University: The University of Birmingham
Name of Sponsoring Professor: Dr Theodoros N. Arvanitis
Department: Electronic, Electrical & Computer Engineering

Postgraduate Research with OPNET

1. VBR Traffic Source Modelling & Statistical Multiplexing for High Speed Networks

Broadband communications networks are expected to support a wide range of multimedia applications including entertainment video on demand (VOD) high definition TV (HDTV), and multimedia teleconferencing. Nowadays we are all aware of multimedia: They consist of voice, data and video all in the same application. These applications generate video and audio streams that have to be transported in a timely manner to ensure coherent reception and playback at the receiver. What we must be aware when dealing with an issue like that is that when multimedia are transmitted over a network - whether it is a fixed or a wireless one - the vast majority of the consumed bandwidth is because of the video traffic. The emitted video streams are typically compressed before being transported to the network. The data compression is required to reduce the storage and bandwidth requirements of the digital video. The compression’s efficiency depends on the video dynamics as well as the underlying compression technique. For constant-quality video, the video encoder generates a sequence of variable-size compressed frames. When the frame generation rate is constant (e.g., 30 frames/s in the NTSC format), the output of the encoder constitutes a variable bit rate (VBR) stream. 

Transporting VBR-coded video streams while guaranteeing a required level of quality of service (QoS) is a challenging problem that has received a lot of attention in recent years. Conventional data networks (e.g., the current Internet) have not been designed to transport traffic streams with QoS requirements and so their performance is poor. While asynchronous transfer mode (ATM) networks are more suited to real-time and guaranteed QoS communications, the complexity of VBR video combined with the diversity of its QoS requirements make it difficult to transport video traffic in a cost-effective manner. This has been a major hindrance facing the economic viability of digital video services over computer networks. Unless efficient bandwidth allocation approaches are devised, the cost of these services will prevent their widespread acceptance among potential customers. 

It is predicted that in the near future the transport of compressed video will pervade computer networks. Variable Bit - Rate (VBR) encoded video is expected to become a significant source of network traffic due to its advantages in statistical multiplexing, gain and consistent video quality. Both system analysts and developers need to assess and study the impact these sources will have on their networks and networking products.

2. Intelligent Portable Camera Network Node

Future digital cameras will be seamlessly networked to enable transfer, distributed processing and storage of images in the network. The network must be capable to manage the users, track the users usage and bill them accordingly in an efficient manner. Security and authentication of the images also has to be taken in to consideration.

The work will be to investigate how the processing of the images can be distributed intelligently between the camera, gateway nodes and the active nodes in the network to reach a balanced processing load between the different nodes. The new proposed architecture is simulated and verified in OPNET.

3. Routing Algorithm for Mobile Networks based on the Small World Approach

With the Mobile IP, the mobile nodes are allowed to send and receive data despite their current point of connection to the Internet. But mobile nodes have to report their every movement in the foreign network to their home networks. This causes huge amount of signalling traffic and disturbing latency during handoffs.

The work investigates how the traditional IP addressing and routing affect networks under mobile environment. It also identifies the problem facing Mobile IP as well as a proposed work on routing protocol based on Small world approach. OPNET is used for simulation and measurements.

4. Swarm Intelligence Based Route Discovery in Mobile Ad Hoc Networks

Mobile Ad-Hoc Networks (MANETs) are wireless communication networks that allow users to access and exchange information regardless of their geographic position. In contrast to the infra-structured networks, all nodes in MANETs are mobile and their connection is performed dynamically in an arbitrary way. Thus, unlike other mobile networks MANETs are equipped with no fixed infrastructure, an advantage, which offers a decentralised character to the network. This feature results in making the networks more flexible and more useful in certain applications such as emergency search-and-rescue operations and cases in which users wish to exchange information and use services carried out in inhospitable terrain.

A very important issue raised with these networks relates to the determination of the rules (protocols) governing the communication between the entities (nodes) constituting the network. More specifically, one important question is how to facilitate the dynamic discovery of the most efficient route between two nodes within the network, while taking into account the mobility of the nodes and the lack of a fixed topology in the network.

This research project aims to exploit concepts from a new form of artificial intelligence, swarm intelligence to the application of dynamic route discovery in MANETs. The swarm intelligence concept exploits principles of self-organisation in biological systems to describe artificial systems as complex self-organising adaptive wholes. Such “intelligent” systems exploit principles of autonomy, emergence, and distributed functioning to replace control, pre-programming and centralisation.

5. Modelling network infrastructure implementation and performance evaluation for PACS: DICOM over Ethernet-based TCP/IP

6. Inter-Vehicle Communication Based on Mobile Ad Hoc Networks

 The term Intelligent Transportation Systems (ITS) describes a collective approach to the problem of enhancing safety, mobility and vehicular traffic handling capacity, improving travel conditions. Communications requirements are central to any ITS infrastructure. An important subset of the communication requirements of an ITS is inter-vehicle communication (IVC), which is essential in realising certain ITS applications such as vehicle cooperative driving, the dissemination of incident warning data and traffic congestion.

Mobile cellular communication systems are not suited for IVC for a number of reasons. Firstly, communication with a fixed infrastructure would introduce unacceptable delays for safety-critical applications. Secondly, there is insufficient capacity to meet the anticipated demands of the ITS applications. Thirdly, as IVC applications require frequent transmissions, this would be extremely costly using a cellular approach. A dedicated cellular system for IVC would not solve the above problems and would also incur the significant added overheads of cost and time in installing the infrastructure.

In order to meet the requirements of IVC a solution is to implement a distributed communication network between the vehicles on the road. Mobile ad hoc networking is a candidate technology that will be investigated within this research, particularly the routing protocol requirements.

Having investigated the requirements of various ITS application scenarios, it became apparent that many of the applications have varying requirements such as QoS and differing regions to which the data is relevant.  With this in mind this research focuses on the development of a routing protocol framework to enable communication between vehicles using a distributed network formed between vehicles. Given the dynamic nature of vehicular traffic flow, vehicles will be addressed both by location and a unique ID. The algorithmic complexity and maintenance overheads of a distributed positioning database within the vehicular environment are considered to be highly costly. A geographic forwarding scheme will form part of the routing framework along with the knowledge of local connectivity, which is adaptive depending on the application requirement.

The routing protocol framework is being developed using the network simulator OPNET. OPNET lends itself to layered protocol architecture development and is, therefore, being used in this research in order to develop a routing protocol framework for the vehicular environment. Although this research focuses on routing protocol development, a MAC model will also be developed which is “fit-for-purpose.” The mobility assigned to each node (vehicle) during simulation within OPNET is an important factor in the performance of the protocol. Each node (vehicle) is assigned a trajectory, which is generated from a microscopic traffic simulator.  This will provide realistic node movement, which emulates the nodal movement of the target environment for which the framework is being developed, allowing assessment of the routing protocol under various driving scenarios for a variety of ITS applications.and scalability

7. State sharing algorithm for network adaptivity

There are two features in modern communication systems. Firstly, the distribution of traffic loads and network topologies may vary from nearly static to very dynamic. Secondly, the number of nodes and links in a network may vary in orders of magnitude. Consequently, network algorithms are desired to be adaptive and scalable. 

People are willing to pay some resources for network state sharing, in order to increase system adaptivity. Such overhead used to be treated as negligible, such as in popular quasi-static routing protocols (e.g., OSPF). However, with the increase of the size and dynamics of networks, the cost of state sharing becomes a performance bottleneck.

This research project aims to exploit the relationship between adaptivity and routing performance in a scalable means. A new state-sharing algorithm is being developed, as well as an OPNET based simulation test bed of dynamic large-scale mesh network.

Published papers

1. T. N. Arvanitis and D Roth, Modelling Network Infrastructure and Performance Evaluation for PACS:  DICOM over Ethernet-based TCP/IP, SPIE Proceedings Medical Imaging, 2001. 

2.  D. Topham, D. Ward, T.N.Arvanitis, and C.C.Constantinou, Inter-Vehicle Communications Based on Mobile Ad Hoc Networks, First International Conference for Sensors, Navigation and Communications for Vehicles Telematics (VehCom 2003), pp.135-141, 2003.

3. D. A. Topham, D. D. Ward, J. Wu, T. N. Arvanitis and C. C. Constantinou, Routing Framework for Vehicular Ad Hoc networks: Regional Dissemination of Data Using a Directional Restricted Broadcasting Technique, In Proceedings of 2^nd International Workshop on Intelligent Transportation, WIT 2005, March 2005, Hamburg, Germany, pp.23-28, 2005.

4. D. D. Ward, D. A. Topham, C. C. Constantinou and T. N. Arvanitis, Developments in Vehicle-to-Vehicle Communications, In Proceedings of 9^th International Forum on Advanced Microsystems for Automotive Applications, March 2005, Berlin, Germany, pp.353-370, 2005.   

5. T. N. Arvanitis, Y. Sun, C. C. Constantinou, A. S. Stepanenko, B. Liu, and K. Baughan, Network Visualisation and Analysis Tool based on Logical Network Abridgment, In Proceedings of MILCOM2005 – Innovation…Fueling the transformation, IEEE/SAIC, Atlantic City: New Jersey, paper 1092, 2005.  

Last updated 7/03/2006