Roundup of Papers in Mobiquitous

Here are some papers presented at Mobiquitous. I’ll introduce them by research areas.

Object Tracking
To track mobile objects (e.g., mobile phones, books, children), one may require objects to report their own location. However, in reporting continuously, objects consume considerable energy. To limit energy consumption: Tobias Farrell puts forward a novel protocol (web); and Nico Debauwe and Peter Ruppel propose to combine GPS and GSM cell-ID positioning. Maximilian Michel of DoCoMo looks at object tracking from a different angle: how to make it scalable; and he does so by handling tracking data in a hierarchical p2p architecture. Another problem of tracking objects is how to index their locations. And this is exactly the problem Robert K. Harle has tackled by adapting and combining two existing methods for indexing indoor locations.

Mobile services
First there was simply “sensors”, now the buzzwords in research circles of mobile services seem to be “participatory sensing“. Dirk Trossen and Dana Pavel rightly say what they mean by those words and list cases in which mobile phones may act as sensors to enable participatory applications (eg, community-based environmental monitoring, vehicle warning). They also propose a pub-sub architecture for participatory sensing.

The problem of how to deploy and discover services in ubicomp has won a significant following among the papers presented. For example, to discover services, Wei Gao of ASU proposes the use of a “hierarchical DHT”: ubicomp devices group into a set of local DHTs whose clusterheads are connected to form a backbone; each DHT then indexes the services that its ubicomp nodes offer (it would be very interesting to know the extent to which such an architecture is robust to device mobility). On the other hand, Artin Avanes proposes different data retrieval strategies (retrieval of device profiles) for a flexible service deployment (a deployment that scales for different network topologies).

Another promising research area is body sensors: “Mr Cowen’s model for running better meetings is hard to beat: put in place a system (he suggests body sensors) that enables participants to signal their boredom anonymously. When everyone is known to be bored, the meeting halts.”(link) :-) In addition to this application of body sensors, there are many others, and getting most of those applications to work requires adaptation to changes in context or in service availability. Sye Loong Keoh proposes a policy-based architecture for doing so.

Anonymity and Trust
In sharing digital content, ubicomp devices bump into the problem of how actually to select content (especially if content varies in quality). Of course, devices should preferentially select high-quality content. To do so, one idea (to which some researchers are coming round) is for ubicomp devices to run trust models. A trust model is a piece of software that keeps track of who provides quality content and who does not. I put forward a way with which mobile devices learn from their past experiences to set initial trust values (TRULLO – pdf and post). And Liam proposes to select content by accounting for both trustworthiness and co-location of content producers (pdf and post) – this makes a lot of sense since ubicomp devices (content producers) are usually mobile.
Finally, consider a base station collecting data from sensors. For some applications, it may be important that the base station does so anonymously (ie, it collects pieces of data without being able to link them to the originating sensors). To this end, James Horey suggestes that sensors do not send what they sense but send what they don’t sense – the complement of sensed data. From those complements, the base station guesses what has been sensed. James showes how this approach applies to traffic monitoring (pdf).

Comments are closed.