Research Students

Mohamed Ahmed

[Home Page] Research:
Supervisor: Stephen Hailes and Peter Kirstein

I'm a completing Ph.D. student and a research fellow at the Department of Computer Science, UCL. I am working within the MARS project alongside Rae Harbird and Alexandr Seleznyov. The project is based at UCL@Adastral with the at BTexact Future technologies group. My academic advisors are Stephen Hailes and Peter T. Kirstein, and Robert Ghanea-Hercock is my industrial advisor. My interests are varied and include: component based system and frameworks, context management and adaptation for mobile systems, multi agent systems, peer-to-peer frameworks for distributed and open systems, policy representation, reasoning, management and languages, trust and reputation management for cooperative systems, risk identification and reasoning for security, social network definitions and application for cooperative and mobile systems.

Lucia Del Prete

Research: Service Composition in Mobile Environments
SupervisorLicia Capra

In the last years two major trends have emerged: a great penetration and evolution of mobile devices and the proliferation of blogs and personal content spaces, revealing a transformation of users from traditional consumers to active producers of content. It will not be long before these two trends will converge, and we will consume - in additions to the traditional services delivered by powerful server machines accessible via wide area networks and aiming to Five 9s availability - services provided by devices in the surroundings that will need to be composed to deliver valuable functionalities to the end user. 

In order to be successful, composite services in mobile environments will have to be perceived by the end user as if delivered  by a unique provider, reachable and available for the duration of the service. To do so it would be required that just service compositions that can be successfully completed would be initiated. With this goal in mind, in my research I study how information about users' historical mobility patterns, together with composition semantics, can be used to maximise the chances of successfully consume compound services in mobile environments. Now, selecting providers based on their predicted colocation will not protect users against malicious ones, nor will enable them to discover services of interest to them. As such, I am  now studying the integration of distributed trust models and recommender systems  to dynamically reason about the trustworthiness of a composition and the likeness that an user may find it interesting or useful.

Ettore Ferranti

[Home Page] Research: Sensor Networks and Mobile Robots
Supervisor: Niki Trigoni, New College, University of Oxford

Recent advances in micro-electromechanical systems (MEMS) are allowing the deployment of small mobile nodes with sensing capabilities within an area for monitoring purposes. These nodes have processing and storage capabilities, and are able to communicate with each other through wireless multi-hop links. They have the potential of sensing the ambient environment, and collaborating with each other to detect high-level events of interest. My research is focusing on sensor network and mobile robots deployment in an emergency sce-
nario, which present critical requirements for reliable and real-time event detection. The objective is to explore a dangerous area after a disaster and gather information about hazards and victims, and to guide first responders to the latter avoiding the former. The main challenges that will be addressed are i) to provide reliable detection of user-defined composite events, even in the presence of intermittent failures in the communication infrastructure and ii) to ensure immediate (real-time) reaction to high-level events by activating nearby sensors, communicating high-level events to interested parties, or raising alarms. At the implementation level, this project will involve evaluating distributed event detection techniques using simulation tools and a real sensor network platform.

Adam Greenhalgh

[Home Page] Research: Power Conservation in Ad Hoc Networks
Supervisor: Stephen Hailes

My research interests lie in the area of mobile networking particularly using routing in Ad Hoc networks to conserve power in low power radio systems. My PhD research is funded by the EPSRC through an Industrial Case with Philips Research Labs; I also have interests in Internet Routing and have contributed to the development of the BGP component of the XORP Extensible Open Router Platform during an internship at ICIR.

Tom Hewer

[Home Page] Research: Large Scale and High Fidelity Vehicular Network Simulation
Supervisor: Peter V. Coveney and Maziar Nekovee

Vehicular communications technology has been supported by the reduction in price of equipment and the increased availability of frequency spectrum allocation. The use of computer equipment in vehicles has been well established and now the manufacturers are investigating ways to transmit vehicle information, for various applications. The expense and size of vehicles means that testing new protocols and systems is unfeasible in the real world. This leads to a requirement for high-fidelity, timely and accurate computer simulation platform. Current simulation tools have been designed to simulate static,  wired telecommunications networks, and as wireless communication and mobility has increased, these tools have become less effective. The size of vehicular traffic networks and the complexity of the mobility patterns requires a large amount of computing power to run simulations in a timely fashion. I will research the use of parallel and distributed grid computing systems to accelerate the running time of simulation, that will operate at large-scales and in time for decision-making and safety applications. This will enable the development and validation of vehicular network protocols and network mechanisms and perform QoS measurement for media and data delivery systems.

 

 

Rae Harbird

[Home Page] Research: Adaptive resource discovery for ubiquitous computing 
Supervisor: Stephen Hailes

I am investigating the application of ad hoc routing algorithms to service discovery and I have designed and developed RUBI, a Resource discovery algorithm for UBIquitous computing. RUBI is unique because it is adaptive, altering the discovery mechanisms used based on the perceived stability of the network.

Research Keywords: ubiquitous computing, pervasive computing, service discovery, ad hoc networks, wireless routing

Selection of Papers:

(pdf) Rae Harbird, Stephen Hailes, and Cecilia Mascolo. Adaptive resource discovery for ubiquitous computing. Middleware for Pervasive and Ad-hoc Computing, 2004.

 

 

Neal Lathia

[Home Page] Research: Trust-Based Recommender Systems
Supervisors: Stephen Hailes and Licia Capra

Recommender systems are becoming ubiquitous: they already have a dominating presence on the web, and collaborative filtering is the prime example of harvesting the wisdom of the crowds. However, the methods used to generate recommendations often do not take into account the temporal aspect of the process, and simply iteratively apply one of the many classifiers available for this problem. My research addresses this gap, by focusing on modelling a recommender system as an evolving community of interacting users who need to decide who to trust.

Afra J. Mashhadi

[Home Page] Research: Coping with unwanted content in mobile ad hoc network
Supervisor: Licia Capra

Web 2.0 has transformed Internet users from passive consumers to active producers of digital content(e.g., personal stories, music, pictures, etc.). Such content can be created and shared worldwide at virtually no cost, by means of social networking sites. As research on popular Web2.0 websites demonstrates, users' interest is distributed in such a way that a tiny portion of content is of interest to many users, and the vast majority of content is of interest to only small groups of users instead. We anticipate that the same trend will move to the mobile ad hoc networks where users carry resource constrained portable devices. My research is focused on how to disseminate content in MANeT, so that it reaches the interested users while not overwhelming others with unwanted content.

Liam McNamara

[Home Page] Research: Content Distribution in Urban Transport
Supervisors: Cecilia Mascolo and Licia Capra

Most people carry wireless electronic devices, including when travelling upon public transport. These devices can play media and wirelessly procure content from neighbouring peers. This can be a laborious, manual process for the user. To improve these systems, I use techniques for wireless connection pattern prediction, involving social network concepts. Incorporating an awareness of urban environments, unreliable networks, malicious users and leveraging trust systems. To test these mechanisms I have developed software for mobile device testbeds and exploratory network simulations.

Daniele Quercia

[Home Page] Research: Trust Models for Mobile Content-Sharing Applications
Supervisors: Stephen Hailes and Licia Capra

Daniele is currently based in London where  he's trying to create some algorithmic weirdness as part of his PhD at UCL.  He writes about mobile 2.0 research and industry on the group's blog. He also speaks at top-tier conferences (such as ICDM and Ubicomp) about the way co-located mobile users will share digital content (e.g., pictures, videos) in the future. He is  a Microsoft Research Cambridge PhD Scholar and MBA Technology Fellow  of London Business School. He worked at the National Research Council in Barcelona and at the National Institute of Informatics in Tokyo. He studied at Politecnico di Torino for his MS in Computer Engineering and, while doing so, he also graduated from Universitaet Karlsruhe (Germany) and from University of Illinois (USA).

Research Keywords: ubiquitous computing, pervasive computing, social networks, security, machine learning, reputation systems

Selection of Papers:

(pdf) MobiRate: Making Mobile Raters Stick to their Word. Ubicomp 2008.
(pdf) Lightweight Distributed Trust Propagation. ICDM 2007.

Elisa Rondini

[Home Page] Research: Collaborative Computation with Bandwidth Control in Wireless Sensor Networks (WSNs)
Supervisors: Stephen Hailes and Wolfgang Emmerich

In Wireless Sensor Networks (WSNs), the tradeoff for nodes between cost, size, and battery lifetime versus computational power may leave individual nodes unable by themselves to perform all the conceivable calculations that might be requested over their lifetime. Techniques for load sharing in traditional computer systems, notably computational Grid Computing systems, have been extensively studied but have not yet been applied in these situations. Moreover, whenever nodes need to collaborate with each other to achieve a global goal, they need to coordinate by exchanging messages while distributing computation in the system. However, resource-constrained devices have a limited amount of bandwidth availability. In addition, wireless communication is prone to network interference and collision. Consequently, in this research, we explore the convergence of the field of ad hoc sensor networking with that of computational grids, and present novel paradigms able to take into account both the computational capabilities of nodes and their local network conditions while distributing jobs computation within the system.

Valentina Zanardi

[Home Page] Research: Uncovering Relevant Content Using Tag-based Recommender Systems
Supervisor: Licia Capra

Our work aims at solving the problem of information retrieval for Web 2.0 scenarios in which disseminated data are completely unstructured and often lack of any kind of categorization. Since no well-defined data model is adopted and explicit ratings are not available to understand users' preferences, we propose a new information retrieval technique capable of leveraging the knowledge of the social network underlying collaborative tagging systems.

Missing

  • Vera Cady
  • Ilias Leontiadis
  • Daniel Madadi
  • Panteha Saeedi
  • Christian Wallenta
  • Mohibullah Wallizada
This page was last modified on 08 Dec 2008.