3. Client Devices for Intelligent Mobile GIS’s

As a starting point in designing intelligent mobile GISs we consider two forms of egocentric spatial information systems: a map and a pointer. The first system we investigate is a mobile mapping system, whose advantage over other maps is the ability to automatically adjust panning, zooming, and rotation based on the user’s real world movements. The second system is based on the concept of pointing. With this system a user can point at a geographic object and receive some form of feedback about the object. The metaphor for the pointer is to function like a magic wand (Egenhofer and Kuhn 1998). To receive information the users point at an object in their surrounding and receive information in a format of their choice. The description of these systems focuses on the advantages they have over current technologies and describes their functionalities.

3.1. Spatially-Aware Map Technology

Maps still provide the main means for understanding spatial environments, as well as for performing tasks such as wayfinding, trip-planning, and location-tracking. Static traditional maps have several disadvantages, such as:

• Fixed orientation: that is the map always faces in one direction (typically north). The map users, however, may be facing any direction at any given moment. Hence, in order to understand the map users need to perform some kind of rotation, either of themselves or of the map to align their frame of reference with the map’s frame of reference. This process puts an immense cognitive load on the users, because it is not always intuitive and may present considerable difficulties, especially in cases of complex, uniform or unfamiliar spatial environments.

• Fixed scale: that is the map cannot be changed to a different granularity level. This limitation is one of the most restrictive aspects of paper maps. The scale determines the level of zooming into a spatial environment, as well as the level of detail and the type of information that is displayed on a map. Users, however, need to constantly change between different scales, depending on whether they want a detailed view of their immediate surrounding environment or a more extensive and abstract view in order to plan a trip or find a destination. Current solutions to the problem include tourist guides that comprise maps of a specific area at many different scales. Tourist guides, however, are bulky books, difficult to carry around, and search time is considerable as they typically consist of hundreds of pages.

• Inability to represent a changing world: that is all spatial environments and the objects that they encompass, whether artificial or natural, are displayed statically although they are actually dynamic and change over time. Artificial spatial objects, such as buildings, may get created, destroyed, or extended, while others, such as land parcels, may merge, shrink, or change character (e.g., when a rural area gets developed). The same holds true for natural entities, for instance, a river may expand or shrink because of a flood. The static 2-dimensional map is restricted to representing a snapshot in time and the information on it may soon become obsolete.

• Limited display of thematic information: that is an inability to show many varying thematic information at the same time. There are many different types of maps, such as morphological, political, technical, tourist-oriented, and contour maps. The thematic content of a static map has to be defined a-priori and is usually restricted to one area of interest. Even then, the information displayed is minimal. For example, a tourist map will indicate that a building is a church or a restaurant, but it is highly unlikely that more information will be available, such as the construction date of the church or the menu of the restaurant and the type of cuisine it offers.

An egocentric intelligent mobile GIS may be defined as a GIS that uses sensors to gather information about the users reference frame and adapts the visualization of the map data accordingly. In abstract terms, this adaptation is defined as a translation of the map reference frame to the absolute reference frame based on the users personal reference frame. The relationship between these three spatial reference frames is illustrated in Figure 3.1.

Figure 3.1: Relationship between the real world, map, and user’s orientation

Sensing the user’s position and orientation allows the map to rotate in accordance with the user’s frame of reference. As a result, the system provides what we call adaptive orientation. The north-orientation convention can be dropped. Furthermore, the system provides intelligent zooming and panning: The user’s location is at all times displayed through a dot on the screen of the mobile computer and the system pans through the map as the user change his or her location by moving in the spatial environments. Zooming is also very easily performed and can be done manually if desired, or automatically, provided that the system has some information about the context of the user (e.g., entities of interest, motion speed, and other parameters).

Since it supports thematic contexts this system is beneficial to any type of user, from map professionals such as surveyors and GIS analysts to technicians and tourists. For instance, when the user is a tourist, the display provides the relevant information and type of map that would serve this type of user best, such as churches and museums. For a sewer technician, however, it would provide a map of the sewer system for the city and the relevant technical information. All the supplied information depends on the underlying geospatial and thematic dataset that is being loaded on the system.

Figure 3.2 shows a prototype interface for such an egocentric mobile map. The sensor output on the right side of the display shows the GPS coordinates and orientation data provided by the sensing hardware. The left side showcases the thematic data types, such as roads or buildings. The map display has a dot highlighting the user’s current location and an arrow showing the direction he or she is facing. In this example, the user is interested in buildings and is located on the central mall of the University of Maine facing in the direction of Fogler Library. The bottom panel of the interface shows some attribute information pertaining to the library. In the bottom right panel the user is given the choice of listening to the attribute feedback. This system works with standard geospatial data types comprised of both raster and vector data.

Figure 3.2: An egocentric mobile map running on a Tablet PC.

Such mobile map technology is useful for people who are used to working with maps. Another form of interaction that might be useful for people less experienced in using maps is a pointing device.

3.2. Spatially-Aware Pointing Technology

A spatially-aware pointing technology will allow users to point at an object or location and learn about it (Figure 3.3). The pointer provides information about x, y and z (or latitude, longitude, and elevation) and pitch, roll, and yaw (angle up from the horizon, angle side to side from the horizon, and compass angle). The sensed positional and orientation information is integrated into a database management system. This database management system contains geospatial data. The user’s position and orientation are used to create polygons representing the egocentric concept of here and there. These here and there polygons are then checked against the geospatial data in the database management system and a list of best-case candidate objects (e.g., building) are identified. The user can then request additional information about the identified objects (e.g., name, date built or history).

Figure 3.3: Representation of pointer use.

A key challenge for the pointing technology is to optimize the spatial information retrieval. The pointer needs to match the real-time measurements of location and orientation with the best candidate object in a geospatial dataset. Integrating location-orientation data with a digital landscape model and developing a plausible computational model that targets granularity is key to the success of the pointer. The pointing technology will exploit the use of orientation sensors so that geospatial datasets are not only user-centered, but also egocentrically oriented. This aspect is relevant to the pointing technology, since otherwise no distinction could be made between such cognitive aspects as in front and behind.

The pointing technology consists of two parts: the physical pointer to sense the user’s egocentric spatial reference (Figure 3.4a) and a mobile computation platform, such as a PDA, to process the sensed data, calculate within a digital landscape model the element to which the user pointed, and generate an auditory or textual response to the user (Figure 3.4b). The pointer will house an orientation sensor unit that captures yaw and pitch; a GPS unit to determine location (x, y, and z); a Bluetooth transceiver for wireless transmission of the sensed data to the PDA within a near range; and a battery.

Figure 3.4: (a) Orientation sensor, GPS, Bluetooth, and battery, and (b) PDA with pointer software, data, Bluetooth.

3.3. Summary

Two mobile spatially-aware client technologies are demonstrated in this chapter. The first is a mapping technology that provides automated panning, zooming, and rotating functionality based on the user’s real world spatial context. These automated attributes alleviate the user from having to put themselves cognitively into map space. The second technology is a pointer that allows users to point and select objects in their surroundings. Once an object is selected the pointer can provide information about it. The two mobile technologies define the requirements for mobile spatial data processing:

• sensor-driven

• distributed, and

• egocentric

The two spatially-aware mobile clients demonstrated in this Chapter are only the first half of a mobile spatial data processing system. Chapters 4 and 5 develop the egocentric spatial data model, which provides a database management system with an insight into the reference frame of the user. This data model extends spatial data management with functionality of incorporating the user’s position and orientation to process the concept of here and there based on sensed data collected from spatially-aware clients like the ones described in this chapter.