"If we knew what it was we were doing, it would not be called research, would it?"

~ Albert Einstien

 

Ever since I became sensible, I was facinated with integrated circuit design. While working towards a Diploma in Electronic Engineering, I have opt for integrated circuit design and also worked on a Z80 microcontroller project. During my undergraduate degree, I have researched into analogue CMOS VLSI inplementation of fuzzy logic circuits. I was later given the opportunity to continue a PhD in the CMOS design of a novel neural fuzzy chip. However, as unpredictable as life can be, I am now working on antenna design for the Passive Picocell systems, funded by British Telecoms. Well.... if I have an opportunity, I might just go back to ANN-Fuzzy CMOS designs or perhaps try something new, like MEMs (microelectrochemical systems) devices.

You are visitor number:

|What is antenna design | PhD project | Undergraduate project |

 

What is antenna design

.......... It is a black art!!!!

Back to top

 

 

PhD project

 

The recent explosion in the communication market has resulted in the emerging of a number of systems operating at different frequencies. Currently, competition in the market aims at providing multiple services within a single communication terminal unit. This sets out the need to design radio links (i.e. antenna) that will not only cope with current existing systems but also be operational for future emerging systems. At the same time, terminal unit also receives tremendous attention in providing complexity and value added functions whist maintaining portability and compactness in size. The design of small antenna, unlike other electronic components such as semiconductors, capacitors and resistors, faces difficulty in realistically miniaturising it. Compromises are often inevitable between different aspects of the antennas� performance and its geometry. The fast emerging application and compactness issues have steered many researchers in the design of antennas for not only multi-band or wideband operation but to also consider integration with the terminal unit (i.e. the radio module). Limitation with small antennas has always been a major challenge for researchers to provide complete coverage of the required radio band with efficient performance.

The aim of this PhD study is to design an integrated antenna system for the passive picocell system. In the course of the research, various antenna types will be assessed for this application and a design on a tentative specification of 8dBi gain from 900MHz to 6GHz will be produced with priority to operating bands of GSM, DECT and Wireless radioLan. Physically small size antennas will be an important research issue aimed in overcoming its inherently small bandwidth and gain. Methods of monolithically integrating the antenna with the EAM on a common semiconductor substrate and development in monolithic microwave integrated circuit(MMIC) will be assessed. This project is supported by British Telecommunication

Back to top

 

 

Undergraduate project (1997)

 

In recent years, many approach to fuzzy processor has been greatly escalated due to the advancement of VLSI technology. Both digital, analogue and mixed mode implementations are equally investigated for specific and general purpose application. It was until recently that alternatives to defuzzification in analogue implementation has been proposed. This is to use a feedback loop to force the divisor for the center of gravity computation to unity.

In this project, several building blocks suitable for fuzzy processing has been identified. 2 architectures for implementing a general fuzzy processor has been proposed and verified through simulation results. Both architectures are capable of parallel fuzzy processing. The defuzzification scheme employs the Normalisation Lock Loop (NLL) technique proposed. A novel inverted membership function circuit suitable for the NLL defuzzification architecture was also designed.

Finally a general purpose fuzzy inference processor has been design. The simplicity of this architecture, circuitry and implementation makes it an attractive processing core for any analogue fuzzy processor. The final ccircuit was design on a 2um by 2um CMOS, N-well process.

Back to top