University of Minnesota Driven to Discover
U of MNUniversity of Minnesota
Center for Transportation Studies
Mohamed Mokbel

Mohamed Mokbel

"Toward Scalable and Privacy-Aware Location-Based Services in Transportation"

Presentation by Mohamed Mokbel, Dept. of Computer Science and Engineering

October 24, 2006

If you have ever wondered how far it was to the nearest gas station, struggled to remember which neighborhood grocery store was open late, or wandered through an unfamiliar area in search of that new restaurant, then you have an interest in location-based information. For centuries, signs have been the primary means of communicating this kind of information; today, however, explosive growth in personal electronic devices, wireless electronic communications, and geographic information systems (GIS) is driving the rapid development of new kinds of location-based services.

Providing location-based services to meet this growing worldwide demand, while ensuring that users’ identities and personal information are protected from prying eyes, is the goal of research by Mohamed Mokbel of the Department of Computer Science and Engineering. Mokbel presented his ongoing work in an Advanced Transportation Technologies Seminar sponsored by the ITS Institute October 24, 2006.

As the required technologies continue to mature, Mokbel foresees the potential for location-based services to penetrate every aspect of life. However, he says, two important challenges must be overcome in order for this potential to be realized.

The first challenge is in the area of privacy. Recent news reports of Global Positioning System (GPS) receivers being surreptitiously used to spy on victims point out the potential for location-aware technologies to enable privacy violations. True location-based services, incorporating users’ personal information to drive information retrieval, present even greater potential for abuse. At present, location-based services adhere to a model in which users give up privacy in order to gain access to information; for example, a user may agree to let the service provider know her exact physical location in order to receive information about nearby travel routes.

In order for users of location-based systems to receive accurate information while at the same time preserving their anonymity and masking their exact coordinates, a new technology must be added to the mix. Mokbel described the operation of CASPER, a “location anonymizer” developed during his research, which acts as a trusted third party between the end user and the location-aware information provider. CASPER enables users to define privacy profiles that control how much information service providers receive about them

CASPER “blurs” users’ private information in two ways. First, users can stipulate that their information must be hidden within a group of other users, with the size of the group controlled by the user; for example, a user may specify that their transmitted location must include at least ten other users. Second, users can define a spatial boundary on their exact position, such as a radius of half a mile. By using these two methods together, it is possible for users to develop a variety of profiles suitable for different situations.

Mokbel described how location anonymizers such as CASPER would be embedded within location-based database servers, enabling such systems to process queries based on blurred spatial regions rather than exact positional information. He went on to describe an algorithmic approach to such a query, along with mathematical proofs of its correctness. He then demonstrated a prototype system in operation using geospatial data on the Minneapolis area.

The second major issue holding back the development of location-based services—scalability—was the subject of the remainder of Mokbel’s presentation. He argued that the traditional “snapshot” approach to database queries, in which a new query is issued every time data must be updated, is inadequate to the demands of location-based systems that require continuous updates. Instead, he proposed an incremental query model in which results are continuously updated with new data. In order to support this model, he explained, database architecture for location-based systems must move from a layered approach (with location-based systems built on top of GIS on top of data storage) to an integrated approach in which location-based query processing is a part of the database itself.

Another important aspect of Mokbel’s proposed architecture is shared execution. Instead of each query being handled by an individual “thread” of program execution, shared execution enables a single thread to handle multiple continuous queries, thus reducing computational overhead.

These principles are implemented in Mokbel’s location-based data server, dubbed PLACE for Pervasive Location-Aware Computing Environments. Working together, PLACE and CASPER form a backbone for future location-based services that are both highly scalable and capable of ensuring the privacy and anonymity of users.

Mokbel closed his presentation with an invitation for researchers to attend an international workshop on privacy-aware location-based services, to be held in conjunction with the Eighth International Conference on Mobile Data Management in Mannheim, Germany, May 7–11, 2007.