Washington State Transportation Centre

The majority of the Seattle part of the tour was based at the facilities of the Washington State Transportation Centre, (TRAC) where I was kindly provided with office space I.T. facilities and access to an impressive library of research documents. As a well equipped research and development organisation, TRAC gets involved in many transport and transport information initiatives including new Intelligent Transport System (ITS) projects and provides a useful link among the government, university researchers, as well as the private sector. Much of the research projects are funded by the State Department of Transport (WSDOT) with TRAC acting as a liaison body between WSDOT, who require new initiatives, and the Universities who are in the best position to conduct research and trial potential projects. The links with the local University of Washington are strong with a good focus on student involvement in transportation research as well as report editing and production.

An introduction to the TRAC facility and the staff was provided by Mark Hallenbeck who then gave an impromptu tour of the Seattle city freeways to best demonstrate the traffic problems caused by geographical limitations together with the added severity caused by roadworks. Current major works underway during my stay included various lane closures on vital bridge links. These were to enable bridge repairs on a support that had been recently damaged in an accident involving a barge.

Mark also introduced me to the vast library of TRAC reports available regarding evaluation of ITS related projects. An overwhelming information source to tap into during my stay at the TRAC offices.

Interviews were arranged with members of the TRAC team, including Catherine Bradshaw Boon, an excellent source of knowledge for many ITS projects and who had a key role in the design of the King County Metro Internet Site.

King County Metro Transit Online

As stated previously, this report doesn't need to focus on the content and perceived quality of Internet sites due to the fact that they can be accessed readily by anyone who has access to them. However, it was valuable to talk to Catherine who designed a lot of the content for the King County Metro Transit Internet site (www.metrokc.gov) and also had access to the feedback received. The Web site was first installed in mid-1998 as part of the Smart Trek Metropolitan Model Deployment Initiative (MMDI), a federally funded ITS program.

At the time of my visit, the state of the Metro King County public transport information Internet site was very much in the format that it was back in 1995, prior to its launch on the world wide web. This can either suggest that the format is acceptable or that no resources exist to freshen the site. Actually, an overhaul of the layout of is pending but not currently flagged as a priority by Metro K.C. Transit who have the final say, and who would be involved in financing such an overhaul. Nevertheless the site is impressive and developed from a system originally designed for employees of companies with a vested interest.

The content has everything you would expect from a public transport information source including a hierarchical route mapping system, timetables and fare information.

The route maps generated are the actual graphics generated for the paper versions of the timetables. The obvious advantage here being the minimal amount of effort required to make them available for the Internet. The multi-tier mapping actually looks very simplistic, mainly because it was developed using older computer systems capable of showing only twelve different colours. Actually, and without knowing it, King County had developed a map displaying good contrasting colours useful for people with visual impairments.

Drawbacks with the system, however, included an initial difficulty to maintain the site due to the manual nature of converting timetable databases into web pages.

This issue has now been addressed and the timetable web pages are generated automatically from the timetable database. This has also had the effect of eradicating the lengthy and tedious proof reading process that existed with manually entering the timetables. The output is kept simplistic and therefore is very easy to read. The fare structures for King County services are fairly simple and so this information is easy to provide in Internet form.

View larger version	View larger version

Criticisms of the system, from both designers and users alike, include a lack of user friendliness in the location descriptions. The timetable headers themselves include descriptions such as Aurora Nth & 85th instead of using landmark names and colloquialisms. The criticism being levelled is that that some knowledge of the road system is required in order to make the best use of the system. From my point of view, a tourist armed with just a few maps, the Aurora and Nth 85th was ideal. Feedback from Seattleites suggests that well known landmarks would be preferred. This raised an interesting and very important point when considering public transport map design for Internet access. As far as I was concerned I didn't know all the local landmarks (actually the Space Needle was probably my limit) and I actually found the site easy to use and quickly had a list of routes, as well as maps, timetables and fares from my accommodation, in the north of the County, into Downtown and beyond.

Ideally the maps and titles would convey landmarks and road intersection information and perhaps if the graphics had been redeveloped for the Internet rather than taken directly from the paper designs then this would have been the case. In my opinion the route maps available, with every single route accessible electronically, offer a lot of help in getting around and about in King County. Seattle, yesterday

If travellers have little idea of where their destination or even current location is then they have little hope in gaining any useful information from the system anyway. Therefore it is surely fair to assume that commuters have some degree of knowledge about the layout of the city. The other salient point here is that the hierarchical mapping system could be put to good use in Nottingham where current service information is accessed from an alphabetical table of locations around Nottinghamshire. Ideally, some national public transport journey planner, processing data entered into forms as queries and outputting road, rail, timetable, fare and mapping information, as well as expected delays and problems, would be nice.

Further development of the electronic timetable system is planned to enable web users to request any timetables which will then be automatically generated on the fly from a database of timetables. The advantage here being that the processes of converting from database to Internet whenever schedule changes occur are totally eradicated as all information obtained from the Internet will be as up to date as the main system database.

The major disadvantage with the current system include the fact that only Metro King County services are catered for. Several inter county services are available but not accessible from this particular information source. However, with the development of a new trip planner it is hoped that at least three of the State's counties will be covered in terms of public transport information. This development is covered later in this chapter. Other major feedback includes requests for text only versions, additional County information, more system maps and cycle route maps.

The fact that such a quantity of feedback was achieved is indeed impressive as we have found such results hard to obtain in Nottingham. There certainly seems to be little interest in responding to our request for qualitative feedback about our Internet site in general yet we are keen to gain access to such feedback to constantly improve and tailor our services. However, it is useful to see any feedback and the responses to the King County information services are valuable in this respect. It is also interesting to note just how so much data was gleaned. Are the citizens of the U.S. more likely to speak their minds or have an opinion or was the solution in the methods used to get the information? e.g. A feedback form published on the Internet or questionnaires handed out to commuters etc.

As I was granted access to basically any research document from the TRAC library I considered the "Survey of users of Metro Online, the official Web site for Metro Transit" a particularly useful one. Actually they all appeared relevant as well as interesting but I figured that the chances of my stay in Seattle being extended by six months were slim to none. Nevertheless some of the reports were available on line in a Members Only area of the TRAC web site. Access to this area was kindly granted by Pete Briglia of TRAC. Anyway, this particular survey, carried out in July 1999, gleaned responses from 571 users of King County Metro Transit services. This doesn't sound many out of a population of 1.65 million and the sample was not representative of the population of users of the Internet site. In addition, it is was not possible to estimate the size of the total user population from the survey responses. That said, the responses carry enough information to be useful in the development of King County Metro services and information systems as well as Internet sites of a similar nature (e.g. www.itsnottingham.info) and public transport companies in the U.K.

The results are often fascinating, especially direct comments about services and / or information provision. The reason I have included virtually the entire report in Appendix (iii) is due to the sheer quality of the feedback. Not just qualitative feedback but extra direct comments about service provision and information access. It was one of the most relevant pieces of feedback relating to the access of public transport information via electronic means and demonstrates wonderfully the differing views that would be gleaned if we could achieve a similar questionnaire response in Nottingham. An additional conclusion to the report may be that it is impossible to provide the perfect journey planning system for everyone but it is possible to come quite close. In terms of aesthetics and interface design there must be literally millions of opinions about what works and what doesn't and what patrons like or prefer. Whilst aesthetics are very important the core of any system is its functionality and if that fails then the system is worthless. Refer to Appendix (iii) to read the report.

TRAC Information Kiosks

Another development experiment in information distribution came with the deployment of e:kiosks (electronic information kiosks) which reflected the services of King County Metro. These were installed in 1995 mainly in office environments and were initially aimed at commuters. Other kiosks were then installed at numerous bus and ferry terminals utilising fast ISDN (Integrated Services Digital Network telecommunications) Internet connections. They have all subsequently been removed from service, the last ones as recently as July 2000, due to the lack of use. At some employer sites access was as low as ten times per month although much higher at establishments such as cafes. Problems with the actual hardware, including the need to frequently reboot, sometimes on a weekly basis, also added to the downfall of the project. Strangely enough, in a country where it is assumed that computers are used for every conceivable application, the kiosks were deemed difficult to use because knowledge of the Microsoft Windows operating system was a requirement. A surprising yet important statistic I feel, highlighting the fact that computer access and windows based operating system concepts, whilst widely used are not used by everyone and skills that many may take for granted have not totally infiltrated society. Yet. With the possibility of such kiosks being located around Nottingham this was an interesting part of the discussion and provided some idea of barriers to be overcome to ensure the success of such a project back home. I tend to think that hardware has improved in reliability and stability since 1995 as well as in information processing performance and communication speeds. It is also fair to say that over the past six years, the usage of desktop computers for both work and domestic use in the U.K., with the Windows operating system, has undergone a massive growth.

Even the state program to deploy traveller information kiosks has been shelved based on the problems encountered with Seattle's pilot scheme.

The kiosks are described in slightly more detail in the Riderlink section later on.

TRAC Trip Planning

Continuing my debate with Catherine, we talked about possible new or planned information systems, not necessarily designed to be Internet applications. The first of which was a multi-modal the trip planner initially designed as a tool for a dedicated Call in Centre. I questioned this fact with Catherine as the Internet continues to grow in the States why not develop such a system for the web as well. The the extra overheads would surely be minimal i.e. the systems would be more or less identical. Strangely enough there apparently exists quite a stigma about providing Internet services that may take employment away from such call centres. Personally I would be very surprised if this didn't eventually develop as an Internet based application and look forward to seeing it as such in the future.

In terms of our discussions back in Nottingham regarding the expansion of traveller information services we have discussed Internet applications, leaflets, call centres, kiosks and in station advertisements, to name but a few. In our fairly educated opinion the Internet does not replace manual interaction but enhances it and in certain cases make information more conveniently accessible. Let's face it, what could more convenient than accessing timetables, fare tables, maps, incident data from the warmth of your own lounge? Granted, not everyone has access to the Internet at home but the technology is spreading as well as evolving (Internet television) and it going to disappear or be phased out.

The main criteria for journey planning systems should be that information is as accurate, timely and as user friendly as possible regardless of the output medium (which ideally would be several different types).

An interesting fact concerning this system is that user feedback on current traveller information systems indicated the desire for such a system. Development is not currently a priority for Metro but it should be available soon.

TRAC BusView

As well as engaging in various debates with various members of staff, I was granted access to their computer systems, which enabled me to evaluate various systems for myself. Such systems included BusView and MyBus which allow users to access real time bus location information.

The King County Internet based BusView system is a Java Applet* specifically designed for Internet browsers. It was, however, designed by software engineers without any user input which has resulted in a system that looks like a technical tool and not a serious end user system, especially as it launches into a mapping system comprising simple wire-frame graphics, with moving white blocks to represent transit vehicles.

As a concept, were talking simplicity itself. You can view current locations of buses or specify your own bus route. The bus information automatically updates approximately every minute. Designing this kind of system without input from end users is generally not perceived as best practice. In this case, the users would probably have chosen to have road names / numbers displayed automatically without having to move the mouse over the intersections to see them (the city is largely grid format so a lot of the bare map sections are difficult to place). However the system isn't overly complicated and it rightly assumes a certain amount of user intuition as any tool like this would. * Java is a fully fledged computer programming language which is often used to write small programs or applications accessed by Internet pages. These programs are called "applets" because they are literally small applications.

There is always a danger that a lot of effort could be spent on systems that are technically excellent, but fall down on the user interface side. BusView has managed to be a useful commuter tool and the front end can be redeveloped without a major technical overhaul of the system. However, there were no plans to do so during the time of my visit.

The display often shows the graphical bus tags overlapping. This can look like a flaw with the product but the display can be altered in a couple of ways. Firstly, hidden or partially hidden tags can be made fully visible using the mouse and automatic drop down menus, even if this does slightly complicate the user interface. Additionally the display can be limited to show just a single route and recent developments include map "book-marking" (i.e. customise the display to suit a particular journey and then bookmark it in with your Internet browser software for quick access later). The software is regularly updated and improved for stability and usability.

Whilst demonstrating ably what is achievable, by linking technologies such as Internet and radio tracking of vehicles, it also demonstrates how difficult it can be to develop a product usable by everyone. Too little map detail and the system becomes of limited use and places the onus on the user to work out what the map is actually showing. Too much detail and updating the map could become a slow affair when refreshing although different Internet techniques could resolve this, for example, updating only sections of the map graphic that have changed. The saving grace for Busview is the fact that it is customisable to whatever routes the user wishes to select and save. For commuters completing the same journeys on a regular basis this is ideal.

The following is a sample of comments gleaned during the Metro Online Survey.

Forty-six percent of the respondents who had used BusView reported that they found it very or somewhat easy to use. Seven respondents also made open-ended comments related to BusView. Three of them were specific suggestions for improvements:

• "You should have BusView in all tunnel stations to show the buses in the tunnel I find it very useful seeing when the next bus is scheduled and how late it is."
• "BusView would be better if I knew that the bus I was looking at was the one I wanted, that is, is it the 4:45 or the 5:05, etc."
• "BusView is not very useful since it only has a few small maps on it. Since my routes/maps aren't shown, it doesn't help me. Once you get the whole system up, then it will be a useful feature."
• Another four respondents generally commented that they had not yet accessed BusView though they were eager to do so. One of the respondents stated that
• "The main reason that I have not found BusView more useful is that I use Netscape* on a Mac**, and on the only occasions I've tried it has not even come close to working. (Java isn't very stable with Mac Netscape). But I love the idea and hope that eventually I will get it to work."
• "Provide more or larger bicycle racks on buses that cross Lake Washington. During the summer a cyclist may have to let several buses go by before one comes that has an open space on its bike rack."
• "The 128 bus is generally 5-8 minutes late in the afternoons - it costs me an hour because I can't get my next route."

* Netscape Navigator - Internet browser software

** Apple Macintosh Computer

TRAC MyBus

MyBus is another Java based Internet application that shows real time service information in the form of a simple table. King County Metro installed their Automatic Vehicle Location (AVL) system in 1992 and the data fed back from this is now utilised in such systems as MyBus. Skills at the nearby University of Washington were utilised to develop several algorithms to assist in the location prediction system using the Metro King County base data as input. The information required can be obtained by selecting from a list of bus stops around Seattle. Maps can also be generated to verify the location selected before the results table indicates which services are on time or delayed. Again, some geographical knowledge is assumed as the user has to select from a list of stops with road name descriptions.

Recent developments in this system have included the availability on Wireless Access Protocol devices. This can be viewed by accessing www.mybus.org/wml/ and selecting MetroKC from the MyBus menu on any WAP capable device. User inputs for this system include an identifying code for the requested location. Locations are available for time-points only and predictions at individual bus stops are currently unavailable. There are, however, over 1000 time-points to choose. As in the full Internet version, MyBus then displays the predicted departure times for buses leaving from the selected stop. If MyBus has no current real-time information about a bus, the scheduled departure time is shown in place of a predicted time.

Internet version.  Step 1 : Choose your stop	Or verify your location on the map
Check for delays

Whilst current WAP displays are concise, and generally have a lesser graphics capability than standard desktop machines, the advantage lies is in the fact that you don't have to be tied to your computer to get the latest bus location information. The display shown here may appear confusing at first but becomes familiar with use. WAP sites will almost definitely develop in the same way that current Internet sites already have. The potential for WAP may be even greater due to the ability to access information on the move. Not only do BusView and MyBus demonstrate different solutions to an ITS problem but also indicate what can be achieved in partnership.

Without the co-operation of Metro King County and the University of Washington it is doubtful that either system would have been developed.

TRAC Riderlink

Not all systems developed by TRAC have developed into fully fledged end user systems. A small part of the Riderlink project had to be shelved, ironically down to the fact that it was so successful. Riderlink was set up with the goal of providing access to public transport information regardless of the public transport provider. This naturally extended to include various modes of transportation e.g. buses, ferries, cycling and even promoted car sharing. Car sharing in Seattle allows commuters to take advantage of the many highway high occupancy vehicle (HOV) lanes that exist. The system, which evolved as an Internet based project, despite this not being the original intention, included form based trip planning whereby the requests would be sent from the Internet to be processed manually. Responses used to be e:mailed back until the sheer amount of requests that Riderlink received snowballed to such a degree that the system was disappointingly scrapped. If anything, the popularity of such a travel information service was at least proven.

Similar methods are currently in use in the Nottingham Travelwise Centre and, presently, the level of requests being made can be handled. Any significant increase may see the service go the same way as Seattle's unless it can be automated. The Riderlink site was launched after being designed and put together by information experts and Internet designers, and also after many focus groups were held with the public to gain valuable input as to what they required from such an information service. It was yet another project where a serious service evaluation followed. The decision to use the Internet came after weighing up the advantages and disadvantages of doing so.

As it happened, the usage and availability of the Internet grew rapidly during the lifetime of the project. It was also identified that the advantages were increasing and the disadvantages decreasing. It also allowed the project to develop in ways not possible without the Internet. What began with a system providing access to electronic transportation information to eight employer sites evolved to include access throughout the Puget Sound region with no additional effort or cost. Another major plus for the Internet project occurred when trying to incorporate and maintain ferry schedule information. Such information was already being provided by the Washington State Department of Transportation (WSDOT), the organisation responsible for the actual ferry service. All that was required then was a link to this existing information source. Since the launch of Riderlink many other government and transportation Web sites have linked to it. This highlights the opportunities to integrate multi-modal transportation information from multiple sources and that it can often be a time and money saving exercise. There was no point in simply replicating the information, it was a bit like reinventing the wheel.

Riderlink demonstrates very well how to develop aspects of transport information services and how information sharing and linking is a powerful tool.

TRAC Riderlink Kiosks

The Riderlink project extended to deploying four kiosks equipped with personal computer hardware with ISDN (Integrated Services Digital Network - digital telephone based communications) Internet connections and touch screen monitors. A standard commercial Internet browser was installed to display the content and a laser printer installed to allow useful information to be available in hard copy form. Concerns from employers included the fact that employees now had access to the entire Internet and considered this as an undesirable feature of the new kiosks. This was potentially overcome by developing software to limit the computers that could be accessed i.e. limited to just the Riderlink information servers. This initially failed due to the industry of hackers at one particular site who, allegedly, managed to bypass this software and have free access to the Internet. The security and firewall settings needed to be enhanced so that access was not directly to the Internet service provider. Other features of the kiosks included an automatic return to the Home Page after five minutes.

The ISDN communication costs were, however, very expensive and paying for this full-time to obtain service connection by the minute was not deemed feasible. To overcome this Metro King County developed a different communication technique using newly developed Digiboard Datafire cards. These were installed in the personal computers, and Digiboard IMACs (ISDN bridged communications) installed in the server, enabling an instant connection to be invoked whenever a new user touched the screen but with an automatic timeout after a specified period of non use.

The cost of each kiosk totalled just under the $20,000 mark - not actually a cheap solution for such limited information outlets.

On the positive side, most users consider Riderlink to be well designed, efficient and informative. Negative comments have also been received, but were earmarked as potential developments of the project:

• A text-only version for users of non-graphical browsers
• Additional transportation information for counties adjacent to King County
• Automated trip planning tools where the user can enter an origin and destination and immediately obtain a recommended bus trip
• An on-line system map
• An on-line bike route map

Other comments were received, providing direct views and suggestions for system developments. Again, I have included such comments as they are certainly useful when evaluating and developing our potential new systems.

• "I was looking for information for my husband who has to start commuting to Boeing Everett and I didn't find anything helpful. The bus routes were not recognised by the system. This would be easy to use if you asked where the user wanted to go and where they wanted to leave from and then just gave them the options. The tech-cuteness makes this too unfriendly."
• "Although the web pages are informative, they are very difficult to use from a non-graphical browser. Good use of alt text on the images, and normal links to supplement the image maps, would greatly improve this service usability."
• "It would be useful to include in the area the e-mail and snailmail addresses of various Metro departments. If this information is here, have not found it yet."
• "I filled out the survey, giving high marks to everything, but I just bombed out trying to find one clear route between the Seattle Center area and the University. There were many listed routes, too many to make any sense of!!! Really, many solutions are much worse than one if you don't know how to get there in the first place. So, what is needed is a "point-and-click function" between any two points on the transit map. The system then figures out a likely best route (or 2) and returns that route map. This will give your programmers something to do that will really . . ."(text exceeded space allowed)
• "Finding the bus route needs to be easier for those who aren't familiar with the area."
• "The bus info is good, but I don't know where all the places listed as near my destination (work) are. I can't tell if it is close enough or not. Really useful bus info section would allow me to put you all out of work. I would enter the starting address, then destination address. The software would give me the closest stops, routes, etc. No wait. But his is loads better than no info at all!!! :-) Thanks."
• "Maps of bike routes, similar to the bus route maps, would be a big cool."
• "These pages would be more useful if set up for those without windows, and Those with a limited text base system. Other than that I'm glad they're here."
• "Could not figure out how to determine whether a certain trip involves one zone or two. Is that missing or did I miss it? Other than that this is a great service!"
• "Great service. You should investigate providing instant response for Matchlist requests."
• "Bike lane maps would be very useful."
• "It would be nice if you provided information on custom bus routes."
• "I'm hoping you are working toward an automated trip planning system, i.e., something that could take a departure and a destination address, then automatically work out the best way to get there by bus, and suggest the first bus to take."

There is much more to the Riderlink site but it is accessible via the Internet to view for yourselves at www.riderlink.gen.wa.us/rl_connect.html

My sincere thanks to Catherine Bradshaw Boon

TRAC & WSDOT

Marty Pietz is the Executive Director of the TRAC operation, as well as being the Director of the WSDOT Research Office. As an authority in the field of traveller information systems it was pleasing that he kindly agreed to a chat.

In terms of developing and launching traveller information systems there were two simple questions that could be put to a potential audience: "What do you want?" and "Will your travel habits change? (mode / time)". Based on the rich and varied responses to these WSDOT and TRAC have been striving to produce the ideal traveller information system.

Among the topics of conversation were future projects for TRAC & WSDOT, having already covered a plethora of current systems I asked what other high-tech information solutions were in the pipeline for Puget Sound.

Feedback already gleaned from online surveys and focus groups had identified a requirement for predicted travel times for specified journeys. Now TRAC had teamed up with staff at the nearby University of Washington to try and identify and produce a set of automatic traffic condition prediction algorithms. Daniel J. Dailey, of the University, was developing the predictive model using current and historical flow information together with current CCTV images. The target for the project is to predict as far as twenty minutes ahead to show what the state of the freeways would be. This would help commuters with travel time decisions, as well as the mode of transport chosen. The flow and CCTV data is fed to the University via Seattle's fibre optic ITS data backbone, which allows several organisations access to available travel data for various uses. Although an ambitious project, development is made possible due to the excellent partnerships with local Academia. Little or no neural net programming skills existed within TRAC but the project lent itself very well to utilising University expertise. An institution like the University of Washington is schooled in areas that can benefit ITS projects e.g. computer science, engineering, communications, institutional studies etc.

In terms of traffic and travel information (TTI), it is considered it a constant process of evaluation - demonstration - deployment - re-evaluation. These being the identified steps to ensure successful deployment and longevity of ITS schemes. There is no doubt that TTI is popular when people get to know about it and use it. It was also considered that administrative barriers don't hinder as the road, rail and sea networks don't all end at County boundaries. The ultimate development is identified as a U.S. wide traveller information service.

In terms of the information: the traditional incident, congestion, roadworks information should be backed up with advisory information for special events such as ball games etc. Information can then be generated on how to get to an event as well as how to avoid it. Current developments include a plan for more comprehensive traffic advisory information for such events to include routes, public transport, pickup points and recommended diversions. Also advocated is the current and future use of WAP phones, palm tops and cell phones - all convenient devices for information. The idea being to maximise the audience for TTI. Roadworks and incidents can easily be text forwarded to these devices and they are ever evolving as products which will then enable more sophisticated information and automatic alerts to be forwarded to them.

Again, there was a strong opinion that the Internet was at the core of current information provision and was certainly not going to disappear. It seems more likely that new applications be built into or around Internet technologies and services. Further, that the wider the area for the information then the more useful it will be to a wider audience and not just the 3 million inhabitants of the four County area (King, Pierce, Snohomish, Kitsap). WSDOT had a lot of information systems running prior to the take off of the Internet. The momentum gained enabled them to exploit the Internet quickly when it became clear that it was going to be a mass end user tool. Demonstrations of various systems already existed anyway so a good platform was in place as technology developed, the congestion map being a prime example of this.

My sincere thanks to Marty Pietz