1. Introduction

1.1 Navigational Aids

International travel has become easier and more cost efficient over the last couple of decades (Taylor 2002). The Internet is also creating a more unified global economy, increasing the necessity for effective ways to aid travelers, both navigationally and culturally (Stea et al. 1996). There are currently many in-the-field GISs and navigation aids, including:

• Guidebooks one of the more comprehensive information providers of the in-the-field GISs. They contain maps, spatial, cultural and historical descriptions (Simonis 2002). Descriptions in written guidebooks are limited, as they typically supply only one or two routes between objects of interest. Guidebooks are useful for people spending a short period of time in an environment. They provide brief and concise descriptions that are stationary and cannot offer users in-depth information.

• Signage is one of the best ways to direct people to a location while minimizing their cognitive load. Signage works best if located at every decision point (Raubal et al. 1997). Signage also needs to be correct, easy-to-find, and easy-to-read. Not only does signage need to help people make decisions at different intersections, signs also need to encourage people between the decision points indicating that they are going the right way. Signage is usually only one-way so if a traveler gets lost it is difficult to retrace his or her steps. Since signage minimizes the traveler’s cognitive load it does not create mental models of a space, so it is difficult to get back on the right path once they become lost.

• Paper Maps supply a graphical panoptic view of the world. Paper maps offer a means to allow travelers to freely decide the best way to navigate (Woodward 1992). This panoptic view helps travelers to create a good mental model of an area quickly. Travelers, however, need a different map every time they require information at different scales, such as a whole city or a building. Travelers also require different maps every time they are interested in different thematic information, such as, attractions versus banks.

• Immovable Maps are attached to physical objects, such as subway stations and bus stations. Travelers cannot fold these maps and put them into their pockets, therefore, they need to look at these maps long enough to memorize the necessary information for traveling to their destination. An advantage these kinds of maps have over paper maps is the “you are here” information. These maps are useful where the travel routes are confined, for instance, in a transportation system.

• Guided Tours offer good culture information, but they are the most spatially constrained information outlets. This constraint is because guides decide where to go and how to get there. The cultural information a guide supplies is useful for wayfinding, because it helps travelers to remember landmarks.

• Route Instructions could be from a book or from a local expert on the street (Franklin 1996). Route instructions are useful, because they provide information that is egocentric, which is easy to understand for travelers. Like signage this form of spatial information does not help the traveler to create a mental model of the space, so once lost it is difficult to get back on the correct path.

• Spatial Exploration allows travelers to create a cognitive map of an area through exploration excursions, which are usually performed in loops. This wayfinding behavior is a good way to create mental model of a space (Golledge 1999). As time passes, travelers will remember more about their local space and their loops will become bigger.

These descriptions of in-the-field GISs and navigation aids show the advantages and problems with current systems. One of the general goals of this work is to develop a GIS that incorporates the advantages of the different systems into one system while minimizing their problems. This classification of current GISs is used in the description of the map’s evolution. Maps have evolved to incorporate more of the tasks proved by these systems. In Chapter 4 a classification is developed of the types of queries users perform in the field. It is this query operation classification, which was developed based on the functionality of these systems, which forms the core of egocentric queries.

Whichever category the navigational aid belongs to, in order to provide the appropriate information to the user it is necessary to understand the user’s point-of-view. A synonym for a user’s point-of-view is egocentric view. The term egocentric encompasses the information that is collected about the environment from the users’ personal sensors. They see (Sholl 1996), hear, smell, and feel their surroundings. Once navigational aids have an insight into the user’s egocentric view they can more easily offer specific information and geographic data, which allows the navigational aids to function as separate information appliances. The idea of information appliances (Norman 1999) is to have easy-to-use, dedicated devices. People have separate devices for blending food and drying hair, instead of one device for all functions in a household. Propagating this idea to the information technology domain, we use intelligent mobile geographic information systems (GISs) to refer to all egocentric mobile geospatial information appliances. Such GISs can provide geospatial information that is more intelligent (Minsky 1985) than paper maps by providing, for instance, route instructions and additional information about any thematic type in an environment to improve a user’s mental model of a place (Timpf et al. 1992). For a system to intelligently adapt to a user’s specific geospatial information requirements, intelligent mobile GISs require an insight into the user’s egocentric view.

In order to understand the benefits intelligent mobile GISs have compared to other navigational aids, we will examine the history of one of the most popular and useful navigational aids, the map, focusing on how maps have changed over time. The evolution of the map shows that maps are sometimes difficult to use due to their inflexibility.

1.2. Evolution of the Map

Before we improve the concept of the map, we need to understand what a map is, what it provides, and how people use it. The history of the map reveals how it has changed dramatically throughout time. At the same time, however, it shows how the map has provided the basic functionality of giving people spatial information that is difficult to gather directly with their own senses.

The concept of the map has existed for thousands of years (Harley and Woodward 1991). From the beginning, maps in various forms have helped people remember events that happened at a certain time and place: for instance cave drawings can be considered maps (MacEachren 1995). For a map to be categorized as useful it needs to provide information that is consistent with reality and this information needs to be presented in a way that is easy for people to understand. Examining this map evolution we see that not only have maps become more flexible, they have also incorporated most of the functionality of other navigation aids.

From cave drawings to paper, maps have been both mirrors and catalysts of their times by reminding people what the human perspective of a geographic area used to be like, and providing an idea that there might be more area to discover (Harley and Woodward 1991). Medieval maps were historical, literary representations of Christian space, often centered on Jerusalem, with an introverted view of the classical world, describing unknown areas with words like, “Beyond here be dragons.” Though Ptolemy introduced the idea of Latitude and Longitude in 150BC, this concept was not widely used as a basis for maps until the renaissance (Bergren and Jones 2000). The Ptolemaic grid created a single, global orientation of north-south-east-west and led to a major conceptual shift in ways of orienting and positioning geographic objects. A global scale was provided by this coordinate system, where Euclidian distances could be measured in real space and translated onto map space maintaining general spatial relations. This concept of maintaining a consistent scale throughout the map is one of the constraining parameters of what and how objects, or themes of objects, can be displayed on a map (Richardson 1992). For example, a map of the town of Orono probably does not show what furniture is in my office, whereas a more detailed map of the Target Technology Center building, or of room 214, would. This framework of scale, orientation, and theme had an immeasurable effect on people’s perception of geographic space. In the next sections we show how improving these aspects of orientation, scale, and theme enhance maps by making them easier to use and appear more intelligent.

1.2.1. Paper Maps

Paper maps are one of the most popular map mediums and the oldest form of spatial information still in use. They provide bird-eye-views of a space showing large-scale relations. Some of the problems with paper maps are their fixed themes, orientations and scales. This makes it difficult for users to find their location on the map. Since paper maps have these fixed attributes there is a high cognitive load for users to calculate the relationship between the map’s orientation and their own orientation (Kuipers 1982). The main difficulty with these framework attributes of the classical paper map is their inflexibility. A paper map has a single scale (Hudson 1992), which cannot be changed unless a new map is created. The user cannot zoom into a different scale to analyze objects in more detail. A paper map also has one orientation, for example, populations in the Northern Hemisphere are used to north being in-line with the top of the map. To navigate effectively, users have to align the map’s orientation with the real world’s orientation. Paper maps also have a fixed set of thematic information. A map of an area’s road network probably will not display what soil types are found in that area. With the constraint of printing a map on paper comes the limitation that if a user wants information about soil types, he or she will need to print a different paper map. A user cannot re-center a paper map, which usually has its center aligned with the center of the mapped physical space (Couclelis 1992). For instance, a map of the State of Maine is centered on Brownville, which may make it difficult for a user in Portland to determine the “you are here” location whereas a map centered on Portland showing some of New Hampshire and Massachusetts would facilitate this task. Many of the problems that paper maps are known to have are addressed with the development of digital maps.

1.2.2. Digital Maps

There have been many attempts to alleviate the problems paper maps are known to have. Digital maps have provided many solutions. Since digital maps incorporate more than just a map display (Kuhn 1991), the modern, encompassing term for these maps is GIS. GISs offer the functionality of overlaying different map thematic layers, which allows users to combine such spatial information as county boundaries, road networks, soil types, and vegetation to derive new information. With a GIS a user can also zoom between multiple scales, gathering different information about the same object. A GIS analyst can also easily rotate and translate the map’s projection to different orientations and shift the map so that it is centered on different locations. This flexibility of alteration allows the map to seem more malleable. The problems that have arisen from GIS use are due to the system’s complexity (Egenhofer and Kuhn 1998). GISs in general have complex human interfaces and their complex processing requires a substantial amount of computing resources (Worboys 1995). This need for powerful computing has made it difficult to bring GISs from the desktop into the field. Currently, GISs are mainly used in the PC desktop and workstation environment, where digital maps form the basis for the production of paper maps, which are printed and brought into the field. The problem with this situation is that a user must know all the information that will be required in the field beforehand.

1.2.3. Mobile GIS

A mobile GIS is a digital mapping system that functions on a mobile computing platform, such as a Pocket PC or Palm device (Chen and Kotz 2000). There is a difference between mobile digital maps and mobile GISs. A mobile digital map has similar functionality as paper maps, but it is more malleable, whereas mobile GISs include spatial analysis tools and data. Both types of mobile mapping systems have many of the advantages of regular desktop systems, but they are used in the field. These systems do not automatically change the map’s perspective to fit the user’s dynamic needs when they explore their surroundings. These mobile mapping systems have their own problems relating to the complexity of the human-computer interface (Rodden et al. 1998). Mobile computing devices are not as powerful as their desktop cousins, so some functionality is lost (Bertolotto et al. 2002). The often complex interfaces that GISs have in the desktop domain do not scale to mobile devices, where displays are much smaller and input devices, such as keyboards and mice, are very limited (Cooperstock et al. 1997). Another problem mobile GISs have is their lack of contextual awareness (Chen and Kotz 2000). For example, the orientation and scale of a mobile GIS are stagnant unless a user directly modifies them. The cognitive load of the users is still heavy since they must differentiate between their physical orientation and the map’s orientation. It is also difficult for the users to find where their physical location is on the map.

Current mobile GISs provide a limited amount of information pertaining to a user’s specific needs. They give general information and do not cater to the user’s egocentric point-of-view. Also, current mobile GISs do not have the ability of automatically centering on the user’s location or aligning with the user’s orientation (Burnett et al. 2001). These systems are virtually desktop GISs that have been ported to handheld devices: the human-computer interaction framework has not been modified for the in-the-field tasks that users are likely to perform (Mackaness 2002).

1.2.4. Intelligent Mobile GIS

Travelers are more confident and relaxed when they know some background information and can understand why an environment is structured in a certain way (Stea et al. 1996). The more relaxed and “at home” people feel, the easier it is for them to function and accomplish their intended travel goals, such as attending a conference or enjoying Hawaii. By extending the basic functionality of a mapping system to allow it to gain some insights into a traveler’s egocentric point-of-view we advocate intelligent mobile GISs that help travelers not only to navigate (Rodden et al. 1998), but also to gather background information regarding the context of their current locations (Schilt et al. 1994).

Based on familiar, non-digital travel aids discussed earlier in section 1.1, such as a “Lonely Planet Travel Guide,” a set of criteria was developed for an enhanced intelligent mobile GIS. An intelligent mobile GIS needs to be diverse with respect to handling spatial and temporal constraints. As many travelers are uncomfortable exploring foreign spaces without first having some background details on a site, the intelligent mobile GIS must provide well-structured and well-presented background information (Rodden et al. 1998). This background information includes spatial details as well as historical and cultural information about an area. While bus and walking tours are typically highly structured, allowing little opportunity for additional exploration by the traveler, a criterion for the intelligent mobile GIS is that it must be spatially and temporally flexible. This flexibility will assist travelers in exploring a foreign space at their own pace, gathering information that pertains to their specific interest.

Ideally, therefore, an intelligent mobile GIS should provide Cicerone-like interaction. A Cicerone is the most user-centric guide. Cicerones are private tour guides that offer very effective user interaction based on face-to-face human contact (Schmidt et al. 1999). An intelligent mobile GIS will take advantage of the traveler’s location, orientation, pointing gestures, as well as indirectly-gathered information about the user’s interests (Rodden et al. 1998). Being aware of such attributes allows an intelligent mobile GIS to more intelligently guide the traveler, both, spatially and temporally.

In the near future maps will seem more intelligent because of a translation between the user’s egocentric view of the world and the panoptic view provided by the mapping system (Egenhofer 1999). This translation will allow the mapping system to more effectively support a user’s needs. These intelligent mobile GISs will have the ability to automatically adapt to the necessary scale based on the user’s needs. Such a GIS can be aligned with the user’s view of reality as well as inform the user of the global orientation. Such maps will also be user-centric and the scale and themes will adapt to the user’s tasks.

Users in the field have a very constrained perception of the environment around them. Their perspective point-of-view provides them with an egocentric understanding of the local geographic space. Current mapping systems’ lack of an egocentric translation makes it difficult for users to interact with the information systems. Intelligent mobile GISs overcome this limitation by sensing when and how to translate between the user’s on-the-ground egocentric view and the bird-eye-view. The core of the translation process is the spatial awareness based on seven degrees of freedom: three location parameters (x, y, z), three orientation parameters (pitch, roll, yaw), and time (Narayanan 2001).

A mapping system that is aware of the user’s spatial context (Dix et al. 2000) will have the data necessary to decide what kind, amount, and format of information will help the user the most. For example, knowing the user’s location and orientation at all times allows the system to derive the user’s speed. Linking this information to data about the environment provides an insight into user’s mode of travel (Majumdar et al. 2003). If the user’s location is on train tracks moving for a prolonged period of time at a certain speed, the system can infer that the user is most likely traveling by train. Knowing the time of day and weather conditions could be used to decide on how to provide information to the user (e.g., as text, in graphical form, or by sound). An egocentric mapping system has the benefit of linking the spatial information around the user’s local environment to cultural, non-spatial attribute information.

Making mobile maps egocentric alleviates some of the problems inherent in their limited computing functionality. Users will have the perception that they are not directly interacting with the computing device, because the mapping system will be able to derive what the user wants from spatial contextual information and interact with the user the way another human would (Schmidt et al. 1999).

1.3. Research Questions

To improve the map and provide people with useful information about their surroundings we investigate ubiquitous and context-aware computing and show how these systems improve mobile GISs. This paper incorporates the egocentric spatial data model into context-aware computing to answer questions users might ask in the field, such as “Where am I?” or “What is that?” To return meaningful results for these user queries we address the following challenging research questions in this paper:

• How would sensors need to be incorporated into a mobile device in order to become aware of the user’s spatial context and what kinds of sensors are necessary?

• How would a database management system use spatial awareness from a client device to process the terms here and there in order to provide information that is in tune with the user’s environmental context?

To process these questions, an information system needs models, and methods that can consider the user’s location and orientation (Lukowicz et al. 2002), as well as the spatial and temporal aspects of their surroundings. In order to improve interaction with a spatial information system, the system needs to provide information that is intuitive for the user. The concept of intuitive is relative to the user’s spatial, temporal, and cultural context. This paper develops a data model that processes the users’ spatial and temporal context, thereby providing information that is more aligned to their needs.

1.4. Goal and Hypothesis

There are currently many limitations within traditional geospatial information tools. The goal of this paper is to create a seamless integration of the user’s sensed position and orientation into a data model for mobile GISs. This goal is accomplished by identifying the role of spatially-aware sensors within mobile information devices, as well as developing a data model that has an insight into the user’s spatial information needs. This data model incorporates the egocentric spatial abstract data type, which has methods to process such egocentric spatial concepts as here and there. Consequently, the Hypothesis of this paper is:

Measurements from position and orientation sensors are sufficient to formulate executable spatial queries about “here” and “there.”

1.5. Approach

Egocentric mobile spatial information systems have two components: first, a spatially-aware client device, and second, a data model that can process egocentric insights into the user’s perspective view of the world. We investigate spatially-aware mobile client devices showing how spatially-aware sensors can improve the map and better aid travelers, examining the functionality they offer and the form of user-interaction they have (Baus and Kray 2002).

In order for intelligent mobile GISs to have a similar usability as paper maps, mobile computing needs to be just as ubiquitous as paper (Abowd and Mynatt 2000). An examination of ubiquitous computing is performed showing spatial-awareness as a core property of many context-aware computing systems (Burnett et al. 2001). Context awareness is also shown as a necessity for a system to appear ubiquitous (Jiang and Steenkiste 2002). This idea is also justified with the existence of other spatially-aware information systems that showcase the use of context-awareness to ubiquitously provide information to the user.

An intelligent mobile GIS affords many more functional properties than other spatial information systems used in the field. We develop two forms of spatially-aware systems: (1) a map system and (2) a pointing system. The map system is spatially-aware, which allows it to automatically pan, zoom, and rotate the map based on the user’s needs. The pointing technology uses spatial-awareness and functions like a computer mouse, but instead of selecting objects on a computer screen to receive information users are selecting objects by pointing at them in their real world surroundings.

Upon the completion of an investigation of these spatially-aware client devices we examine a second aspect of intelligent mobile GISs, which is an egocentric spatial data model. Such a data model has two components: first, a mechanism to represent the egocentric attributes, such as position and orientation and second, methods of the abstract data type to process the concepts of here and there. Abstract data types are developed to represent the user’s egocentric attributes of position and orientation and show how the standard database query language can be used for egocentric queries. An execution model is then developed for egocentric spatial queries with query re-write rules demonstrating the algorithms behind the queries’ functionalities. This term query re-write means the process of transforming user queries into executable statements.

1.6. Scope of Paper

The setting for this paper is a system that is aware of the user’s position and orientation. The data are sensed at high precision. All user queries are based on a typical travel scenario, so they are constrained to information within a single city. Other scales and granularities are not considered. The egocentric spatial data model works well even when it is not constrained, though current sensor accuracy does not allow us to test the model in more general situations at a global level. Another assumption is that the system has knowledge about objects within the environment, not only their location, but also non-spatial attribute information.

There is no human subject testing perform for the prototype. The prototype developed for this paper should not be considered a fully functional commercial intelligent mobile GIS application. We are only interested in providing spatially-aware map displays. Though we describe the functionality and specifications of an egocentric mobile query system, its implementation is considered future work.

Other attributes that are discussed, but are not incorporated into our system at this time, are the use of distance and spatial data representations displayed in both, two-dimensional and three-dimensional space.

1.7. Assumptions

Many assumptions are made upon which the egocentric spatial data model is based. These assumptions are because in order to develop a robust data model, certain database management system traits must exist. They include:

• Object-Relational DBMS: An object-relational DBMS is a relational database system that has been extended to support object-orientated principles. This assumption is necessary because the egocentric data model is developed as a data object.

• Open DBMS: A database is considered open if developers can create their own abstract data types. This abstract data type creation is done to enhance the inherent modeling power of databases beyond built-in types such as integer and character.

• Spatial Data Support: The egocentric data type processes spatial data; therefore, in order for a database to incorporate the egocentric data type it needs to support the spatial geometry data type.

• Object Inheritance: In object-orientated design a subclass inherits properties from a super class. To ensure that the egocentric data type supports spatial data it should inherit properties from the spatial data type.

1.8. Results

Based on the framework for egocentric query operations some innovative examples of spatially-aware prototypes are described. These descriptions show how current geospatial information systems can be enhanced to more effectively improve user’s knowledge and interaction with foreign environments. The geospatial information systems are enhanced by applying knowledge about the user’s location and orientation.

An egocentric spatial data model is also developed integrating the concepts of here and there into a database management system. This integration is achieved by obtaining position and orientation data and translating these data into actual database queries. This data model is necessary for an intelligent mobile GIS to function as an effective information system providing information that is aligned with the egocentric view of its user.

1.9 Intended Audience

This paper is intended for researchers and developers interested in the design of future geospatial information appliances, in particular mobile spatial information systems. Such an audience includes researchers concerned with Location-based Services, ubiquitous computing, and intelligent spatial appliances, as well as computer scientists investigating how mobile computing affects human-computer and human-environment interaction in the spatial sciences. The developed prototype environment provides a test bed that may be of interest to scientists who want to perform experimental human subject tests or experiments on how people navigate and gather information pertaining to a foreign environment.