Research Group Projects and Descriptions

Viral Communications focuses on constructing infrastructure-free agile, scalable, collaborative systems that permit uncontrolled growth, minimal power use, and maximum ability to intercommunicate, with viral architectures moving the intelligence from the trunk to the leaves. The goal is purely photonic communications—optics and radio, no wires, with minimal costs for innovation and flexible architectures. Primary work addresses both the basic mechanisms of radio and the applications that embed communications in the bits and pieces of daily life, from clothes, to dog collars, to furniture.

Increasingly, more devices with communication capabilities and computational power are carried on our body, or are embedded in the environment around us in the form of "furniture"—desktop computers, cars, wall screens, and TVs. However, these devices lack the ability to seamlessly discover and communicate with each other. Cerebro provides a communication platform across a large-scale, heterogeneous set of devices that can potentially cover a large residential area. A basic but extensible set of services is offered, such as presence of users and/or objects in the physical area around us, and a data transportation mechanism that is optimized for efficiency and scalability across a mixture of mobile and static devices. Finally, a framework is offered for development of new applications, ranging from a simple chat to secure identity negotiation between mobile phone users.

Comm.unity is a platform that implements a wireless, device-to-device information system that bypasses the need for any centralized servers, coordination, or administration. A key feature of the platform is that it combines knowledge, awareness, and learning of the user's social relationships and integrates this information into the communication protocols and network services. Comm.unity is designed to work on as many devices as possible, and with as many different radios as possible (WiFi, Bluetooth, IR). It is designed as a platform over which many different networked applications could be developed with ease. SnapN’Share and additional applications in development are intended to be used in upcoming field studies to collect information about user behavior and their social interactions, and aid in fine-tuning the platform’s learning capabilities.

We explore secure, cooperative storage mechanisms in peer-to-peer networks, where a group of peer nodes form a contributory storage infrastructure. In this collective storage platform, each peer node can access a large pool of information by only contributing a small portion of the total required storage. The basic idea is that we fragment a given data-object into several segments and store them in the peer nodes based on DHT (distributed hash table). Each segment is associated with a partial key, and the master key can be retrieved if the number of assembled partial keys reaches a given threshold. In practice, this means one can enlist one's peer group to help with one's storage needs, and vice versa; it is also secure even when one loses one's password.

The Digital Aura project deals with ways to give end users better control over their personal information, leveraging principles of social and physical proximity. People are quite good at establishing a social style and using it in different communications contexts, but they do less well when the communication is mediated by computer networks. It is hard to control what information is revealed and how one's digital persona will be presented or interpreted. In this project, we ameliorate this problem by creating a "Virtual Private Milieu," a "VPM," that allows networked devices to act on our behalf and project a "digital aura" to other people and devices around us in a manner analogous to the way humans naturally interact with one another. The dynamic aggregation of the different auras and facets that the devices expose to one another creates social spheres of interaction between sets of active devices, and consequently between people. We demonstrate our prototype of the "Social Dashboard," a readily usable control for one's aura. Finally, we present "Comm.unity," a software package that allows developers and researchers easy implementation and deployment of local and distant social applications, and present several applications developed over this platform.

Ego establishes a distributed social-networking platform whereby users retain ownership and control over their personal information and social connections. Rather than posting your interests and social connections to Facebook, your location to Google Latitude, and your status to Twitter, post it all to yourself. You produce such data over a lifetime, so you should own, control access to and, if necessary, set the price for it! We demonstrate a society of "agents" representing both humans and non-humans to each other and to Web 2.0 applications, such as Facebook. The architecture of Ego-based agents resembles that of a P2P file-sharing network, allowing user applications that operate by crawling Ego's distributed network, such as a way to gather contextual information (e.g., location, mood) and share it with agents of friends, or an application run by a company that provides incentives such as discounts to agents in exchange for revealing subsets of their personal data.

Fluid Voice is a group communication platform that allows both real-time and asynchronous group activity participation. Users can form local audio broadcast networks for group discussion, similar to a conference call. Disconnected group members can be included in these discussions asynchronously: the audio is recorded and sent to them opportunistically. In this way, membership of a group persists in the face of disconnection caused by communication failures and mobility. Our platform supports other asynchronous applications: polling and voting, sharing of wish lists, and text messaging. Our current work focuses on usability and how to leverage opportunistic interactions in a group-oriented network in a dynamic, mobile context. We aim to understand how to propagate relevant data to every member of a group in an efficient manner.

Alumni Contributor(s): Durga Prasad Pandey

Focal Point turns any screen in the environment into a shared space for personal or social use. Walking up to a screen populates it with pictures, calendars, or other desired items. Multiple people can share the screen simultaneously, and can even have a joint work space for convenient collaboration. In either scenario, the display is locally contained as data is only displayed if the user is within line of sight.

iCityQuest is a location-based, people-driven resource finder that takes the form of a mobile scavenger hunt. Unlike simple wish lists or inventory databases, iCityQuest actively engages the collective power of the local community using a competitive multiplayer game format and monetary rewards. With iCityQuest, users can send requests to the local mobile community to locate parking spaces, discover needed shopping items, or simply to ask for a favor within the neighborhood, and provide rewards to the best responders for their help.

The purpose of this study is to explore the creation of Internal Knowledge Markets. Our goal is to bring an electronic distributed market (decentralized information creation) inside a firm or a laboratory in an effort to (i) measure the value of knowledge created and (ii) link this to productivity and/or profits. A key advantage of this study will be a direct appeal to information economic theory to design an information market. In particular, we will appeal to "two-sided network theory," principles of "information asymmetry," and also "price theory." We seek to bring the rigor of this discipline to real world application and measure the results.

Public and shared displays are starting to permeate the cityscape. Open Spaces opens these displays as part of our mobile experience. As a result, we can exploit our immediate environment as an aura embedding our mobile experiences, freeing our personal data from tiny mobile devices and screens. We can also transform a personal, individual activity, like a phone call or a mobile device interaction, into a companionable experience—sharing a calendar, ordering a cab, exchanging namecards and pictures, or finding objects and acquaintances of common interest. We are creating an architecture and protocols that support such experiences, called the Third Cloud, a framework for using mobile devices to orchestrate interactions among nearby resources and cooperative people.

The Open Transaction Network (OTN) extends the utility of social networks to allow a community of consumers to collaboratively traverse the market for consumer needs through a mobile device, enabling more informed financial decisions in a geo-local context, based on the experience and advice of friends. A specific mobile application allows one to log one's wish list and itemized list of transactions to form a social network around the list of interests. Individuals can share this data to inform and guide others in a timely, personal, and contextual manner when they are shopping for a product or seeking a service. It can also help people connect opportunistically in a local area to make group purchases, pick up an item for a friend, and perform reverse auctions.

With the tremendous explosion of community photos, how can we present a coherent visual model for a large-scale community photo collection? In this project, we aim to build a photo space explorer for effective image browsing, retrieval, and annotation on top of our existing, distributed, community storage mechanism. The key is to identify object and feature points relations among community photos in an effective and scalable way, and to distribute the storage through an ensemble of mobile devices. The photo space explorer also enables users to navigate community photo collections using a 3-D browser.

Mobility in urban environments is often coupled with a quest for efficiencies such as a nearby parking spot, a ride to some destination, a date for dinner, a ticket to the theater, or any other kind of service that one might need while on the move. The idea is an ad hoc, mobile "Craig's List." We demonstrate a system that provides an easy way to advertise services or goods offered or sought by means of mobile phones and ultra-mobile PCs, and we are concerned with issues of ensuring trust, security, and anonymity for all participants.

The future of mobile communications will certainly be one where devices will opportunistically connect to a variety of networks, obtaining and providing services on the fly. Our portable devices will carry our identity and act as intermediaries between users and their rich environments. The Amulet embodies this vision. We are developing this new device which instantiates a user's identity and provides a computing model that allows the device to securely exploit its context and surrounding infrastructure in an open, secure, and adaptable manner.

Twopons is a Twitter-based viral coupon service which helps eliminate unnecessary spam by allowing only those people who actually make purchases in certain stores to share the coupons with their peers. It is a social-network-oriented coupon system that runs on top of the Open Transaction Network. It also incorporates geographic filtering to make it only available in locations of interest.