Intelligent services in mobile GIS applications

Intelligent services in mobile GIS applications


Evangelos kotsakis
Evangelos kotsakis
Institute for the Protection and the Security of the Citizen (IPSC),
Joint Research Center
European Commission, ITALY

Michalis Ketselidis
IPSC, Joint Research Center,
European Commission, Italy
[email protected]

Wireless and mobile environments bring new challenges to users and service providers when compared to fixed, wired networks. Smaller and more powerful mobile communication devices improve connectivity in a wireless environment. Mobility brings additional uncertainties, as well as opportunities to provide new services and supplementary information to users in the locations where they find themselves. It is reasonable to expect that such devices will play a major role in shaping what is known as “ubiquitous computing”. Much of today’s research is focused on improving the mobile devices (reduce their size, cost, power consumption etc.), enhancing the communication technologies (bandwidth, coverage, connectivity etc.) as well as facilitating data management (efficient caching, querying, deliver of data).

Although all of the research efforts mentioned above are central to the successful development of ubiquitous computing applications, an equally important issue is the modeling of con-text sensitive behavior for location aware services. Context awareness is the driving force of ubiquitous computing, as it requires the devices to be aware of the environment as well as of the tasks the user will perform in the near future. Context aware applications range from an intelligent notification system to a “smart space”, which is a selfadaptive environment that adjusts itself according to who is present and what they are doing.

Context awareness imposes significant demands on the knowledge maintained by the system and the inference mechanism used. It requires an internal representation of the user’s preferences and role as well as a sophisticated mechanism for monitoring the environment and tracking user actions while assisting the user in performing the required task in real time.

In our approach, location-based data is described as a set of Location Data Records (LDR). Each LDR corresponds to an alternative in the decision process. Our approach utilizes fuzzy clustering to generate a number of fuzzy classes containing LDRs and then it prompts the user to state her interest. User preferences take the form of goals and constraints. These are in turn assigned a grading value for each LDR. This forms a decision matrix, which can be used for making decisions either by employing a Max/Min evaluation or a weighted vector in case the goals and the constraints are of varying degrees of importance.

Fig 1 A mobile client (pocket personal computer) running an embedded GIS

System Architecture

Client/Server Architecture
The system is based on a client/server paradigm with the clients being the mobile devices (like pocket PCs, personal digital assistants etc.) and the server running on a desktop computer in a Local Area Network (LAN). The communication link that physically connects the client with the server is based on either a wireless LAN or a General Packet Radio Service (GPRS). The interaction between the clients and the server is realised via a web HTTP interface through CGI and Java technologies (Servelet containers).