Helsinki Arena 2000
-GIS for everybody
Risto Linturi, MSc
Multimedia Technology Director
Helsinki Telephone
GIS for everybody - it sounds great, but GIS is not now a popular term like internet, virtual reality or electronic shopping. This can be changed because GIS affects our lives in many different ways.
Let´s define GIS a bit unorthodoxly to see what happens. DEFINITION: GIS is for us a way to collect, organize and view information based on its geographic coordinates. The abovementioned information can correlate to the coordinates by physical or conceptual way.
We used to collect GIS information by walking, looking or feeling and then remembering that information. We also altered the physical environment to include conceptual border. Now we use different mechanical and electronic means to store, retrieve, convert and distribute that information when and where needed. We also use lots of symbols to express information content in an abstract manner to avoid lengthy explanations or limitations of our equipment. This has become awkwardly complex and GIS for everybody must mean simpler things.
Era of complex notations and difficult retrieving of information may however be passing. Edp experts, science fiction fans and many internet advocates have started to talk about virtual reality as a serious possibility. I personally fall into each three categories.
Most professional GIS people tend to categorize all graphical representations as mere visualizations of the GIS data. This might lead to serious underestimation of the nature and possibilities of virtual reality technologies.
Networked virtual reality is not a mere visualization. It is also a way to collect and organize geographic data. Besides it supports by its nature both real time GIS and past, history and future planning information.
Virtual reality houses also phenomena existing only in electronic form. These could be claimed to fall outside GIS. I however recommend to do otherwise. In the future virtual and real environments mix more and more. If users use geographic information as a backgroud and orientation point then this should be considered GIS even though part or all of the considered things would happen in electronic form inside the computers only. There are many other contractual or otherwise abstract things that we place on our maps which are considered GIS as well though not physical in nature.
Advances in mass market technology
Internet has popularized a language for describing three dimensional models. This language is called VRML (Virtual Reality Modelling Language). VRML is very quickly gaining critical mass and becoming a real mass market standard for three dimensional environments.
VRML-model can be layered and it can include links to other models or other elements of internet content. Common spesification for VRML has been accepted by nearly all members of the software industry. Current finalized version is VRML 2.0 <1>. This includes many features that help in creating realistic and versatile virtual reality models for various purposes.
VRML is not a simple 3d model. It facilitates layering and linking system where a user can be transferred to another model or a user can interact with agents that start information flow to and from some other program or computer in the network. VRML should thus not be considered as a mere 3d-visualization model but as a linked information system where information flow is defined directly as a three dimensional and distributed VRML-database. VRML could be considered as a standard file format for 3d mapping and linking.
VRML viewers are usually distributed freely and can be used with popular internet viewers such as Netscape Navigator or Microsoft Internet Explorer. Some viewer developers have created enhancements so that users are represented by animated avatars and can interact with other users of the same model.
Networked virtual model can be used for telepresence where the viewer is connected to one or two way video communication thus being able to view the specified geographical area directly in real time. Besides video, telepresence can be achieved by animated avatars as I mentioned earlier. Avatars or other animations can be also used to supply GIS -information. One example being simulated weather and animated measurement devices linking information flow from the real world to the virtual reality.
Avatars require only modest bandwidth and can be implemented realistically in internet. Satisfactory video communication requires guaranteed bandwidths from ISDN:s 128 kb/s to DLS-modems 2Mb/s <2>,<3>.
Helsinki Arena 2000 -project
Helsinki Telephone Company and the City of Helsinki have started a project to create a virtual reality model of the city and its surroundings. This VRML model will be used as the user interface and environment for local electronic services such as video commerce and other electronic shopping, multipoint video telephony, cultural services, other public services, entertainment, virtual meetings, and viewing the city and its public places in real time. Many forms of other GIS information will also be included in the model for both professional and popular purposes. The project has gathered wide support including all seven regional art and science universities, largest edp-companies and other large corporates.
Virtual Helsinki will be networked virtual reality where users can interact. The project is based on internet standards but internet itself has to be locally enhanced to facilitate high quality for virtual reality and multimedia interaction. Internet by itself has no means to guarantee either response time or bandwidth and thus quality of real time video telephony or audio interaction is not acceptable without enhancements. Users need to connect directly to Helsinki Arena 2000.
The Helsinki Arena 2000 -network will be a combination of fiber and copper lines. Home and small office users are connected via 128 kb/s ISDN-lines or 2 Mb/s DSL-lines to the nearest telephone switch. Service providers, gateway providers to other networks and corporate users are connected with ATM-lines.
Users can navigate the City model and activate links which connect him to service providers equipment or other users PC:s and regular telephone conversations. The model is distributed so that only external model of the city and limited amount of video streams viewing the public places of the city are provided centrally. Each service provider or individual user will supply the necessary models when user gets inside their buildings. No constraints are enforced for the interior models and so they can either follow industry standards and resemblance to the physical world or be very experimental in both technical requirements and contentwise.
First results from project experiments
The project is in its starting phase but initial feasibility studies have yielded some results already in many project areas.
Based on expected decrease in transmission equipment we have calculated that 2 Mb/s connections can be provided to households before year 2000 at reasonable costs. Based on expected increase in processing power and video capabilities we expect that ISDN-level video conferencing equipment will be commonplace and near tv-quality receiving will be possible with equipment in household affordable prices at the same timeframe. Based on world wide software investments and market behaviour we expect that VRML and its PC-based viewers will even earlier be developed into efficient and suitable tools for modelling virtual worlds and meeting points.
Our first experiments show that there seem to be many architect tools suitable for modelling a virtual city but in practice the requirements are very specific and most of the proposed alternatives have been ruled out.
Fully automatic means such as stereo photography or 3d digitizing equipment are unsuitable for this purpose because there is a need to minimize information content in the model and at present this requires human desicion making in each typical picture element. Knowledge or intuition is needed so that resemblance could be achieved with minimum amount of unnecessary details. This is required so that the virtual reality environment could be navigated with mass market equipment in real time.
Basic VRML-tools are ruled out because their production level is too low. Even usual architect tools were ruled out for the same reason. When a whole city is modeled the tools need to be very productive and suited for the spesific task. They need to include buildings as objects and building constructs as object libraries. VRML should be generated in several levels of detail from the objects and links should be automatically formed and updated to the VRML-model when and as needed.
Main criteria in Helsinki Arena 2000 is to create a familiar setting which is close enough to reality to support visual navigation and close enough to support storage and linking of various GIS information with modest requirements. A suitable tool was found and a large scale test was performed to check the production speeds and quality of the resulting models. The tool used was originally designed and used for city planning and it is Autocad based KCAD- from Arcus Software. KCAD will be developed into an environment planning software according to conversations with its chief designer Immo Teperi.
Experience from the pilot project showed that most buildings could be constructed very quickly from the object libraries. Building mass, roof model, colour and window typing were routinely handeled and some spesific features were easily added when needed. Spesific landmarks were created from more primitive objects and thus needed more work.
Creation of the piloting VRML-model was a success but first utilization experiments succeeded only partially. A large fair environment with many pavillions and other coordinate based content and communication facilities was widely distributed and used in internet using VRML-technology. Test was successful and showed that versatile services can be implemented in closed 3d models using todays technology.
The open environment city model is however too complex for todays computers as was originally expected. Present VRML-viewers are in addition early versions and don´t support all features required by the spesifications. Especially harmful was lack of support for the level of detail (LOD)-feature which is absolute requirement for large open views. To get a complete city model to work in PC:s we need this corrected and some speed and efficiency enhancements in both PC:s, their peripherals, system software and VRML viewers. These enhancements combined are expected to yield efficiency gains in the order of two magnitudes by the year 2000 which would make Virtual Helsinki very quickly a pleasant and easy environment for the home PC-users. For the GIS professionals it will provide enormous possibilities for popularizing GIS and its applications. Virtual worlds are not visualizations or user interfaces - they are worlds by themselves even when they mix and match telepresence, imaginary avatars or real world information. Updating and creating geographic information will be more important than ever as access to GIS information will become commonplace and easy.
[1] ISO/IEC CD 14772 The Virtual Reality Modeling Language
Specification, version 2.0. ISO/IEC, 1996.
[2] Männisto E. (ed.), Clarke J., Defee I., Murphy P., Sauvala S.
Report of RACE DIAMOND video-on-demand field trials in Helsinki and Sligo:
“Evaluation Report”. RACE R2105 DIAMOND Deliverable 35, 1996.
[3] Poikolainen P. (ed.), van de Haar P.G., Koponen M., Varonen T.
Report of RACE DIAMOND video-on-demand field trial in Helsinki:
“Prepared Network and Service Ready for Trialing”. RACE R2105 DIAMOND
Deliverable 30, 1995.