Investigation of the Sequential Accelerator on the Perceptron for Pattern Recognition

In machine learning methods, when the input data becomes extremely large, the current direct methods
require too large learning times and memory. This project investigates one sequential method to overcome this
problem. It is quite a simple method to implement and is tested using the Perceptron as the base classifier. The
perceptron converges very slowly. So it will be interesting to find out if the proposed accelerator can improve
significantly the computational times of this simple classifier. The student will try various investigations on
different very large data sets and to measure their computational complexities. It has been shown that this
sequential method is very fast and only need a small subset of the large data set to complete the learning.

Investigation of the Sequential Accelerator on LDA for Pattern Recognition

In machine learning methods, when the input data becomes extremely large, the current direct methods
require too large learning times and memory. This project investigates one sequential method to overcome this
problem. It is quite a simple method to implement and is tested using the well-known LDA classifier. The student
will try various investigations on different very large data sets and to measure their computational complexities. It
has been shown that this sequential method is very fast and only need a small subset of the large data set to
complete the learning.

Equalization of Fading Channels using RBF Networks

Equalization of Fading Channels using RBF Networks
Summary: Equalization is the process of recovering the true input data from the received data which is corrupted
by noise and passes through a nonlinear communication channel. Earlier work by the supervisor and his group has
successfully used RBF neural networks for complicated nonlinear channels that are stationary. In this project, the
performance of the RBF equalizers will be investigated for time varying ( fading ) channels ( mainly for Raleigh
Fading Channels) and an implementation scheme for the equalizer will be evolved. Performance comparison with
other conventional equalizers will also be made.

Hierarchical K-Winner Machines with Fuzzy Memberships

The aim of this project is to develop a hierarchical K-winner machine classification system with fuzzy
memberships in the highest level. The K-winner machine makes use of supervised and unsupervised learning
techniques. The hierarchical approach allows us to improve the classification speed during the application stage
and at the same time finely partitions the pattern space in the highest levels. In order to further improve the
performance of the hierarchical K-winner machine, we introduce the fuzzy membership assignment to the
prototype vectors. The developed system will be tested on some standard datasets and its performance will be
compared against competing approaches.

Generalized Sidelobe Canceller (Gsc) for Wireless Communications

The Generalized Sidelobe Canceller (GSC) has been a useful tool in spatial filtering. The GSC structure is
also widely used for realizing adaptive array beamformer for communication systems. Many researchers have
proposed modifications to the GSC structure This project involves the design of the quiescent beam pattern,
adaptive nulling of interference and the technique of designing the blocking matrix for the adaptive portion of the
GSC structure. Students with strong analytical skills, Mathematics and proficient in Matlab programming are
preferred.

Portable Fluorescence Detection System,Embedded System Project

Integration of Micropump, Microfluidic Card, Optical Detection and Data Acquisition. The objective of
this project is to construct a prototype of portable fluorescence detection system for microfluidic chip. It involves
the integration of a micropump, microfluidic chip, optical detection, one-dimensional scanning system, and data
acquisition. Micropump and scanning system will be controlled by a PC, laptop or microcontroller. The output of
optical detection will be amplified, stored in memory, processed and displayed. Micropump and microfluidic chip will be made from polymeric materials. Photodiode or avalanche photodiode will be used as optical detector. Required background knowledge: Photonics and electronics, particularly data acquisition and controlling either using PC or microcontroller. Please contact Dr Rudi Irawan (erirawan@ntu.edu.sg) to discuss if you are interested.

System Prototype for Personal Mobile Health Monitoring

Infocommunication technology has been employed in health care for many years with good success.
However, the upgrade of existing medical instruments and the design of new medical applications as a result of continuous advances in information technology should not lead to a neglect of the practical needs of patients and physicians. Although significant progress has been made with respect to the creation of new medical instruments, it is also important to direct efforts to the integration of these instruments in operational information systems where all the devices can be integrated into a single framework of health-care resources. It should be observed that this aspect is particularly critical since different medical devices and hospital data bases using different protocols and data representations may be unable to interact automatically, thus failing to provide efficient diagnostic support. Although the current technologies offer the necessary means to support this type of health care, it should be possible to integrate selected health care resources offering a continuous, cooperative health care system and tools for personalized health monitoring.
This project focuses on the development of a versatile system prototype for personal mobile health monitoring on mobile phones operating on Symbian OS including the use of open standards such as Bluetooth, GPRS, etc. The aim is to achieve an optimized system design for efficient data acquisition and online analysis for a 3-stage distributed mobile system that consists of the front-end integrated biosensor, the personal mobile phone and the back-end server. The primary goal will be on the data storage and processing of ECG and PCG signals. Requirements: Wireless Communications, Embedded Systems, Programming (Symbian OS, C++).

Analysis of the quality of fingerprint image features

The most popular fingerprint feature used for matching is the feature points, called minutiae. Several approaches have been proposed which in a way, involves detecting the skeleton image to locate the minutiae. Subsequently, the geometrical and other information are extracted and used to match the minutiae. However, the performance of the minutiae extraction varies. Problem arises when there is noise or the fingerprint image is not clear. This could cause the introduction of false minutiae or the omission of valid minutiae. In this project, the main aim is to devise methods to analyse the quality of the minutiae so that the overall matching is improved. This
will involve analysing the property of the minutiae to provide suitable confidence level and providing consistent information that can be extracted from the minutiae. Good knowledge in C/C++ or Matlab programming is essential for this project.

Neural Network based Color Image Segmentation

Breast cancer is the most common cancer among women, and is the second leading cause of cancer
deaths in women today. Basically, the diagnosis procedure of breast cancer consists of two steps; (1) mammography
based breast abnormality detection; (2) biopsy based diagnosis. Biopsy is the only definitive way to determine
whether cancer is present.
In this project, biopsy color Image segmentation based on neural networks will be studied. The objective of breast
biopsy image segmentation is to segment cells and blood vessels in the image.
As the second part of the project, color image segmentation software will be developed.

Fast Image Synthesis of 2D Infrared Facial Images

Pose determination of human faces plays an important role in face recognition. The traditional way of
dealing with pose-variation is to use a number of representative images in different poses. In this project, we will
synthesize virtual images at different pose using few reference images in the pose space. Fast methods to find
correspondence between reference images will be explored. To deal with the intensity variation of IR images
among individuals and to further deduce computation, modeling method will be adopted. The synthesized IR
images can be used to interpolate virtual views between real views to provide more “samples” or to produce a
standard frontal view for recognition.

Development of a Matlab-based Control System Laboratory

The aim of this project is to develop a Computer Aided Control System Design (CASCD) environment that
are typically utilised in an undergraduate controls laboratory. Due to their popularity and availability, MATLAB,
SIMULINK and the Real-Time Workshop toolbox are chosen as the prototyping environment. The following issues
should be addressed:

1. Standardisation: a consistent hardware interface to various laboratory apparatus and a consistent user
interface, for the following task: modelling, control design, data collection, parameter estimation, and real-time
experiment.

2. Control experiment via Internet: While the local real-time control could be extended for remote control via the
Internet, there will be some issues that are peculiar to Internet lab, e.g. all Internet experiments need to be selfresetting.
In addition, safety, security and user flexibilty are issues that need to be addressed 

Wavelet-based Deconvolution for System Identification

In many practical applications, we are given access to the input and ouput of a system whose
characteristics are unknown. The term ‘deconvolution’ is used to describe the operation of separating the input
function from the characteristics of the system we intend to identify. Conventional techniques for deconvolution
based on direct Fourier or Laplace transfoms suffer from the ill-conditioned nature of the problem in the presence
of noise. The aim of this project is to develop and implement a robust deconvolution technique based on wavelets
and study its performance as compared with other conventional techniques. An appreciation of Matlab and
numerical computation skills would be desirable.

Active Contours for Image Segmentation

Active contour models are physics-based deformable models used to model the appearance and behavior
of a physical object being imaged, or to simulate some image analysis task. Using a Larangian formulation of the
energy functional, active contour models can be adopted for image segmentation by the design of appropriate
energy terms and the subsequent minimization of the total energy. The aim of this project is to develop and
implement appropriate active contour schemes for the segmentation of noisy images. An appreciation of Matlab
and numerical computation skills would be desirable.

Spline Interpolation

The objective of this project is to automatically extract both 3D range data and 2D intensity image from
a stripe-light range camera. By texture mapping the two data formats, a realistic 3D textured view of the imaged
object can be obtained. Usually the 3D range data contains a significant level of noise (both random and impulse
noises) which have to be culled away before the data can be useful. In addition, there will also be many missing
data points which have to be recovered through interpolation. In this respect, the candidate should be firmly
grounded in mathematical concepts of splines (cubic, bezier curbes, NURBS) as well as a good working knowledge
on Visual C/C++ programming.
Project Nature: Software base

Model Predictive Control (MPC) on a Chip Embedded project

Model Predictive Control (MPC) has become an established control technology in the petrochemical
industry. Its use is currently being pioneered in an increasingly wide range of high bandwidth applications, such as
ships, aerospace, road vehicles and “Lab-on-Chip” devices. MPC outperforms other control strategies through its
ability to deal with constraints. This requires on-line optimization, hence computational complexity can become an
issue when applying MPC to complex systems with fast response times or to embedded applications where
computational resource may be a major constraints. We are seeking students with suitable background and interest
to help us realize the vision of ``MPC on a Chip'', which is also an ASTAR research programme.
There are several projects along this line of research:
1. To develop a scalable and modular Matlab/Simulink model of constrained MPC algorithm
2. To develop a C model of (1) based on the interior point method
3. To develop a C model of (1) based on the active set method
4. To implement (2) and (3) on a suitable embedded processor or FPGA platform.
5. To exploit parallelism and/or multiple date width model to achieve area-time efficient FPGA implementation.
Students who wish to be considered for the above projects should have good programming expertise and are
interested in mathematical algorithms. Experiences with embedded control or FPGA would be an advantage.

Protect your mobile phone from unauthorised use

MOBILE SHIELD
Protect your mobile phone from unauthorised use or theft using this simple circuit. It can generate a loud chirping sound when somebody attempts to take away the mobile handset. The added feature is that
the circuit also works as a mobile charger. The circuit is powered by a step-down transformer X1 with rectifier diodes D1 and D2 and filter capacitor C1. Regulator IC 7812 (IC1) along with noise filter capacitors C2 and C3 provides regulated power supply. The circuit utilises two NE555 timer ICs: One as a simple astable multivibrator (IC2) and the second as a monostable (IC3). The astable multivibrator has timing resistors R1 and R2 but no timing capacitor as it works with stray capacitance. Its pins 6 and 2 are directly connected to a protecting shield made up of 10cm×10cm copper-clad board. The inherent stray capacitance of the circuit is sufficient to given an output frequency of about 25 kHz with R1 and R2. This arrangement provides greater sensitivity and enables the circuit with hand capacitance effect. Output pulses from the oscillator are directly given to trigger pin 2 of the monostable. The monostable uses a low-value capacitor C6, resistors R3 and preset VR1 for timing. The output frequency of the monostable is adjusted using preset
VR1 such that it is slightly less than that of the astable. This makes the circuit standby, when there is no
hand capacitance present. So in the standby mode, the astable’s output will be low. This makes the trigger
input of monostable low and output high. The warning LED1 and buzzer are connected such that they become active only when the output of the monostable sinks current. In the standby state, the LED1 remains ‘off’
and the buzzer is silent. As somebody tries to take the mobile phone from the protecting shield, his hand comes near the shield or makes contact with the shield, which introduces hand capacitance in the circuit. As a result, the astable’s frequency changes, which makes the trigger pin of the monostable low and its output oscillates. This produces chirping sound from the buzzer and also makes the  LED1 blink. The circuit can also be used as a mobile charger. It provides output of 6V at 180 mA through regulator IC 7806 (IC4) and resistor R5 for charging the mobile phone. Diode D3 protects the output from polarity reversal.
The circuit can be wired on a common PCB. Enclose it in a suitable case with provision for charger output

  Comment for circuit diagram

A Highly Efficient DC Lamp Dimmer

The simplest lamp dimmer circuit consists of a rheostat, in series with the lamp, which one
may adjust to obtain the required brightness. Such linear regulators are quite inefficient since
a lot of power is wasted in them. Moreover, in the rheostat the moving contacts are likely to
get damaged in the long run, as its value is frequently adjusted by moving the slider. Such
linear control circuits provide an overall efficiency of no more than 50 per cent. This wastage
of power can be avoided if one uses pulse width modulation (PWM) which can be made to
control an electronic rheostat. The circuit shown here is based on PWM principle. Gate N1
and its associated components constitute an oscillator producing oscillations of approximately
200 Hz with a pulse width of 0.1 ms. This output is fed to transistor T1 for level shifting. At the
output of this transistor is a potentiometer VR2, using which a DC component can be added
to the pulses emerging from transistor T1. By adjusting this potentiometer/trimmer, one can
have a good linear control of the lamp brightness from completely off state to 100 per cent on
state. The signal is inverted by gate N2 and fed to MOSFET 12N10. IC CD40106 provides six
inverting buffers with Schmitt trigger action. The buffers are capable of transforming slowly
changing input signals into sharply defined jitter-free output signals. They are usually used as
wave and pulse shapers. IC CD40106 possesses high immunity and low power consumption
of standard CMOS ICs along with the ability to drive 10 LS-TTL loads. In this circuit loads up
to 24W can be connected between MOSFET drain and 12V supply without using a heatsink.
The loads can even be DC motors, miniature heating elements, etc. If one uses a low RDS
(on) MOSFET, a higher efficiency can be achieved. By using the components as shown in the
circuit, an efficiency of approximately 95 per cent can be achieved. The flexibility of the design
makes it possible to change the MOSFET with a similar one, in case of non-availability of
12N10. The circuit by itself does not draw much current when the load is disconnected.
Ensure proper ESD protection while handling the MOSFET to prevent damage. Lab note: The
circuit was tested using MOSFET IRF640 with RDS (on)=0.18 ohm.