Wireless Networks 3 (1997) 421–433
`
`421
`
`Cyberguide: A mobile context-aware tour guide
`
`Gregory D. Abowd a, Christopher G. Atkeson a, Jason Hong a, Sue Long a,b, Rob Kooper a and Mike Pinkerton a
`a Graphics, Visualization and Usability Centre, College of Computing, Georgia Institute of Technology, Atlanta, GA 30332-0280,USA
`b Wink Communications, Alameda, CA 94501, USA
`
`Future computing environments will free the user from the constraints of the desktop. Applications for a mobile environment should
`take advantage of contextual information, such as position, to offer greater services to the user. In this paper, we present the Cyberguide
`project, in which we are building prototypes of a mobile context-aware tour guide. Knowledge of the user’s current location, as well as a
`history of past locations, are used to provide more of the kind of services that we come to expect from a real tour guide. We describe the
`architecture and features of a variety of Cyberguide prototypes developed for indoor and outdoor use on a number of different hand-held
`platforms. We also discuss the general research issues that have emerged in our context-aware applications development in a mobile
`environment.
`
`1. Introduction
`
`Future computing environments promise to free the user
`from the constraints of stationary desktop computing, yet
`relatively few researchers are investigating what applica-
`tions maximally benefit from mobility. Current use of mo-
`bile technology shows a slow evolution from our current
`desktop paradigm of computing, but the history of inter-
`action shows that the adoption of new technology usually
`brings about a radical revolution in the way humans use
`and view technology [11]. Whereas the effective use of
`mobile technology will give rise to an interaction para-
`digm shift, it is difficult to predict what that shift will be.
`We follow the advice of Alan Kay, therefore, and choose
`to predict the future by inventing it. Our approach is to
`think first about what activities could be best supported
`by mobile technology and then determine how the tech-
`nology would have to work. This applications focus is
`important to distinguishing our work in mobile comput-
`ing.
`In April 1995, we formed the Future Computing Envi-
`ronments (FCE) Group within the College of Computing
`and the Graphics, Visualization and Usability (GVU Cen-
`ter) at Georgia Tech to promote such an applications focus.
`Our group is committed to the rapid prototyping of appli-
`cations that benefit from the use of emerging mobile and
`ubiquitous computing technologies. Quick development of
`these futuristic applications allows us to predict and shape
`what our everyday lives will be like when today’s novel
`technology becomes commonplace.
`Applications for a mobile environment should take ad-
`vantage of contextual information, such as position, to of-
`fer greater services to the user. In this paper, we present
`the Cyberguide project, a series of prototypes of a mobile,
`hand-held context-aware tour guide. Initially, we are con-
`cerned with only a small part of the user’s context, specif-
`ically location and orientation. Knowledge of the user’s
`current location, as well as a history of past locations, are
`used to provide more of the kind of services that we come
`
`J.C. Baltzer AG, Science Publishers
`
`to expect from a real tour guide. We describe the archi-
`tecture and features of a variety of Cyberguide prototypes
`developed for indoor and outdoor use on a number of dif-
`ferent hand-held platforms. We also discuss the general
`research issues that have emerged in our experience of de-
`veloping context-aware applications in a mobile environ-
`ment. Some of these research issues overlap with those
`that we have considered in applying other applications of
`ubiquitous computing technology.
`The general application domain which has driven the
`development of Cyberguide is tourism, but we have found
`it necessary to be even more focused in our research. The
`initial prototypes of Cyberguide, therefore, were designed
`to assist a very specific kind of tourist – a visitor in a tour
`of the GVU Center Lab during our monthly open houses.
`Visitors to a GVU open house are typically given a map
`of the various labs and an information packet describing
`all of the projects that are being demonstrated at various
`sites. Moving all of the paper-based information into a
`hand-held, position-aware unit provided a testbed for re-
`search questions on mobile, context-aware application de-
`velopment.
`The long-term goal is an application that knows where
`the tourist is, what she is looking at, can predict and answer
`questions she might pose, and provide the ability to inter-
`act with other people and the environment. Our short-term
`goal was to prototype versions of Cyberguide on commer-
`cially available PDAs and pen-based PCs in which context-
`awareness simply meant the current physical position and
`orientation of the Cyberguide unit (and since it is hand-held,
`this locates the user as well). Position information improves
`the utility of a tour guide application. As the prototypes of
`Cyberguide evolve, we have been able to handle more of
`the user’s context, such as where she and others have been,
`and we have increased the amount in which the tourist can
`interact and communicate with the place and people she is
`visiting.
`
`IPR2020-00408
`Apple EX1005 Page 1
`
`(cid:211)
`

`

`422
`
`G.D. Abowd et al. / Cyberguide: A mobile context-aware tour guide
`
`1.1. Overview
`
`This paper is an extended version of an earlier report on
`Cyberguide [7], we discuss the evolution of the Cyberguide
`design and prototype as well as what future research areas
`our experience has uncovered. We begin in section 2 by
`describing scenarios for the use of context-aware mobile
`applications. In section 3, we provide context for our re-
`search within the area of applications-centered mobile com-
`puting. The generic architecture of Cyberguide is explained
`in section 4. We will describe in section 5 the initial re-
`alization of the generic components of the Cyberguide ar-
`chitecture, a series of prototypes developed for the Apple
`MessagePad. We will then describe in section 6 how the
`initial indoor prototypes were extended for use outdoors
`and for greater interaction with the environment. We con-
`clude in sections 7 and 8 with a discussion of significant
`issues for context-aware applications development and how
`our past experience will influence our future development
`plans.
`
`2. Scenarios for a mobile context-aware application
`
`This section outlines some possible uses for future mo-
`bile context-aware applications. Some of these uses are
`currently being implemented and some are futuristic. We
`begin with our initial assumptions about what technology
`we expect Cyberguide to use. Tourists are usually quite
`happy to carry around a book that describes the location
`they are visiting, so a reasonable packaging would be in
`the form of a hand-held device. The ideal hand-held de-
`vice will have a screen and pen/finger interface, access to
`substantial storage resources – possibly through an internal
`device such as a CD drive, or through substantial commu-
`nication and networking resources (cell phone, pager, data
`radio interface) providing access to other storage servers
`(such as the Web) – an audio input and output interface
`with speech generation and potentially sophisticated voice
`recognition, and a video input and output interface. The
`video input (a video camera) could be pointed at the user
`to interpret user gestures, or pointed at the environment to
`interpret objects or symbols in the environment. The video
`output could be integrated into the main screen or be a sep-
`arate video display device, such as an attached screen or
`heads up display on glasses worn by the user.
`One major application of mobile context-aware devices
`are personal guides. Museums could provide these devices
`and allow users to take personalized tours seeing any ex-
`hibits desired in any order, in contrast to today’s taped tours.
`In fact, many museums now provide portable devices for
`just such a purpose, but what we are envisioning is a device
`that would allow the tourist to go anywhere she pleases and
`be able to receive information about anywhere she is. Walk-
`ing tours of cities or historical sites could be assisted by
`these electronic guidebooks. The hand-held devices could
`use position measurement systems such as indoor beacons
`
`or the Global Positioning System (GPS) to locate the user,
`and an electronic compass or inertial navigation system to
`find user orientation. Objects of interest could be marked
`with visual markers or active beacons or recognized us-
`ing computer vision. Some objects, such as animals at
`a zoo or aquarium, might be difficult to mark but could
`be recognized with simple computer vision and some as-
`sistance from the environment (indications that this is the
`elephant cage, for example). The personal guide could also
`assist in route planning and providing directions. Some of
`these functions are currently being provided by automobile
`on-board navigation systems.
`There are other ways to assist users. Consider a traveler
`in Japan that does not speak or read Japanese. The hand-
`held device could act as a pocket multilingual dictionary,
`actually speaking the appropriate phrase with the appro-
`priate pronunciation to a taxi driver, for example (or even
`showing the appropriate Kanji and an associated map on
`the screen). A device that included video input or a scan-
`ner could assist in reading signs or menus. A device that
`could show stored images might be able to show a shop-
`keeper the desired object or favorite meal. Another more
`futuristic use is to assist the user by recognizing faces at a
`cocktail party and reminding the user who people are.
`Real-time communication allows a personal device to
`act as an agent for the user. A personal guide to a theme
`park could make reservations at particular rides, and alert
`the user when the reservation was available. The device
`could also tell the user which rides had the shortest lines.
`Similar approaches are currently being used for automobile
`traffic management in major cities.
`An important application of context-aware devices is en-
`hanced reality. A heads up display could provide “X-ray”
`vision for the user. While surveying a building for renova-
`tion, the location of hidden plumbing or electrical conduits
`could be indicated to the user, based on information from
`sensors and/or building plans. At an archeological site a
`visitor could be provided with various overlays indicating
`what used to be above the current ground level as well as
`what is below the current ground level.
`Context-aware devices can also be used as tools. Simple
`sonar devices are used to make room measurements today.
`It would not take much to have a hand-held device that
`both videotaped and mapped a room along with user com-
`mentary. An ecological field study or an archeological dig
`could be assisted by a device that automatically recorded
`the context of a particular find, including noting the sur-
`rounding objects. Consider an electronic field guide that
`assisted the user in recognizing plants or insects.
`One of the most interesting applications of context-aware
`devices is to support group interaction on a tour or in a
`classroom, for example. Participants in a live demonstra-
`tion of some new technology could use their personal device
`to help steer the demo using majority voting or consensus
`among the viewers. Each participant could run a person-
`alized version of the same demo by expressing their own
`choices. In this case context is which demo a participant
`
`IPR2020-00408
`Apple EX1005 Page 2
`
`

`

`G.D. Abowd et al. / Cyberguide: A mobile context-aware tour guide
`
`423
`
`is participating in or attending to, and the personal ma-
`chine may switch to another context if it detects the user is
`attending to that context instead.
`Many tourists take records of some sort of their travel-
`ling experiences, either by taking pictures or videos or by
`composing a travel diary. Imagine the possibilities if the
`recording of these experiences could be more efficiently
`and accurately recorded. A drive across the country could
`result in a trail superimposed upon a map, and clicking on
`the trail would reveal an image of what you could see at
`that moment – an automatically-generated spatial index into
`your memories.
`These are but a few of the possibilities we can imagine
`that a context-aware application can provide for the tourist.
`We have investigated many of these possibilities already
`and report on them later.
`
`3. Related work
`
`In thinking about and developing a location-aware ap-
`plication, we were greatly influenced by work such as the
`PARCTab at Xerox PARC [15],
`the InfoPad project at
`Berkeley [8], the Olivetti Active Badge system [14] and
`the Personal Shopping Assistant proposed at AT&T [3].
`We wanted to build useful applications that might take ad-
`vantage of the hardware developed in the PARCTab and
`InfoPad projects. We did not want to build our own hard-
`ware, so we have a different focus from all of these projects.
`There are a number of commercially available and relatively
`inexpensive hand-held units that would suffice for our pur-
`poses, such as the Apple MessagePad with the Newton
`Operating System1, a MagicCap2 machine or a pen-based
`palmtop/tablet PC. We chose to work initially with the Ap-
`ple MessagePad 100 with Newton 1.3 and pen-based PCs
`running Windows for Pen Computing 1.0. Each platform
`was available for $150–500 with relatively powerful de-
`velopment environments. This low cost of hardware was
`critical to the success of Cyberguide because it made it
`possible to put a number of units in the hands of many
`students, all with different ideas that they were allowed to
`investigate.
`For positioning, we considered the Active Badge system,
`but rejected it for reasons of cost and long-term objectives.
`The Active Badge system combines position detection with
`communication. For room-level granularity of position, this
`is reasonable since the communications range is on par
`with the position resolution. With Cyberguide, it is not
`clear that positioning and communication systems should
`always share physical resources. Certain versions of our
`prototype did; other prototypes did not. We provided for
`the separation of the wireless communications capabilities
`from the positioning system, so we could seek out more
`cost-effective solutions for both.
`
`1 MessagePad and the Newton Operating System are registered trademarks
`of Apple Computer, Inc.
`2 MagicCap is a registered trademark of General Magic, Inc.
`
`We tried to pay attention to the higher level concep-
`tual design of Cyberguide, but we have not been as gen-
`eral in our handling of context-aware mobile objects as has
`Schilit [13].
`
`4. Architecture of Cyberguide
`
`From the beginning, we have viewed Cyberguide as a
`family of prototypes and not just a single prototype, so it
`made sense to think about a conceptual design, or archi-
`tecture, that captured the essence of the mobile tour guide.
`We have divided the system into several independent com-
`ponents, or building, and have found it useful to present
`those components both in terms of the generic function and
`personified in terms of the people a tourist would like to
`have available while exploring unfamiliar territory. The
`overall system serves as a tour guide, but we can think of
`a tour guide as playing the role of cartographer, librarian,
`navigator and messenger. The services provided by these
`components are:
`(cid:15) Cartographer (map component). This person has in-
`timate knowledge of the physical surroundings, such
`as the location of buildings, interesting sights within
`a building, or pathways that the tourist can access. This
`component is realized in our systems by a map (or maps)
`of the physical environments that the tourist is visiting.
`(cid:15) Librarian (information component). This person pro-
`vides access to all of the information about sights that a
`tourist might encounter during their visit. This would in-
`clude descriptions of buildings or other interesting sights
`and the identities of people associated with the areas.
`The librarian can answer specific question about certain
`sights (“Who works in that building?” or “What artist
`painted that picture?” or “What other demonstrations
`are related to what I am looking at?”). This component
`is realized as a structured repository of information re-
`lating to objects and people of interest in the physical
`world.
`(cid:15) Navigator (positioning component). The interests of
`the tourist lie relatively close to their physical location.
`Therefore, it is important to know exactly where the
`tourist is, in order to show the immediate surroundings
`on the map or answer questions about those surround-
`ings (“What am I looking at?”). The navigator is respon-
`sible for charting the location of the tourist within the
`physical surroundings. This component is realized by
`a positioning module that delivers accurate information
`on tourist location and orientation.
`(cid:15) Messenger (communications component). A tourist will
`want to send and receive information, and so the mes-
`senger provides a delivery service. For example, when
`visiting an exhibit or demonstration, the tourist might
`want to speak with the owner of the exhibit.
`If the
`owner is not present, the tourist can leave a message. In
`order to find out where other tourists are located, each
`
`IPR2020-00408
`Apple EX1005 Page 3
`
`

`

`424
`
`G.D. Abowd et al. / Cyberguide: A mobile context-aware tour guide
`
`Figure 1. The map (left) and information (right) interfaces of the initial Cyberguide MessagePad prototype.
`
`tourist can communicate her current location to some
`central service that others can access. It might also be
`desirable to broadcast information to a set of tourists
`(“The bus will be leaving from the departure point in
`15 minutes.”). This component is realized as a set of
`(wireless) communications services.
`The utility of this architectural decomposition for Cyber-
`guide is that it provides an extensible and modular approach
`to system development.
`It is extensible because we can
`always add further services. For example, we have con-
`sidered adding an historian whose purpose is to document
`where the tourist has been and what her reactions were to
`the things she saw.
`It is modular because it has allowed
`us to change the implementation of one component of the
`system with minimal impact on the rest of the system. For
`example, we have implemented different versions of the
`navigator and the librarian without having to alter the other
`components. Of course, these components are related in
`some ways; for instance, position information ultimately
`has to be translated into a location on the physical map.
`Defining standard interfaces between the components is the
`means by which we achieve separation between and coor-
`dination among the various components.
`
`5. The indoor Cyberguide
`
`In this section, we describe how each of the separate
`modules in the conceptual architecture have been realized
`in the initial series of prototypes developed on the Apple
`MessagePad for use indoors during GVU open houses.
`
`5.1. Map component
`
`The initial map module, shown on the left side of fig-
`ure 1, contains a map of the entire GVU Center. Pas-
`sageways and demonstration stations (stars in figure 1) are
`
`shown. Only a limited view of the lab can be seen at any
`given time. The user can scroll the map around and zoom
`in and out to see alternative views. There is an icon to
`show the user’s location on the map. Using information
`from the positioning module, we implemented automatic
`scrolling of the map. If desired, the user’s position is up-
`dated automatically and the map is scrolled to ensure that
`the user’s current position remains on the visible portion of
`the map.
`
`5.2. Information component
`
`The information module (shown on the right side of fig-
`ure 1) contains information about each of the demos on
`display at the GVU open house. This includes abstracts of
`the project being demoed, background information on those
`involved with the project, as well as where to get further
`information. The location of each demo is marked on the
`map by a star. The user selects the star icon for a demo
`to reveal its name. Selecting the name brings up the infor-
`mation page for that demo. The user can also go directly
`into the information module and search for information for
`specific demo pages either by category or by project name.
`One version of the information module was hard-coded,
`providing very fast response but requiring a recompilation
`every time demo information needed to be updated. An-
`other implementation used Newton files, called soups, to
`store information. The use of soups avoided hard-coding
`data into the application and simplified demo information
`updates, but did not have adequate response time. Our
`third implementation of the information module used New-
`ton Books, the Newton platform documentation viewer, to
`store the demo information. The use of Newton Books
`improved our access time considerably, allowing for an au-
`tomated information update process without requiring data
`be hard-coded directly into the application. Throughout all
`
`IPR2020-00408
`Apple EX1005 Page 4
`
`

`

`G.D. Abowd et al. / Cyberguide: A mobile context-aware tour guide
`
`425
`
`(Motorola 68332) interfaced the IR receiver to the serial
`port of the Apple MessagePad. We deployed an array of
`remote controls hanging from the ceiling (figure 3 right),
`each remote control acting as a position beacon by repeat-
`edly beaming out a unique pattern. The 68332 translates
`the IR pattern into a unique cell identifier that is sent to
`the Apple MessagePad’s serial port. As the tourist moves
`around the room and passes into the range of a new cell, the
`position (indicated by an arrowhead) is updated on the map.
`Keeping track of the last recorded cell location provides a
`good guess as to the location the tourist is heading, so we
`indicate an assumed orientation by pointing the position
`icon accordingly.
`The remote control system is too expensive for large
`scale use as the cost of the 68332 microcontroller is roughly
`equivalent to that of the MessagePad.
`
`Figure 2. Questionnaire using communications module for delivery.
`
`6. Extending the initial prototype
`
`three versions of our information module, we were able to
`modify the information module independent of the develop-
`ment efforts of the other modules, validating the modularity
`of part of our design.
`
`5.3. Communication component
`
`Our initial implementation of a communication module
`consisted of a wired Internet Appletalk connection from
`a Apple MessagePad through a Unix Appletalk Gateway.
`We designed an application level protocol on top of a pub-
`lic domain implementation of the Appletalk protocol for
`Solaris [4]. This allows us to open a connection-based Ap-
`pletalk stream from the Apple MessagePad to a UNIX plat-
`form. We then invoked our gateway application to repacke-
`tize Appletalk packets into TCP/IP packets for transmission
`over the Internet. This allowed for TCP/IP connectivity
`from a Apple MessagePad via an Appletalk connection.
`We could then fetch HTML documents as well as send and
`receive e-mail. We utilized this functionality within Cyber-
`guide by providing a questionnaire for users to complete,
`which was sent to the developers as an e-mail message (see
`figure 2).
`
`5.4. Position component
`
`Position is the obvious starting point for a context-aware
`mobile device. We considered several methods for sens-
`ing the user location. Outdoors, continuous services, such
`as GPS, can be used. Indoors, however, GPS signals are
`weak or not available. We considered RF for indoor po-
`sition measurement, but found off the shelf solutions too
`expensive.
`One solution for an indoor positioning system was to
`use infrared (IR). Our first positioning system was based
`on using TV remote control units as active beacons, and
`using a special IR receiver tuned to the carrier frequency
`((cid:25) 40kHz) of those beacons (figure 3). A microcontroller
`
`The first Cyberguide prototypes were completed within
`6 months. To test out the genericity of our architectural
`approach, we decided to develop further prototypes that
`altered one or more of the major components described in
`section 4 and increased overall functionality. We describe
`these extended prototypes here.
`
`6.1. Outdoor positioning
`
`There were several motivations for building a Cyber-
`guide prototype for outdoor use (figure 4). First, we wanted
`to use Cyberguide over a wider area than the relatively
`small GVU Center. We also wanted to test the modularity
`of our design by having to change critical features. The
`two features that were changed on this prototype were the
`underlying map and the physical positioning system. We
`obtained a different map and inserted that into the map mod-
`ule without any problems. For positioning, we replaced the
`IR positioning module with a Trimble GPS unit attached
`to the Apple MessagePad serial port.
`(see right side of
`figure 4). The GPS unit sends a position in latitude and
`longitude which was then translated into a pixel coordinate
`representing the user’s current position on the map.
`The outdoor positioning system has been tested by two
`prototypes. We first built a proof of concept tour of the
`Georgia Tech campus (shown in figure 4). We also devel-
`oped a more functional outdoor prototype that covered three
`surrounding neighborhoods of the campus, described later.
`
`6.2. Alternate platforms
`
`In order to verify the platform independence of our con-
`ceptual design, we initiated two separate efforts building
`pen-based PC versions of Cyberguide. These limited func-
`tionality PC versions were written using Borland’s Delphi
`environment and Microsoft’s Visual Basic. Both were ini-
`tially installed on Dauphin DTR-1 palmtops running Pen
`for Windows Computing 1.0.
`
`IPR2020-00408
`Apple EX1005 Page 5
`
`

`

`426
`
`G.D. Abowd et al. / Cyberguide: A mobile context-aware tour guide
`
`Figure 3. IR positioning prototype (left) and the array of positioning beacons in the GVU Lab (right).
`
`Figure 4. The outdoor Cyberguide (left) with GPS unit (right).
`
`The Delphi version implements the map and informa-
`tion module. Web pages containing demo information are
`stored locally as database objects using a stand-alone Bor-
`land database engine. The information base is extracted
`from the collection of existing Web pages for GVU projects
`but stored locally. This information is viewed using a pub-
`lic domain Delphi HTML viewer. Though this provided a
`very fast response for information queries, it is a long-term
`disadvantage to have the information base stored locally.
`Too much in our environment is subject to change. A lo-
`cal and static database is only slightly more useful than a
`book. This approach to static information storage is used
`currently for on-board navigation systems on certain rental
`cars.
`
`The Delphi prototype uses vector-based maps, allowing
`for arbitrary scaling and rotation of the map and well as
`path generation. While there are several sources for ob-
`taining vector maps for outdoor regions, it is not so easy
`for indoors. Consequently, there is a trade-off between the
`easily generated but limited functionality of bitmap images
`and the highly functional but hard to generate vector maps
`for indoor use.
`The Visual Basic prototype, shown in figures 5–9 real-
`izes all four components of Cyberguide, including two-way
`communications, which is discussed next. We implemented
`historical context by predicting when a user had visited a
`demo, based on time spent in the area of the demo and
`interaction with the map. In figure 5, a visited demo is in-
`
`IPR2020-00408
`Apple EX1005 Page 6
`
`

`

`G.D. Abowd et al. / Cyberguide: A mobile context-aware tour guide
`
`427
`
`Figure 5. The main map interface of the PC Cyberguide. Checks on demo sites indicate the user has been to visit that demo already, indicating history-
`sensitive interface.
`
`Figure 6. The information browser interface of the PC Cyberguide.
`
`dicated on the map by a checkmark. There is also a separate
`panel that lists the demos visited. This information could
`be used, for example, to generate a summary of the day’s
`visit to GVU open house and then mailed off to the visitor.
`We again made use of a publicly available HTML rendering
`component to display the project descriptions, still stored
`locally.
`
`6.3. Increased communication
`
`A number of interesting possibilities are enabled for the
`tourist in a wireless communication mode. We have spent
`a good deal of effort building an indoor, low-cost wireless
`communications infrastructure for Cyberguide. We have
`built a serial IR network using inexpensive modules from
`
`Sharp (the same modules used in the MessagePad). We
`have written UNIX server software and client software for
`both the MessagePad and PC. Figure 7 shows how our
`homemade network connects mobile units to the departmen-
`tal network. We have defined a simple protocol to support
`three different kinds of messages: mailing out from a mo-
`bile unit to the network (as shown in figure 8); broadcasting
`from the network to all mobile units (figure 9); and updat-
`ing positioning information. We implemented this protocol
`over serial IR instead of some other standard so that we
`could immediately use all platforms (UNIX, Newton and
`Windows). Given the appropriate hardware and protocol
`support (perhaps IRDA), we sencould provide the same
`functionality more robustly.
`
`IPR2020-00408
`Apple EX1005 Page 7
`
`

`

`428
`
`G.D. Abowd et al. / Cyberguide: A mobile context-aware tour guide
`
`Figure 7. Home-made IR units allow cross-platform communication.
`
`Figure 8. The mail interface of the PC Cyberguide. Addresses are automatically filled in and message templates are supplied upon request.
`
`6.4. Increased interaction with environment
`
`All of our applications of Cyberguide so far have re-
`stricted the role of the tourist to browsing, but it is likely
`that as she visits some place, the tourist will want to keep
`a record of her experience as advice for herself or others
`later on. With this idea of increased interaction with the
`environment and recording in mind, we created another Cy-
`berguide prototype, called CyBARguide, to assist a tourist
`in pursuit of refreshment at neighborhood establishments
`in Atlanta. This prototype covers approximately 12 square
`miles of midtown Atlanta, using multiple maps at varying
`levels of detail.
`Figure 10 shows the map interface on the left and a view
`of a user-modifiable database for interesting establishments
`on the right. The tourist can indicate a desired destina-
`tion and as she moves around, CyBARguide automatically
`chooses the map of the highest detail that contains both
`
`the traveler (indicated by a triangle in figure 10), and the
`destination (the beer mug with the emboldened border in
`figure 10). Along the way, if the tourist eyes another in-
`teresting establishment that is not currently highlighted on
`the map, it could be added. Each establishment has a user-
`modifiable database entry associated with it that reflects
`both objective (e.g., availability of parking, average price
`of drinks) and subjective (e.g., ambiance or other com-
`ments) information that can be used in the future to plan an
`evening’s excursion. querying of a large amount of infor-
`mation and some minimal routing facilities. We also plan
`to make the data within the information module modifiable
`so the user can add personalized information including per-
`sonal impressions that may be useful for future reference,
`a type of virtual graffiti. We envision the use of CyBAR-
`guide in a mode in which both tourist and proprietors are
`able to modify the information base. A simple searching
`
`IPR2020-00408
`Apple EX1005 Page 8
`
`

`

`G.D. Abowd et al. / Cyberguide: A mobile context-aware tour guide
`
`429
`
`Figure 9. A demonstration of receiving a broadcast message from the Network to an individual Cyberguide unit.
`
`Figure 10. The CyBARguide interface. The left shows the interactive map indicating the user’s location (the triangle) and the location of establishments
`previously visited (the beer mugs). The user modifiable database shown on the right supports the long-term development of touring information for a
`location.
`
`capability can assist revellers in search of a certain kind of
`entertainment experience, and additional contextual infor-
`mation, such as the knowledge of where the traveller has
`been already, the time of day, and what special events (in-
`formation provided by the proprietors) are currently sched-
`uled, can be used to deliver suggestions for where to go.
`
`7. Issues
`
`Our experience over the last year developing versions
`of Cyberguide with different features on different platforms
`has given us a certa