helpful, they had to ship the system back and forth to the vendor due to software issues. Las Vegas also experienced a delay in schedule while working to get permission from the patent holders of this countermeasure.

Lesson #7-Budget Ample Time for Deployment and Coordinate with the Appropriate Jurisdictions

Developing and implementing a comprehensive pedestrian safety program and plan requires a multi-year time frame. Deployment of a wide variety of countermeasures in numerous locations across an extended period of time presented a challenge to the field teams. The Las Vegas team faced additional challenges, as the deployments spanned three separate jurisdictions, including the City of Las Vegas, Clark County, and the Nevada Department of Transportation. Several ways to mitigate project delays due to deployment and coordination include:

• Coordinate deployment activities with local intersection construction plans. The San Francisco team had to repair countermeasures damaged by major intersection construction that occurred just following countermeasure deployment.
• Budget additional time for implementation when deploying on roads owned by the State but operated and maintained by the city or county. The Las Vegas team needed to have designs approved by NDOT and then deployment scheduled with the city or county.
• Maintain persistence when obtaining permits to deploy pedestrian safety improvements. The Las Vegas team faced unexpected obstacles in obtaining a permit from NDOT for a pedestrian refuge island. The team had to address multiple objections but eventually received the permit after six months. The Las Vegas team was able to overcome the objections by contracting with a private engineering firm to produce the needed drawings. This increased the project cost and added four to five months to the schedule.
• Consider the trade-offs of deploying in-house versus using an outside contractor. The Miami team had a hard time getting the local agencies to install various countermeasures when they were supposed to. The team thought that having its own contractor for deployment would be more effective, but it proved to have its own problems. As a result of these difficulties, schedule adherence suffered.

The San Francisco team also considered some of the trade-offs. Overall, they felt it was better to handle the deployment in-house, but contracting would have had some advantages. There are many traffic engineering and public work projects going on in a city like San Francisco, and it can get very complicated. When a project like this is managed in-house, it is easy to get caught up with other priorities and project needs, and turn-over in staff can cause further complications and delays due to the time needed to replace staff. As they moved into implementation, the San Francisco team needed an engineer to manage the project. While getting a contract in place for an outside contractor offers complications and takes time, once a contract was in place, a consulting firm could have handled the deployment faster that the City was able to.

Lesson #8-Consider How the Timing of Countermeasure Deployment May Impact the Experimental Design and Evaluation

As part of the project, the field teams were required to conduct an evaluation of the countermeasures they deployed. To do so, the teams selected experimental designs for their evaluation. The Las Vegas team's deployment plan involved implementing a variety of countermeasures at particular sites in a staged approach, and their evaluation plan involved evaluating the impacts of each deployment. The San Francisco and Miami teams' deployment evaluation plans involved using a staggered approach to evaluation. This staggered approach made use of two deployment sites, each serving as a "control" site for the next at different points in the evaluation process.

• Consider a quasi-experimental design if the installation of the countermeasure cannot be controlled. The San Francisco team had planned to use a staggered design but ended up using a before and after design. First, they found it was difficult to get an exact timing for the installation of countermeasures at the two deployment sites for each countermeasure. When the agency was putting in the countermeasure at one site, they generally wanted to put the countermeasure in at the other site at the same time. In effect, the team asked the agency to delay installation at one site for six weeks, which created an extra burden on them. Therefore, the team had difficulty controlling the amount of time that elapsed between when the before data were collected and the installation of the countermeasures. Perhaps more important, the San Francisco team suggests that even if they could have pulled off the staggered design, they are not sure how much it would have worked. Their experience was that the real "control" was what was going on in the city in general, not at an intersection across town that was selected as the control. A single intersection might not always help control for confounding factors.
• The Las Vegas team's deployment plan involved implementing a variety of countermeasure at a particular site in a staged approach, evaluating the impacts of each new deployment. The team was initially going to deploy in five different stages; however, much time and attention was required to manage vendors and sequence the staged deployments in multiple jurisdictions. As a result of the deployment delays, they had to combine countermeasure deployments, which impacted their ability to evaluate the individual countermeasures. If multiple countermeasures were inadvertently deployed within the same time period, the team had to scramble to meet the data collection needs. In the end, they were able to evaluate the individual countermeasures only because early in the project they had deployed some of the countermeasures separately.

Lesson #9-Consider the Unique Aspects of Collecting and Reducing Pedestrian Safety Data

When safety evaluations are conducted over time periods too short to rely on crash data, it is necessary to use surrogate measures of safety to determine the impacts of the deployments. In this study, the field teams used a variety of safety surrogates, including: pedestrian behaviors (e.g., violating the pedestrian signal, crossing against traffic, looking before crossing) and driver behaviors (e.g., blocking the crosswalk, yielding, coming to a complete stop). In addition to the safety surrogates, teams collected data related to pedestrian mobility (e.g., pedestrian volumes, traffic volumes, pedestrian delay), customer satisfaction, and demographic information associated with observed pedestrians. Collecting this amount of data, including behavioral data of both pedestrians and drivers, presents some unique challenges. Lessons learned during their data collection activities include:

• Remember that vehicle and pedestrian peak periods do not necessarily coincide. When planning data collection activities that involve observing both driver and pedestrian behaviors, it is important to consider the peak periods for both vehicular traffic and pedestrian traffic. Collecting data only during the peak periods for vehicular traffic may not result in the optimal observations of pedestrians.
• Keep MOEs simple and repeatable. Clear, consistent definitions of MOEs are helpful, but difficult to achieve. In particular, there is no universal, accepted definition of "vehicle/pedestrian conflict." When assessing the effectiveness of the countermeasures or pedestrian safety program, select MOEs that can be defined and for which data can be collected consistently across staff, locations, and time. Regardless of what MOEs are selected, a data collection protocol and corresponding training should be developed and implemented.
• Provide field data collection personnel with a formal letter explaining their purpose and contact information. Collecting behavioral data in the field can require multiple personnel, video cameras, and other equipment. Police in Las Vegas received a call from a citizen concerned with the "suspicious" activities of several individuals with clipboards and video cameras on a street corner. When the police came to investigate, the Las Vegas team was able to produce a letter that instructed the officer to call a lead researcher at the university. The lead researchers confirmed the names of the individuals participating in the data collection, which relieved suspicion.
• Consider the advantages and disadvantages of using field data collection versus video data collection. Pedestrian safety data can be collected by observers in the field, or using video cameras, and there are advantages and disadvantages to both:
  • While video data may allow for greater accuracy by giving precise time stamps of events and allowing analysts repeated viewings, it requires substantially more time to analyze. The labor requirements for tabulating video recorded events were several times greater than for manual data collection.
  • The San Francisco team reported that while using video data complicated the project, it made it more interesting.
  • Video is very good in allowing for review and in obtaining precision, but is not as good at observing subtle details like gender and age and whether people are looking left or right. In addition, the video field of vision was often restricted.
• Pre-plan and pilot test the location and angles of the video cameras. The San Francisco team had difficulty selecting locations that were well-suited for a particular countermeasure and getting the right camera angles. As a result, they recommend developing a protocol for pilot testing the video data at each location, including site review for optimal camera placement, obtaining video, and spending some time analyzing the pilot version to make sure the camera angles are appropriate for obtaining the MOEs for the countermeasure. While this process may add one to two days of work per location, in the long run it saves time and resources.
  The Miami team experienced similar difficulties with video data collection. Their initial test of the technology happened to be done at a location conducive to this type of data gathering, while most study locations proved much more difficult and complex. In retrospect, a more detailed field review would have been advisable and, given the difficulties experienced, using observers may have been preferable to using video.
• Allot extra time for reducing video data. The San Francisco team estimated that they needed four to five hours to reduce one hour of video data. They further estimated that the decision to use video added 20 - 30 percent to the costs and time needed to complete the project, adding nearly one year to the schedule.

The field teams had a large amount of data to collect at a variety of different sites. Each team considered live, on-site data collection, as well as the use of video cameras. There are a number of advantages and disadvantages associated with both data collection approaches. Each team used video to some extent to collect data, and they found a number of ways to expedite the reduction of the video data:

• Develop a customized/automated video analysis software tool to extract data from video images. The San Francisco team developed a customized video analysis software tool. Users type F keys to mark different events, and the tool records the events with timestamps directly into a spreadsheet. The tool allows users to play the video at different speeds and to go frame by frame. The tool greatly improved the ability to extract data from the video images. The Las Vegas team found that one hour of video required analyzers to rewind the video 8 to 10 times. Automated templates offered a more efficient process for data analysis.
• Link signal cycles to video time stamps with automated methods. Linking the signal cycles to the video time stamps was necessary to record several of the measures of effectiveness. For a couple of reasons, the signal timing information was incorporated once the video data were collected, rather than have the data recorders observe both the signal and the pedestrian/driver behaviors. First, it was difficult to position the camera so as to see the signals and the roadway behaviors. Second, even when the cameras could be positioned to see the traffic signals, it was difficult for data recorders to judge who was where and during which signal phase. To link the signal cycle information to the video data time stamps, the San Francisco team developed a computer program which saved the team a lot of time. Once it was set up, which added about one week to the schedule, it allowed a good deal of precision and took burden off of the observer.

COUNTERMEASURE-SPECIFIC LESSONS LEARNED

This section presents lessons learned regarding specific countermeasures. Lessons learned associated with two countermeasures in particular are highlighted below.

Lesson #1-Strategically Place In-street Pedestrian Signs to Reduce the Chance of Them Being Hit by Vehicles and to Maximize Their Effectiveness

In-street pedestrian signs are placed in the center of the roadway prior to or at the crosswalk. Their purpose is to alert drivers that they must yield to pedestrians. The placement of the signs in the roadway, close to drivers, was expected to increase yielding behaviors over the more traditional signs that are placed on the side of the roadway.

All three field teams installed and evaluated in-street pedestrian signs, and all had issues with the signs being hit and destroyed in many locations. There are a number of ways to reduce the occurrence of the signs being destroyed by vehicles:

• Place signs on medians to reduce damage by motorists. The Miami and San Francisco teams recommend placing the signs on raised medians, and the Miami team suggests using low-level foliage on the medians to reduce the percentage of motorists that strike them. Locations with a pedestrian median or refuge are likely to be most effective.
• Place only one sign per approach and at the crosswalk. The Miami team conducted a parametric analysis of the relationship between sign location and yielding behavior to determine if placing multiple signs along an approach would lead to larger increases in driver yielding behavior. The team placed signs in multiple locations on the uncontrolled approaches to three two-way stop-controlled intersections. Based on the results, the Miami team recommends using only one sign per approach, as well as placing the sign at the crosswalk. Overall it appears that installing the signs at the crosswalk line is as effective as or more effective than any other location along the approach or than installing three signs on one approach.
• Do not use signs in locations with high truck or bus traffic. At locations with high truck traffic in Las Vegas, the signs lasted between 48 hours and 2 weeks. For example, the Las Vegas team deployed a sign near a cement mixing plant where cement trucks were frequently turning. The lifespan of the signs at that location was less than 24 hours. At locations with fewer trucks, the signs remained intact. The San Francisco team recommends considering bus routes when determining where to install the signs.
• Carefully consider turning movements and lane width when determining locations for sign installation. The San Francisco team experienced high damage rates to signs located near left-turn paths at intersections. They recommend placing the signs carefully by taking into account car and truck turning movements as well as lane widths in order to reduce damage to the signs.

Lesson #2-Consider the Technical Issues Surrounding the Use of Automated Pedestrian Detection

Pedestrian detection technologies, including microwave and infrared detection devices, provide the means to automatically detect the presence of pedestrians in the targeted curbside area. Technologies may also be used to detect pedestrians moving in the crosswalk. When used at the curbside area, pedestrian detection may either replace or augment the standard push button used to activate the pedestrian call. When pedestrian detection is used to detect pedestrians in the crosswalk, the purpose is to detect the presence of individuals requiring additional time to cross and, accordingly, to extend the clearance interval and to provide more time to cross.

Both the San Francisco and Las Vegas teams implemented pedestrian detection technologies in their studies and offer some lessons regarding the technology:

• Consider In-Surface Activation Devices (ISADs) as an alternative to bollards for infrared pedestrian detection at busy street corners. San Francisco used ISADs instead of two waist-high bollards to activate flashing beacons at a corner where there were a lot of crossings.
• Keep corners with automated pedestrian detection clear of parked cars to avoid false detections. In Las Vegas, parked cars were detected by the pedestrian detection device. To mitigate this issue, the team made the no parking zone near the detectors more obvious by painting the curbs red.
• Ensure that the signal control logic and the detection device are configured for proper communication. San Francisco had significant issues to overcome in making the signal controller logic compatible with the notification to extend the green from the pedestrian detection device.
• Consider building the detection equipment in-house. The Las Vegas team had difficulty getting responses from vendors for automated pedestrian detection systems. As a result, UNLV worked with its partner, the City of Las Vegas Department of Public Works, to build the detection equipment in-house. The sensors and radio transmitters were purchased.

Additional lessons learned regarding other countermeasures are shown in Table 83.