Rodger Lea – research

Core Interests

  • Ubiquitous and pervasive systems and the Internet of Things (IoT)
  • Distributed systems, multi-media and rich media experiences
  • Technology transfer and applied structured research

Current projects

I am currently involved in a number of projects covering the UbiComp and IoT areas. These are carried out either at Lancaster University, the HCT lab at the University of British Columbia or via the Urban Opus project.


Internet of Things (IoT) and Smart Cities

We are using Smart Cities as a broad umbrella for a number of research projects. Our IoT platform, WoTKit, has been extended with opendata support to act as a smart City data hub and used in two smart city projects:

  • UrbanOpus – a broad collaboration exploring smart cities based in Vancouver Canada.
  • SmartStreets An investigation of IoT technologies for transport and Highways based in the UK

Our latest IoT platform, the Web of Things Toolkit (WoTKit) is a web centric IoT platform designed to make it easy to rapidly prototype and build IoT applications and mashups. The WoTKiT platform can be accessed here and the best paper describing it was presented at the 2012 Internet of Things conference in Wuxi, China BIB PDF.

We have spun out a company to commercialize this technology, Sense Tecnic Systems Inc.

  • MB2 and the RESTBroker. The WoTKit leverages our earlier work on an IoT broker originally called the MAGICBroker which was described at the 2010 IoT conference BIB PDF. In turn the MB was built on our earlier work on UbiComp platforms, PlaceMedia which was a context aware platform for developing Ubiquitous computing applications with a media centric focus built based on learnings from our original UbiComp Common Model

Hybrid Cloud Computing

We are exploring the issues raised by deployment to hybrid (public-private) clouds. In particular we are developing algorithms and tools to allow semi-automatic partitioning of cloud applications. This work uses run time code analysis (both binary and byte code levels) to determine module dependencies and build application graphs. It then applies graph partitioning algorithms – adapted to handle cloud constraints – to determine optimal placement of modules across a hybrid cloud deployment.

  • Cross-Tier Application and Data Partitioning of Web Applications for Hybrid Cloud Deployment. Paper
  • Adaptive Scalable Resource Profiling for the Cloud in the Cloud. PDF BIB
  • MANTICORE: a Framework for Partitioning Software Services for Hybrid Cloud. PDF BIB

 


Pervasive Displays is an umbrella for a variety of projects exploring the use of large displays in private (home) and public (urban) settings. Our particular focus is on new interaction techniques using personal devices such as cell phones as well as lightweight infrastructure for interaction.

Three representative papers that details some of the research contributions are:

There have been several sub-projects as part of our Pervasive Displays work, ranging from our original research in 2006 funded by Panasonic Research through to the NSERC funded Personal Screens, Public Interaction (PSPI) project.

Additionally we have been involved in several external projects that have either used our technologies or we have collaborated on research with. These include the UK eCampus project (Lancaster University) and the follow-on EU funded PD-Net project as well as the Canadian research collaboration MUSE.

  • PD-Net started as a Lancaster University project that is deploying public screens on campus. eCampus aims to explore issues relating to interaction with large public screens, screen management and scheduling and designing for multi-modal interaction. Details at: eCampus
  • MUSE is exploring mobile media centric applications and services for the urban environment. It has a strong focus on culture based content and is funded by Heritage Canada. The project is a collaborative research network including researchers from UBC, SFU and BCIT. It also includes local small companies as well as IBM and Nokia.

 


Previous significant projects

Ambient Intelligence: (mostly carried out by my lab members at the New Media Innovation Centre).

While at NewMIC my team carried out a number of initial investigations into areas related to Ambient Intelligence. These included explorations of context aware multi-media messaging, social networking applications and collaborative media interaction.

Home Media related projects (mostly carried out by my lab members at Sony USA Research labs)

During the period 1996-2002 I headed up a research lab for Sony in San Jose, CA. During that period my lab carried out a number of research projects in the areas of home networking, home media, broadband and interactive TV, personalized media and multi-media operating systems. HAVi – Home Audio Video Interoperability: an attempt by the major Consumer Electronics companies to develop a common interoperability architecture for home devices. Based on IEEE 1394 (FireWire) it offered automatic plug and play, backward and forward compatability and a full resource and AV stream management system. Although the architecture was never adopted as a full standard, aspects can be found in a range of devices from Matsushita, Hitachi etc and parts of the specification ended up in DVB MHP, OpenCable and the Java.TV spec. (org.havi.ui)

Personalized media synthesis engine; Blendo, developed as an internal engine for new forms of multi-media, Blendo derived its roots from the VRML2.0 spec and open inventor. The Blendo team (maily hired in from SGI) explore a range of advanced interactive TV centric media experiences using a XML derived markup language.

Home media server; The lab developed a number of advanced prototypes for broadband services in the home. many of these relied on an advanced home media server that combined features of a media engine and home gateway. We prototyped a variety of these hardware and software platforms using technologies such as the Open Services Gateway (OSGi), HAVi and Jini.

Broadcast Virtual Worlds Was an example of a possible broadand service that used advanced visulisation technology to create a synthetic 3D scene into which real live video was composited. We experimented with a number of scenarios including a virtual car racing service that combined video feeds from Nascar with camera and race track data toi allow users to switch between real TV Video and synthesied 3D animations based on actual track data. Users could track the race from virtual view points, replay action, explore car and driver statistics etc. Another prototyped created a virtual concert hall where viewers could switch between real music video and virtual scene.

Aperios Aperios was the production version of the meta object based research operating systems developed in Sony Tokyo Labs. During 1996-2000 Sony developed a commercial version of this operating system and deployed it in a range of devices including the Sony robots, Sony digital set top boxes and home media servers. Eventually we decided to drop our own in-house OS as Linux became more viable and the developer chain and libraries for Linux matured.

Networked VR & distributed Multi-media During the period 1994-1996 I ran a research team for Sony research in Japan, partly at the Computer Science Lab and partly in the corporate labs. We developed several prototypes including:

Community Place: an early attempt a a virtual world service a-la Second Life. Sony built community place on the emerging VRML standard which was co-authored by my group and developed a player and authoring tool. The Community Place system was eventually productized and run as a Japan only service for 2 years as Sony experimented with online virtual worlds and associated business models.

OS related My original research was in the area of distributed systems and particularly distributed operating systems. Three projects of note that I’ve been involved in are:

  • Apertos: A meta object based operating system developed at Sony CSL. Apertos, originally developed by a team led by Yasuhiko Yokote, eventually evolved in Aperios, a commercial offereing by Sony and ended up in a variety of set top boxes and even early AIBOs! . Apertos papers are here
  • Chorus/COOL: I led the team that developed the 2nd version of COOL, the Chorus Object Oriented Layer built on the Chorus Micro-Kernel. This project was joint between Chorus systems and Marc Shaprio’s group at INRIA.
  • Commandos: The Commandos project was a EU funded (Esprit) research project looking t at the development of distributed object oriented platforms.