A Transport Information Monitoring Environment (TIME): Event
Architecture and Context Management (TIME-EACM)
A
framework (TIME) for research, application development and deployment for
transport monitoring in the city of Cambridge will be announced by the end of
2004, with the backing of a number of large industrial partners including BT
(our project partner on TIME-EACM), Boeing, Oracle, Vodafone and
The
research will be carried out jointly by groups at the
The
IAN LESLIE is Professor of Communications and the Pro
Vice-Chancellor for Research at
The
DR
KEN MOODY and DR JEAN BACON are Readers in Distributed Information Management
and Distributed Systems respectively. They lead the Opera research group,
working on the design and deployment of open, large-scale, widely distributed
systems comprising multiple domains. Recent EPSRC research grants include: Global
Computing using Events, with Nortel; OASIS role-based access control; Access
control policy management;
DR
DR
RICHARD GIBBENS is a Senior Lecturer in the Computer Laboratory with interests
in the modelling of computer and communications networks. Prior to joining the
Computer Laboratory in 2001 he was both an EPSRC-funded Research Associate and
then a Royal Society University Research Fellow based in the Statistical
Laboratory within the Department of Pure Mathematics and Mathematical
Statistics. During this time he collaborated with Professor Frank Kelly
(who is currently partially seconded to the Department for Transport as Chief
Scientific Advisor) on a variety of research projects including investigations
of dynamic routing strategies, statistical measures of effective bandwidths in
multi-service networks and congestion pricing in packet-switched networks. Dr
Gibbens was Principal Investigator of grant GR/M09551/01 and is co-investigator
of grants GR/S86266/01, GR/T10510/01 and EP/C510712/1 (forthcoming).
The
PROFESSOR
ANDY HOPPER is Head of the Computer Laboratory (CL) and leads the LCE which has
pioneered location aware, sentient computing, see www-lce.eng.cam.ac.uk. DR ROB HARLE is a postdoctoral researcher
in LCE, specialising in positioning systems (indoors and outdoors), sensor
interpretation, context-aware and sentient computing, wearable computing and
autonomous navigation. DR ALASTAIR BERESFORD is a Research Associate with Alan
Mycroft in the CL, specialising in system-level security and privacy issues in
ubiquitous and pervasive computing. His first degree was in the CL. He then
spent a year with BT, working with our collaborators, then took a PhD in the
LCE.
The
DR
NIKI TRIGONI obtained her PhD with the Opera group at the
There are two distinct aspects of our motivation for
this proposal. One, the application
level, is to increase transport efficiency through monitoring and processing;
the other, system research, is to investigate ways to develop context-aware
applications through open access to information bases derived from multiple
sensor streams.
Road congestion in the
It is therefore vital
that we improve the performance of the
Several transport information
systems already exist [Hou95, Woo95], for example to gather data on traffic
density on the M25 motorway, to provide information displays at bus stops, to
control traffic signals to ease congestion or to give a fast route to emergency
vehicles, to display the number of empty spaces in car parks. However, they are single-theme, vertically
integrated projects that as yet do not fully exploit the potential of fixed and
wireless networks. There is typically one organisation, e.g. the City Council, that
controls the gathering, processing, and delivery of transport information.
We believe that a
functional pinch point, much like the delivery of IP packets in the Internet,
will allow innovation in sensor development independent of applications, and will
provide application developers a common, open interface on which to
build that will be robust to changes in the underlying technology and that
allows the gathered data to be shared in a controlled way. This functional pinch point is the event
based middleware, together with the derived-context models, as
described below, and their design and implementation lie at the core of this
proposal. With such an infrastructure in place, diverse applications can be
developed through controlled subscription to the infrastructure services. No
such infrastructure currently exists. An important aspect of the project is to ensure
that, as required by law, the privacy of individuals is not violated. There are
many possible applications, including congestion detection and projection,
car-park status, bus arrival time displays, free taxi location, support for
emergency services, none of which could, on their own, justify a generic
approach. The algorithms developed for statistical analysis of data and
inference of behaviour will also assist policy makers in long-term planning,
for example for decisions on congestion charging.
The goal of the
project is therefore to investigate, design and provide a secure but open
interface to support the controlled sharing of monitored data. The concrete outcomes of this
work will be an event-based middleware that hides low-level sensor aggregation
from applications, integration of high-level context models with query support,
and an evaluation of this support based on prototype, but real-world,
applications that fully exploit the architecture.
The LCE has long been involved in cutting-edge
research into location systems, radio networks, high performance networking and
the use of sensors to infer context [Ber03, Ber04, Har03,04, Mad03, Pol03, Sco03, 05,
Vid03,04]. They have experience deploying and evaluating large
sensor systems and interpreting the data with reference to context extraction
and security and privacy.
Alan Mycroft provides a programming language
viewpoint on abstracting high-level information from sensor networks and on
controlling such information. Recent work is Katsiri's PhD thesis “Middleware
support for context awareness in distributed sensor-driven systems” and associated
publications [Kat03,04]; there, a framework (SCAFOS) is
proposed containing novel
approaches to summarising a vast rate of low-level sensor data into forms which
can be used by high-level applications. Also relevant is work with Scott (PhD
thesis “Abstracting Application-Level Security Policy for Ubiquitous
Computing”) and Beresford [Sco03, Mad03], and Beresford&Stajano's work
[Ber04] on anonymising sensor data to preserve confidential information without
losing the data-effectiveness of higher-level information.
The Opera group have worked on event-based middleware (
Dr Gibbens is currently exploring road
transport data supplied by the Highways Agency as part of the MIDAS project to
collect per minute traffic measurements from the core motorway infrastructure.
This work, in collaboration with Professor Frank Kelly and Dr Gaurav Raina who
is currently an EPSRC-funded Research Associate, is looking at studies of speed
flow relationships, e.g. the PEMS project at Berkeley [Cho02, Bic01], within
the motorway network as well as approaches to statistical visualisation of
traffic data. These pilot studies are proving a rich and challenging area for the
application of statistical approaches.
Some of the work of Bill Fitzgerald's group [Mor94, Kam99,00,
Pun01,02,03] concerning new statistical methods based upon Sequential Monte
Carlo and optimal (particle) filtering has, in the present context,
applications ranging from the tracking of objects in images, as well as the
tracking from sensors, such as acoustic, radar and radio, to the ability to
detect changes sequentially in time. The framework also enables information
from different sensors and sensor arrays to be fused together in an optimal way
to enable probabilistic inferences to be drawn from the observed data as well
as the ability to extract relevant features from images which can then be
classified (i.e. bus not car) using machine learning approaches.
Niki Trigoni has worked on several
aspects of query processing and data dissemination for sensor networks. In
collaboration with the Cougar group at
Our project partners BT, as well as being
expert in communications and networks, have worked on event-based middleware and
security for pervasive computing environments [Bri04, Sop04]. They are sponsors
and inaugural members of the
We shall build an infrastructure to provide as much
transport monitoring information as possible, in real time, to interested
parties. The city will become sensor-rich with a number of static sensor
networks, as well as sensors and electronic identifiers on mobile objects. The
challenge is to manage the high volume of sensor output, and present
information in a high-level form useful for achieving decision support and
planning. Areas that would benefit include congestion control and the
reliability and predictability of public transport.
In outline, the major research dimensions are as
follows (see Fig. 1):
·
Sensor network (wired/wireless) and aggregation technology including
sensor data provenance
·
Volume and heterogeneity management – aggregation algorithms for sensor
data
·
Event-based middleware (EBM) architecture,
event composition, event logging, event databases.
·
Evaluation of support for applications to control devices using the EBM.
·
Provision of context models. Visualisation, (e.g. maps) for application
use, of some models as appropriate.
·
Query specification and maintenance on databases, context models and
composed events, for decision support and planning. Integration of push, pull
and hybrid models of data and query dissemination.
·
Support for automating the management of the network
·
Access control policies and mechanisms for applications’ use of event
services and contexts. Use of secure communication mechanisms. Conformance with
privacy law. Performance and economic costs of security.
·
Inference of application behaviour. Its placement within the
architectural levels.
·
Real-time detection of occurrences through monitoring and inference.
·
Short-term projection from data; prediction of conditions and events.
·
Statistical analysis of historical data to support long-term policy
decisions as well as decisions in real-time.
As we deploy applications and understand their
requirements more clearly we shall evaluate and evolve this generic support
that abstracts above potentially changing technology.
The
main objective is to identify a sensor network architecture that is tailored to
the needs of the particular application. Since scalability is an important
consideration in traffic monitoring, we plan to investigate layered
architectures that include both resource-constrained and resource-rich nodes.
Resource-constrained nodes, which have limited communication, computation,
storage and energy resources, will operate in low duty cycles and will be
primarily targeted at sensing tasks. Resource-rich nodes will have high duty
cycles and will perform more routing, in-network storage and in-network
processing. We will consider hybrid
architectures, not only in terms of node capabilities, but also in terms of
node mobility. Our sensor network will contain primarily static nodes, but
these must be able to communicate with mobile nodes, e.g. sensor nodes,
application-running nodes or electronic tag nodes mounted in vehicles.
An important part of our work will be to
provide reliable data transmission from the source nodes through the network to
the middleware services. A lot of research has recently been done on routing
algorithms for wireless sensor networks. In this work package, we will study
the efficiency of different routing algorithms in our application scenario. We will
interact with WP2 and WP3 in order to ensure that the in-network processing
tasks (dictated by query and event patterns) are well-integrated with the
underlying
This
work package is concerned with query processing and data management techniques
for sensor networks. Due to the limited bandwidth and, in some cases, limited
energy resources, it is impossible to forward large amounts of raw sensor data
to a powerful gateway node, to be processed there in a centralized manner. This
advocates the need for pushing part of the data management services into the
network, and performing them at the nodes in a distributed manner. Each node
can be viewed as a mini-repository of sensor data generated and stored locally
in its memory. The network can therefore be treated as a distributed sensor
data management system (SDMS) [Int00, Mad02, Dem03, Des04, Geh04, Woo04]. SDMSs
differ from traditional distributed databases in that they are limited by
stringent processing, bandwidth and, often, energy constraints.
We plan to leverage our previous experience on
query processing and multi-query optimization [Tri03, Trig04, Tri05] for sensor networks in order to provide suitable
data management techniques for our traffic monitoring scenario. Important
issues that must be addressed are definition of query language, in-network
query processing and in-network storage.
This work package is to design and create an event-based
(publish/subscribe) middleware (EBM), above the distributed, raw sensor data
and make it available, conveniently and efficiently, for application developers
to build on. A research issue is where in the architecture to place aggregation
and composition of raw event data to create higher-level events. For example,
the raw data may be widely distributed and contain complex sensor and vehicle
IDs, whereas the high-level events will need to indicate application-meaningful
locations and vehicle types, such as buses and taxis. But the fact that event
publication must happen in real time argues for the placement of at least some functionality
in the infrastructure rather than as a service above the middleware. We will
interact with WP1 and 2 to resolve these issues. We will interact with WP4 and
5 over the publication of events by databases and context managers.
An
application package will include support for secure use of the middleware, and
related services, in terms of application roles by application developers. This
will allow publishers to advertise and publish event types and subscribers to
browse available event types and subscribe to those of interest. The package
will also manage, on behalf of applications, the security credentials necessary
for using the middleware. At a higher level, the application will be offered
detection of patterns of events (composite events) of interest and will be able
to express (persistent) queries for the various databases and context
representations that emerge.
The objective of this work package is to
integrate, manage and provide querying tools for large amounts of information
coming from different and potentially heterogeneous sensor sources. The
database system will store two types of information, application domain data
(e.g. high-level traffic-related events) and network control data (.e.g.
information about network data flows, failures, node energy reserves and
message latencies). The application domain data will be provided to a number of
client applications with diverse information needs. The network control data
will be used in order to identify potential failures or malfunctions in the
sensor network(s) and react to them accordingly.
This work package is concerned with
developing programmable, secure and privacy-preserving techniques that abstract
the vast amount of raw sensor data into deliverable high-level events. Deliverables
include extension and implementation of selected ideas and approaches from
Katsiri's and Scott's PhD theses [Kat03,04, Sco03,04], reworked for
transport-related events. LCE will continue their work on context design and
presentation, e.g. in the form of marked-up maps.
This work plan is concerned with the
statistical analysis and modelling of traffic data and the enabling real-time
decision support systems. The incoming data needs to be analyzed to provide the
accurate estimation within the constraints of delay and available computational
resources. This work will be broad in
scope including techniques of statistical data analysis and estimation, using large
scale data sets, and probabilistic modelling of traffic flow relationships
within a road network.
This work package is concerned with the
modelling of image data as well as data from various sensors, to enable
real-time decisions to be made concerning any anomalous events and changes that
may have occurred. Also, the data from various sensors will be fused in order
to be able to obtain greater reliabilities. The classification of objects
present in the various data streams will also be investigated.
This
work package will evaluate the potential for controlling the environment,
through real-time decision support, arising from the monitoring infrastructure
and analysis tools developed throughout. There are many examples: potential
traffic congestion could be foreseen and avoided or existing congestion
alleviated. Traffic light control could give a clear run to emergency services,
fire and ambulance. In-car and environmental sensors could detect dangerous pollution
levels and attempt to reduce them by redirecting traffic. Displays at bus stops
could provide information on bus locations.
The software developed by this project will allow experiments to be set
up to provide data on which to base long-term planning of traffic policy. An
example is the issue of whether to introduce congestion charging.
This
work package will study security and privacy issues including the analysis of
risks and threats. WP3 will produce an event-based middleware secured by
role-based access control so that applications run in a controlled way and
visibility of data is strictly controlled. We will provide tools to allow
system managers to express policy about which roles of which applications can
access which (attributes of ) which data in order to conform with legal
requirements. WP9 will study privacy issues arising from the visibility of
monitoring data gathered from the city, and its subsequent analysis, and
possible inference of the behaviour of individuals. We will also investigate
anonymity achievable with various technologies, such as RFID tags and 2D targets;
also pseudonimity and inference of identity, interacting with WP7. The legal
implications and requirements of a monitoring system will be studied. This WP will
also influence WP4 on database integration – what can be stored in databases and
what applications may see, arising from the monitoring?
Initially,
we will evaluate the monitoring potential of the cameras and sensors already in
place in the city. The 2-D barcodes (targets) developed by LCE will be
considered for attachment to buses and other traffic in the city, for
recognition via cameras. We will design
experiments based on the use of the in-car sensor systems of the LCE for this
project. We will use the selected technology in initial small experiments as an
aid to establishing requirements and guiding software development. An example
is to monitor the new University bus service; its clients find it irregular
while the service providers claim all is well. The University owns property
along the length of the bus route. Also, we will explore the potential of
mobile phones for detecting people and vehicles for our initial experiments.
By three years into the project we will
have made the event-based middleware available, with integrated database and
context-management services. We will deploy this software above a selected
sensor-ed and networked area as the major experiment to prove the conceptual
and design work. We will evaluate the software infrastructure and make any necessary
changes. In addition, the TIME framework
programme is likely to include application developers by the later years of the
TIME-EACM project. We will encourage and assist in their use of our
software/deliverables.
Because
of the broad scope of this work, the need to produce application support and
the need to evaluate this for a variety of applications, we request funding for
five years. We will have regular project meetings, at least six-monthly, with
interim interaction by email, as well as frequent subgroup meetings. As
companies join the TIME Framework consortium we will make regular presentations
to them and assist in their use of the software we develop.
Staff: A. Beresford
will be RA throughout the project, liaising with Birkbeck on all aspects of
sensor and network technology and experiments, and with the BT-funded RA and
staff on security and privacy. Dr
Gibbens and Professor Fitzgerald request RAs for three years. We will recruit
an RA to work on event-based middleware, e.g. the Opera group has people becoming
available. Birkbeck will have a PhD student plus two years of RA support (a
possible extension for the PhD). BT will fund an RA for two years, in the first
instance, with primary interest in WP3 and 9. We have indicated a spine 9 (£24,820)
starting point for the RAs to recruit high caliber, experienced people (in
expensive Cambridge and London), to allow for increases in staff costs for
those who start a year or so into the project, and because of the age of possible Opera group recruits. We anticipate
that our PhD students (not funded by the project) will be attracted to work in
its general area and are focussing on funding for a strong RA core (252 person-months
plus BTs 24).
Equipment: We
request general purpose computers for software development, analysis and simulation
plus additional storage for the project, total £12,000. We expect company
interest and support for the demonstrator(s) in years 4 and 5 but are setting
aside, for sensors and networks, £10,000 in
Travel and subsistence: £40,000 to allow the grantholders, RAs and
PhDs to present the research at conferences, and to companies, and to visit other
groups working in the area. (A contribution rather than full cost to keep the
budget at £1M).
Please
refer to the enclosed letters of support. BT are committed partners in this
research project (TIME-EACM), as well as in the wider TIME framework programme
for Cambridge, which already includes BT, Boeing, Oracle, Vodafone and
The investigators have a good track record
of publishing papers at major conferences and in journals and will continue to
do so. Our project partner BT, and other companies involved in the
The event-based middleware developed for
the project will be made available as open source software. It has potentially
wide application, for other traffic monitoring projects and for a wide range of
event-driven applications where sensors are used to monitor state. The models and algorithms resulting from the
theoretical studies of the monitoring data will be a further output. Professor
Kelly’s involvement as Chief Scientist at the DTI, and his direct knowledge of
the project, will ensure a fast track for any results to policy makers.
[Bac95] Bacon J, Bates J, Hayton R and Moody K, Using Events to Build Distributed
Applications, Proc IEEE SDNE Services in Distributed and Networked
Environments, pp.148 - 155, Whistler, British Columbia, June 1995.
[Bac01] Bacon J., Moody K. and Yao W.: Access
Control and Trust in the Use of Widely Distributed Services. In Middleware 2001, Lecture Notes in Computer
Science 2218, pp.295-310, Springer 2001.
[Bac02a] Bacon J. and Moody K.: Adaptive middleware: Toward open, secure,
widely distributed services.
Communications of the ACM, 45(6):pp59-63, June 2002.
[Bac02b] Bacon J., Moody K. and Yao W.: A model of OASIS role-based access control
and its support for active security. ACM Transactions on Information and
System Security (TISSEC), 5(4):pp492-540, November 2002.
[Bac03] Bacon J., Moody K. and Yao W.: Access control and trust in the use of
widely distributed services. Software - Practice and Experience, 33(4),
pp375-394, April 2003.
[Bel03a] Belokosztolszki A., Eyers D. M., Pietzuch P.
R., Bacon J., Moody K.: Role-based access control for publish/subscribe
middleware architectures. In International Workshop on Distributed
Event-Based Systems (
[Bel03b] Belokosztolszki A., Eyers D. M., Wang W.,
Moody K.: Policy storage for role-based access control systems. Proc.
12th IEEE International Workshops on Enabling Technologies (WETICE'03),
pp196-201, June 2003.
[Ber03] Beresford A.R, Stajano F.: Location Privacy in Pervasive Computing,
Pervasive Computing, pp.46-55, Vol 2 No 1, 2003.
[Ber04] Alastair R. Beresford, Frank Stajano, Mix
Zones: User privacy in location-aware services. Pervasive Computing and
Communication Security (PerSec), Pages127-131, IEEE, March 2004.
[Bic01] P J Bickel, C Chen, J Kwon, J Rice, P
Varaiya and E van Zwet Traffic Flow on a Freeway network, Proc. MSRI Workshop on Nonlinear
Estimation and Classification,
[Bri04] R Briscoe The implications of pervasive
computing on network design BT
Technology Journal, vol. 22, no. 3, July 2004 www.btexact.com/publications/papers?doc=73147
[Car01] Carzaniga A, Rosenblum D, Wolf A, Design and evaluation of a wide-area event
notification service. ACM
Transactions on Computer Systems 19(2001) pp.332-383
[Chl02] T Chloe, A Skabardonis and P Varaiya, Freeway Perfomance
Measurement System (PEMS): An Operational Analysis Tool,
[Dem03a] Alan
Demers, Johannes Gehrke, Rajmohan Rajaraman, Niki Trigoni and Yong Yao. The
Cougar Project: A Work-In- Progress Report. In SIGMOD Record, vol 34, no 4,
Dec 2003.
[Dem03b]Alan
Demers, Johannes Gehrke, Rajmohan Rajaraman, Niki Trigoni and Yong Yao. Energy-Efficient
Data Management for Sensor Networks. In 2nd IEEE Upstate
[Des04] Amol
Deshpande, Carlos Guestrin, Samuel Madden, Joseph Hellerstein, and Wei Hong. Model
Driven Data Acquisition in Sensor Networks. VLDB 2004. Best Paper Award.
[Eug03] Patrick Th.
Eugster, Pascal Felber, Rachid Guerraoui, Anne-Marie Kermarrec The many
faces of publish/subscribe.
ACM Computing Surveys 35(2),
pp.114--131, June 2003
[Geh04] Johannes
Gehrke and Samuel Madden. Query Processing In Sensor Networks. Pervasive
Computing, 3(1) pp.46-55, 2004
[Har04] Robert
Harle, Andy Hopper, Dynamic World Models from Ray-tracing. Proceedings
of the Second International Conference on Pervasive Computing and
Communications, IEEE, March 2004.
[Har03] Robert
Harle, Andy Hopper, Building World Models By Ray-tracing Within
Ceiling-Mounted Positioning Systems. Proceedings of the Fifth International
Conference on Ubiquitous Computing (UbiComp 03), October 2003.
[Har03] Robert
Harle, Andy Hopper, Using Personnel For Indoor Autonomous Environment
Discovery. First Conference on Pervasive Computing and Communications
(PerCom 2003), March 2003.
[Hou95]
[Int00] Chalermek Intanagonwiwat, Ramesh Govindan,
Deborah Estrin. Directed Diffusion: a Scalable and Robust Communication
Paradigm for Sensor Networks. ACM/IEEE International Conference on
[Kam99a] Alvin H.Kam and William J.Fitzgerald: Unsupervised
Multiscale Image Segmentation, Proc.
10th International Conference on Image Analysis and Processing (ICIAP'99),
[Kam99b] Alvin H.Kam and William J.Fitzgerald:General Unsupervised
Multiscale Segmentation of Images Proc. 33rd Asilomar Conference on Signals,
Computers and Systems, October 24-27, 1999, Pacific Grove, California, USA
[Kam00a] Alvin H.Kam and William J.Fitzgerald: Image Segmentation:
An Unsupervised Multiscale Approach Proc. 3rd International Conference on
Computer Vision, Pattern Recognition and Image Processing (CVPRIP 2000), vol
II, pp.54-57, 2000
[Kam00b] Alvin H.Kam and William J.Fitzgerald: Motion Segmentation
of Colour Image Sequences Using Mean Shift Clustering and Bayesian Multiscale
Classification IEEE Conference on Computer Vision and Pattern Recognition
(CVPR 2000),Hilton Head, South Carolina, June 13-15, 2000.
[Kam00c] Alvin H.Kam, Tian T. Ng, Nick G.Kingsbury and William J.Fitzgerald:
Content Based Image Retrieval through Object Extraction and Querying
IEEE Workshop on Content-based Access of Image and Video Libraries
(CBAIVL-2000) Hilton Head, South Carolina, June 16, 2000.
[Kam00d] Alvin H.Kam and William J.Fitzgerald: A General Method for
Unsupervised Segmentation of Images using aMultiscale Approach
6th European Conference on Computer
Vision (ECCV 2000),
[Kat03] E. Katsiri
and A. Mycroft. Knowledge Representation and Scalable Abstract Reasoning for
Sentient Computing Using First-Order Logic, in Challenges and Novel
Applications for Automated Reasoning CADE-19 Workshop, July 28, Miami,
USA, 2003
[Kat04] E. Katsiri,
J. Bacon, and A. Mycroft. An Extended Publish/Subscribe Prtocol for
Transparent Subscriptions to Distributed Abstract State in Sensor Driven Systems using
Abstract Events, in Proc. Third International Workshop on Distributed
Event-Based Systems
[Mad03] Anil Madhavapeddy, Alan
Mycroft, David Scott, Richard Sharp, A Case for Abstract Security Policies. Fifth
International Conference on Security and Management, June 2003.
[Mad02] Sam Madden, Michael Franklin, Joe Hellerstein
and Wei Hong.
[Mor94a] R. D. Morris and W. J. Fitzgerald., Stochastic and
Deterministic Methods in Motion Picture Restoration.Proc. International
Workshop on Image Processing,
[Mor94b] R. D. Morris and W. J. Fitzgerald., Replacement Noise in
Image Sequences - detection and interpolation by motion field segmentation. Proc.
ICASSP 94,
[Pie02] P. Pietzuch
and J. Bacon, Hermes: A Distributed Event-Based Middleware Architecture,
Proceedings of the 1st International Workshop on Distributed Event-Based
Systems (
[Pie03a] P.
Pietzuch and J. Bacon, Peer-to-Peer Overlay Broker Networks in an
Event-Based Middleware. In H. Arno Jacobsen, editor, Proc. 2nd
International Workshop on Distributed Event-Based Systems (
[Pie03b] P. Pietzuch,
B Shand and J. Bacon, A Framework for Event Composition in Distributed
Systems. Proc.4th ACM/IFIP/USENIX Int. Conf. on Middleware
(Middleware '03), pp.62-82,
[Pie04] P. Pietzuch, B Shand and J. Bacon, Composite
event detection as a generic middleware extension. IEEE Network, 18(1): pp.44-55,
February 2004.
[Pol03] C. Policroniades, R. Chakravorty, P. Vidales,
A Data Repository for Fine-Grained Adaptation in Heterogeneous Environments.
3rd ACM Workshop on Data Engineering for Wireless and
[Pun01] E Punskaya, C. Andrieu, A.
Doucet, W.J. Fitzgerald, Particle Filtering for Demodulation in Fading
Channels with Non-Gaussian Additive Noise, IEEE Trans. Communications, vol.
49, no. 4, pp. 579-582, 2001.
[Pun02a] E Punskaya, C. Andrieu, A. Doucet, W.J. Fitzgerald, Bayesian
Curve Fitting using
[Pun02b]
[Pun02c] E Punskaya, A. Doucet,
W.J. Fitzgerald, Bayesian Curve Fitting with Applications to Signal
Segmentation, IEEE Trans. Signal
Processing, 50(3)
pp.747-758, 2002.
[Pun03] E Punskaya, A. Doucet,
W.J. Fitzgerald, Particle Filtering for Joint Symbol and Parameter Estimation
in DS Spread Spectrum Systems (with A. Doucet), Proc. ICASSP 2003.
[Rat02] Sylvia Ratnasamy, Brad Karp, Li Ying, Fang
Yu, Deborah Estrin, Ramesh Govindan and Scott Shenker. DHT: A Geographic
Hash Table for Data-Centric Storage. ACM International Workshop on Wireless
Sensor Networks and Applications (WSNA) 2002.
[San96] R.S. Sandhu,
E.J. Coyne, H.L. Feinstein, and C.E. Youman. Role-Based Access Control
Models. IEEE Computer, 29(2): pp.38--47, Feb. 1996.
[Sco05] David
Scott, Richard Sharp, Anil Madhavapeddy, Eben Upton, Using Camera-Enabled
Personal Devices to Access Bluetooth Mobile Services. To Appear. Mobile
Computing and Communications Review, ACM, January 2005
[Sco03a] D. Scott,
R. Sharp Specifying and Enforcing Application-Level Web Security Policies
Invited contribution to Jul/Aug 2003 IEEE Transactions in Knowledge and Data
Engineering (TKDE)
[Sco03b] David
Scott, Alastair Beresford and Alan Mycroft, Spatial Policies for Sentient
[Sco03c] D. Scott,
A. Beresford, A. Mycroft Spatial
Security Policies for
Presented
at FASE 2003 (part of ETAPS 2003) in
[Sop04] A Soppera and T Burbridge Maintaining
privacy in pervasive computing -- enabling acceptance of sensor-based services
BT Technology Journal, vol. 22,
no. 3, July 2004 www.btexact.com/publications/papers?doc=73141
[Tri04] Niki Trigoni,
Yong Yao, Alan Demers, Johannes Gehrke, Rajmohan Rajaraman. WaveScheduling:
Energy-Efficient Data Dissemination for Sensor Networks”. International
Workshop on Data Management for Sensor Networks (DMSN), 2004.
[Trig04] Niki Trigoni,
Yong Yao, Alan Demers, Johannes Gehrke, Rajmohan Rajaraman. Hybrid
Push-Pull Query Processing for Sensor Networks. GI Workshop on Sensor
Networks (WSN), 2004.
[Tri05] Niki Trigoni, Yong Yao, Alan Demers, Johannes Gehrke,
Rajmohan Rajaraman. Multi-Query Optimiazation for Sensor Networks.
International Conference on Management of Data (SIGMOD), under submissions,
2005.
[Vid04] Pablo Vidales, Leo Patanapongpibul, Glenford
Mapp, Andy Hopper, Experiences with Heterogeneous Wireless Networks -
Unveiling the Challenges. Second International Working Conference on
Performance Modeling and Evaluation of Heterogeneous Networks (
[Vid03] Pablo Vidales,
Leo B Patanapongpibul, Rajiv Chakravorty, Ubiquitous Networking in
Heterogeneous Environments. 8th International Workshop on
[Woo95] K Wood and R T Baker: User
guide to the "gating" method of reducing congestion in traffic
networks controlled by SCOOT. TRL, February 1995.
[Woo04] Alec Woo,
Samuel Madden, and Ramesh Govindan. Networking Support for Query Processing
in Sensor Networks. CACM 47(6) Special Issue on Sensor Networks, pp.47-52,
June 2004.
[Yon05a] Yoneki, E.
and Bacon, J.: Distributed
Multicast Grouping for Publish/Subscribe over
[Yon05b] Yoneki, E.
and Bacon, J.: Dynamic
Group Communication in
[Yon04] Yoneki, E.
and Bacon, J.: Content-Based
Routing with On-Demand Multicast. Proc. ICDCS –
Workshop on Wireless Ad Hoc Networking (ICDCS – WWAN 2004), pp.788-793,
Tokyo, Japan, March 2004.
[Yon03a] Yoneki,
E.: Many Aspects of Reliability in
a Distributed
[Yon03b] Yoneki, E.:
[Yon03c]Yoneki, E.: Pronto: