.

PhD projects

Combined Source and Channel Coding Schemes for Distributed Storage Networks

PhD Student
Tim Porter
Supervisor
Dr X. Peng
Sponsor
EPSRC/Xyratex

My area of research will be focused around "Combined Source and Channel Coding Schemes for Distributed Storage Networks" which will include streaming methods for distributed on-demand video, channel coding for redundancy, efficient source coding, and the application of post processing techniques such as error concealment if the channel conditions are very poor.

The source coding and channel coding techniques should be combined to create an efficient and well documented method of video transport over lossy networks, in particular, the internet.

The methods implemented should allow for seamless dissemination of media content in distributed storage networks.

Developing extended life wireless sensor networks using a traffic-based approach.

PhD Student
Qasim Raza Iqbal
Supervisor
Dr. D.J. Holding
Sponsor
EPSRC

Ad hoc wireless sensor networks (WSNs) are formed from self-organising configurations of distributed, energy-constrained, autonomous sensor nodes. The service life of such nodes depends on the power supply and the energy consumption, which is typically dominated by the communication subsystem. One of the key challenges in unlocking the potential of such data gathering sensor networks is conserving energy so as to maximize their post-deployment active lifetime. In this research we study different network topologies.Quantitative analysis shows that in ad hoc WSN's there is a complex and subtle relationship between various approaches to minimising inter-node communication while maintaining an acceptable QoS and functionality. Concurrent research into energy aware networking using system-level modelling has shown that, if novel network management and network protocols are used as energy management techniques, it is feasible to extend the system life significantly. The need is to demonstrate, through detailed modelling of the network management techniques and message-packet modelling of the protocols, that such extended life systems can be achieved.

This research explores how the use of optimum node transmission ranges can increase WSN cluster head lifetimes and we develop comparative models of equal-grid, COTS, and optimised WSNs. We have modified the network simulation tool NS2 to accommodate nodes with different transmission ranges, and develop detailed simulation and energy models of the three simple linear wireless sensor networks, including our transmission range adjusted (optimised grid) WSN network. Our detailed, packet-level, simulations show that the cluster head lifetime can be improved by 40 to 60% if the radio range is adjusted appropriately compared with the best case where equal radio range and COTS systems are used. The research also shows that QoS is further improved from 74% to 93% by much better throughput. The QoS parameters include much higher packet delivery ratio, latency lower by a factor of 11 and much faster packet in-arrival time. Current work includes improving the total network lifetime by re-modeling the network to include intra-cluster traffic, effects of node density, increased path loss as well as employing a two dimensional model of a network.

For more information please click here.

Multipath distribution for digital storage networks

PhD Student
Richard Haywood
Supervisor
Dr X. Peng
Sponsor
Xyratex and EPSRC

My current research looks into the possible benefits which could arise through the use of multipath distribution in congested networks. With the rise in acceptance of peer-to-peer networks, the use of multiple-path technology will be a viable and important method in the future for a number of reasons. When looking at peer-to-peer technology, the obvious advantage is the reduction on the load at the server. There is also the benefit that the data will be transported through many different paths to the destination with increased reliability and efficiency, which is being investigated currently in my work. My research is split into a number of parts - these include simulations in JSim as well as real world testing though the use of PlanetLab. My investigation into the use of multiple paths shows that there are both advantages and disadvantages in terms of loss and delay performances. This work will is be extended to combining the multiple-path strategy with the erasure coding technique.

Implementing SMAC Protocol on FPGA

PhD Student
Reza Moridifar
Supervisor
Prof. K.J. Blow
Sponsor

Wireless sensor networks are most interested technology to monitors physical and environmental condition such as temperature, pressure and etc.

In a wireless sensor network, large number of tiny nodes will cooperate together to achieve a common purpose. Each node has a simple processor, limited resources of memory and a transceiver.

Each node is battery operated to make a simplify deployment and because of the large amount of nodes which are involved, changing the batteries is difficult, impossible and expensive.

Insufficient battery life, limitation in size and cost of sensor will make protocol design complicated in term of using normal techniques used in other protocols.

Recent Researches are focused on designing a long term and independent operation of sensor network to rectify above issues.

The goal of this project is to find out the requirements, design and implement a basic sensor node. The sensor node must be simply adapted to suit a more specific application. The sensor node developed will have to be strong to stay alive in the extreme conditions environments. In addition to developing the wireless sensor nodes, this project also involves classifying the strengths and weaknesses of the other schemes in term of energy efficiency, routing, network management and etc.

Consequently, there are three purposes that have to be fulfilled within the capacity of this project. Firstly, to develop a basic sensor node by using a FPGA core. The second purpose is to analyse the strengths and weaknesses of the design and method used to develop the sensor node. And the last is to test, modify and optimise the protocol in terms of energy efficiency and performance. for more information please visit my website.

Characterisation of Voice services over Next Generation Wireless IP networks

PhD Student
Ahmed Dastagir Shaikh
Supervisor
Dr. M.A. Eberhard, Prof. K.J. Blow
Sponsor
DTI

Voice services have long been dominant over the cellular technology market, but with the advent of the IMS and the growth of 3G networks, there will be strong necessity for voice services to be provided on the same technology as that of the multimedia services with better quality.

This brings us to the era of Voice over IP services over wireless networks, the demand of which keeps increasing rapidly. There are several advantages of Voice over IP services over the traditional PSTN or 2G cellular systems. First and foremost, they are a lot cheaper. Secondly they are portable. A user can login from anywhere so long as he or she has access to the IP world. Besides a VoIP service can provide almost all features which a client can desire from a current generation PSTN or 2G cellular system. However there are drawbacks as well. Because of the nature of the IP network, the Voice Packets can have variable transit time through the network. The delays experienced are larger than that of the Circuit Switched network and if the networks are congested it would mean that there are lost packets. Providing VoIP services over today's networks as well as future networks has been a challenge because of the many issues which need to be tackled such as efficient bandwidth management at the edges of the cellular and other wireless networks.

This PhD project is part of the Aston HIPNet project, which aims at addressing some of the problems related to voice and other media services over next generation wireless IP networks. This project will also specifically focus on voice models for next generation wireless networks and will also look into greater depth the impact of the various languages, dialects and accents of the users on these models and how this could create a change in the pattern of voice traffic and improvement of efficiency of spectral resources in the next generation wireless IP networks. Also part of the research work will be dedicated to the SIP signalling and the analysis of the many voice codecs associated with VoIP for the next generation wireless IP networks and relevant conclusions and proposals for improved models will be made.

Cooperative diversity in wireless communications

PhD Student
Guchun Zhang
Supervisor
Dr X. Peng (academic), Dr. Xuanye Gu (industrial)
Sponsor
BT and Aston University

Cooperative Diversity in Wireless Communications is a cutting-edge technology for next generation broadband wireless networks. In broadband wireless networks, users can download video clips, listen to music, play on-line games, and even make video calls. These multimedia applications require much wider data bandwidth, thus the so-called "spatial diversity" technology would be needed in these systems. The diversity technology is capable of overcoming the particular physical constraints in wireless environments, such as, fading, multipath, inter-channel interference. As an important part of "spatial diversity" systems, cooperative diversity involving multihop data relays is a solution to improving propagation performance, expanding coverage and enhancing system capacity in such wireless environments. Cooperative diversity is the new area that involves the concept of "virtual array of antennas" rather than physically deploys multiple-antennas, since for end users with small platforms multiple antennas cannot be accommodated. In a cooperative diversity system, the "third party" mobile devices acting as relays to help the main transmission link improve the performance such as bit error rate (BER), data rate, coverage. The involvement of the "third-party" complies with some particular rules called "protocols". The PhD project aims to develop new coding techniques and protocols, and investigate the performance of the cooperative diversity systems by means of mathematical modelling, computer simulation, and laboratory experiments.

Investigation of TCP/IP in Mobile Wireless Environments

PhD Student
R Taank
Supervisor
Dr X. Peng
Sponsor
EPSRC

The aim of the research is to investigate the feasibility of developing an optimised version of the Transmission Control Protocol (TCP), which governs today's major communications networks, from the wired backbones to the heterogeneous networks.

The main problem is that packet losses in traditional wired networks are usually as a result of congestion somewhere in the network path, a situation which conventional TCP can resolve efficiently.

In wireless networks, however, such environments are prone to packet losses due to high bit error rates (BER), unstable channel characteristics, and host mobility induced disconnections. TCP interprets these packet losses as an indication of congestion and inappropriately invokes congestion control mechanisms (i.e. retransmissions and increasing delays), leading to unnecessary degradation of network performance.

To rectify the associated problems, there have been several attempts at creating enhanced TCP implementations that work better in wireless environments, however, each solution is targeted towards specific network orientations/applications only.

A more general and universal solution is required.

Application layer coding for optimising lifetime in ad-hoc networks

PhD Student
D Georgoulas
Supervisor
Prof. K.J. Blow
Sponsor
EPSRC

The project will assess and optimise the use of application level coding in ad-hoc mobile networks and investigate the applicability of agent technologies towards that goal. By building example networks based on TinyOS and motes it will be possible to measure accurate energy use as a function of network traffic for different compression schemes and agent frameworks. Analytic models will be developed and network simulations will also be performed.

Please see my web site for further details.

Quality of Service (QoS) for Wireless Networks

PhD Student
S Mukherjee
Supervisor
Dr X. Peng
Sponsor
EPSRC

Currently Quality of Service is not provided over popular Ethernet based WLAN standards such as IEEE 802.11g and IEEE 802.11b. By providing QoS over the wireless medium we can guarantee service levels. This in turn will allow WLAN to be used for bandwidth critical applications such as Real-time Audio and Video Distribution. Possible future applications include Healthcare environments for monitoring, where a guaranteed bandwidth is essential. The research involves working with the IEEE 802.11e working group, simulation and prototyping using FPGA in addition to the investigation of other IP based protocols. The goal of the research is to innovate a QoS architecture appropriate for structured infrastructure deployments and adhoc networks.

Please see my user area for more details.

Adaptive Algortihms using FPGAs

PhD Student
M Mylonas
Supervisor
Prof. K.J. Blow
Sponsor
EPSRC

Investigation of the roles of dynamic hardware in enabling active networking at the physical layer. The project looks at the design and implementation of dynamic schemes in FPGAs.

Distributed processing, reconfigurable processes, and active network

PhD Student
Dr. Y Peng, completed 2004
Supervisor
Dr. D.J. Holding
Sponsor
Aston University

Study of Topology Management and Node Cluster Operation in Wireless Adhoc Sensor Networks

PhD Student
A.Y. Maaroof
Supervisor
Dr. D.J. Holding
Sponsor
Aston University

The aim of this project is to study node cluster operation in wireless adhoc sensor networks and the effect of clustering algorithms on overall network topology. A principal metric of concern is efficient use of node energy and how this can be improved amongst other metrics such as connectivity. Studies of existing clustering schemes such as geographic and dual radios are examined and improvements are suggested while maintaining the characteritics properties of distributed operation, self configuration and adaptability to network dynamics. Results obtained will form the basis of an improvement on existing clustering techniques in Adhoc sensor networks.

Masters projects

Heterogeneous Internet Protocol Networks (HIPNet)

MSc Student
Suro Avoseh
Supervisor
Dr. M.A. Eberhard
Sponsor

With the tremendous growth in IP technologies and the increase in wireless data traffic has lead to large telecommunication companies considering it for future carrier network backbone for converged networks. However, quality of service issues arises in a best effort networks. If IP technologies are to be successful in replacing legacy networks, there is a need to put in place architecture with protocols that could handled end-to-end same level of performance as seen in legacy networks. Internet Engineering Task Force (IETF) Session Description Protocol (SIP) used for signalling system for packet control in IP-based voice network has been adopted to be the signalling system to replace the International Telecommunication Union-Telecommunication (ITU-T) Signalling System Number 7 (SS7) in future core network. This is adopted by third generation partnership project (3GPP). Typically, the important indication of performance in telephony network is defined by the signalling or delay in call setup. This project evaluates performance of SIP signalling system in Voice over IP (VoIP) network as target toward the next generation networks (NGN). The results were carried from measurements of live SIP calls; simulate SIP traffic in Networks Simulation tool (NS-2) and modelling SIP signalling messages basic call flow in SIP traffic generator SIPp with SIP Express Router (SER). In particular, the project investigates end-to-end delay in SIP call setup, user agents' registration and performance characteristics of SIP signalling messages Interarrival time. Also, impacts of transport layer protocol on SIP traffic transmission, the behaviour of SIP traffic in overload/congested networks where SIP traffic alone is congested or competing with other background traffic. And investigation of SIP messages packet length properties.

Multimedia Service Characterisation

MSc Student
Abdukkabir Ajia
Supervisor
Dr. M.A. Eberhard and Dr S. Gresty (Ericsson)
Sponsor
Ericsson

The great push towards making IP the prime bearer of services, coupled with advances in chip technology which has enabled many people to make and transfer their own content is placing enormous pressure on current communications networks. There is therefore a huge amount of activities geared towards the design and development of IP-based Next Generation Networks (NGN) that will deliver multiple services (video, voice, gaming, and data). Multimedia Service Characterisation is a sub-project of a larger project-The Heterogeneous IP Networks (HIPNet) project that has amongst its aims, the creation of a model to characterise and validate multi-service NGNs for the UK. Multimedia Service Characterisation aims to characterise video and gaming services traffic as they are currently delivered over IP, develop models for them, and incorporate these models into Discrete Event Simulations in OPNET so that they can be used to test, model and develop these NGNs. This was done by measuring live traffic (Inter- arrival times (IAT) and packet lengths) of video and gaming data over the internet, analysing and processing this database of traffic using Microsoft Excel, processing the analysed data with C-scripts, Perl-scripts and a C-program, and plotting the results with Gnuplot. The measured data was subsequently modelled mathematically by fitting curves to their Cummulative Distribution Functions (CDF) and Quantiles (inverse CDF) using Matlab and Mathematica, before being implemented in OPNET. Some activities were also dedicated to investigating the effects of network impairments on video services by sending real video packets across a simulation environment, and observing the effects of network impairments introduced on the video packets in the simulation, at the other end.

QoS for Ad-Hoc Wireless Networks

MSc Student
H Haque
Supervisor
Dr X. Peng
Sponsor
Aston University

An ad-hoc network is a dynamic wireless network that is established by a set of mobile nodes interconnected by multihop communication paths. It has no fixed network infrastructure or administrative support. The topology of the network changes as mobile nodes join or depart the network or radio links between nodes are unavailable. How to ensure quality of service (QoS) in such a complex environment is a challenging issue for future generations of wireless communication systems. In this project, based on a survey of the latest developments in this area, investigation and testing on the QoS related performance such as delay, throughput and bandwidth requirement will be carried out within a wireless LAN for both ad hoc and infrastructure modes. Different network configurations in terms of network topology, traffic patterns and number of users will be considered in the tests, and the results obtained will be properly analysed.

Lean Sensor Node Protocol Stack

MSc Student
P Kumar (MSc Project 2003)

To design and demonstrate an energy-efficient and lean communication protocol stack for sensor networks.

Employable Graduates; Exploitable Research