Closed-loop Connection Admission Control Scheme

Connection Admission Control (CAC) is a preventive traffic and congestion control mechanism. It determines whether or not a network should admit a new connection request. CAC is essential for providing Quality of Service (QoS) supports required by multimedia applications, which are becoming the most important revenue-generating sector in telecommunication services.  The complex statistical characteristics of traffic present big challenges for CAC. If a CAC scheme is based on a complex traffic model with too many parameters, both the performance analysis and obtaining accurate model parameters from either user or online measurements are difficult. On the other hand, simple traffic model often cannot fully characterize the traffic.

Currently almost all CAC schemes are open loop. Its performance relies on accurate traffic and network models, accurate model parameters, and accurate loss performance analysis. As a quick review of the existing literature indicates, no traffic model can be claimed to be accurate for all traffic sources (e.g. voice traffic, data traffic, video traffic) under all network conditions (e.g. heavy traffic or light traffic, LAN or WAN). Modelling errors, traffic parameter errors and errors in performance analysis are inevitable. An open-loop CAC scheme lacks the ability to account for these errors and adapt to real network environment to achieve an optimum performance.

This project will investigate a novel closed-loop CAC scheme where performance feedback is provided to CAC scheme to enable it to adapt to real network environment. The closed-loop CAC is able to overcome inherent drawbacks in an open-loop CAC to achieve better performance in a complex, changing network environment.

 Network Quality of Service (QoS) Management

Connection Admission Control (CAC) is a preventive congestion control mechanism. It determines whether or not a network should accept a new connection. A new connection can only be admitted if the network has enough resources to accommodate the new connection without violating the Quality-of-Service commitment already made by the network to existing connections. This project aims to investigate the statistical multiplexing effect of traffic sources. Considering the difficulty for the user to tightly characterize such parameters as mean cell rate and maximum burst size, and the difficulty for the network to accurately measure such parameters as instantaneous traffic rate and autocorrelation functions, this project will investigate a hybrid dynamic CAC scheme employing only simple traffic parameters that can be obtained reliably and easily from traffic descriptors and on-line measurements, while at the same time capable of achieving robust QoS guarantees and high bandwidth utilization.

One of the most critical elements within the optical transport networks based on wavelength-division-multiplexing is the optical cross-connect. In the near future this optically routing device will be in charge of network management in the optical layer, with potential throughput of terabits per second. At the core of many proposed optical cross-connect architectures is a set of optical space switch capable of switching a large amount of input and output fibers. This project will investigate the design and implementation of an intelligent reconfigurable optical switch with a large number of input and output fibers.

Gyroscope Subsystem and Data Acquisition & Supervision Subsystem of Attitude and Heading Reference System (AHRS)

AHRS is widely used on ships, submarines, vehicles and airplanes for navigation purpose. This project aims to design a navigation system for naval vessels, it also provides mathematical platform for weapon system in naval vessels. Gyroscopes and linear accelerators are used as basic sensors for the system. Gyroscopes are used to sense angular velocity. Linear accelerators are used to sense linear accelerate. The output of gyroscopes and linear accelerators are current signals and voltage signals respectively. These output signals are passed to A/D and I/F converters in which they are converted to digital format. The digital signals are then fed into the microprocessors. Complex mathematic calculations are performed in the microprocessor, which include noise handling, drift handling, attitude matrix calculation, control of gyroscopes, etc. The output signals are attitude matrix of naval vessels and control signals which are then used for control of the system, including control of the gyroscopes. 

Multi-channel High-Precision Data Acquisition System

This project aims to investigate a multi-channel high precision measurement instrument. The instrument can be used to measure voltage with an accuracy of 1uV. In combination with pressure sensor, temperature sensor, etc. the instrument can be easily adapted to measure these physical quantities by modifying its software.

© 2005 Last revised November 16, 2007