1- Reducing Microwave Radiation for Mobile Phone Handsets and Base Stations
Basic Specifications:
An important current issue of major concern to users
and providers of mobile phone systems is the effect of microwave radiations
from mobile phone handsets and base station masts on the people. The aim
of this project is a critical study of the various method of reducing the
harmful effects of radiation: from the use of CDMA where theoretically
the energy can be thinly spread over a relatively larger bandwidth, to
the use of beam forming antennas and the polymer radiation absorption materials.
The specific objectives are the followings:
1. Literature survey and study of state of the art mobile
phone systems and antenna systems, their radiation levels and propagation/absorption
patterns, and the effect on human body.
2. A study of WCDMA and its impact on the level
of microwave radiation
3. Methods of designing handset and base-station antennas
to reduce radiations.
2- Improving the Quality and Intelligibility of Speech From Mobile Cellular Telephones in Moving Car/Train Environments
Basic Specifications:
The objective of this project is to improve the perceived
quality and intelligibility of noisy speech from a car environment. Improving
the speech quality requires that the natural’ temporal-spectral composition
of speech be restored. The subband approach provides a suitable platform
for restoring the spectral-temporal composition of the speech signal. Furthermore
the subband approach can be used to develop methods for utilising the auditory
masking effect and the psycho-acoustics of hearing for the perceptual processing
of the noisy speech.
The main objectives of this project are the followings:
(a) Further development of a subband linear prediction-based
perceptual noise reduction method, and its application to enhancement of
noisy speech from mobile car environments.
(b) Analysis of the effects of noise reduction
on the perceived spectral-temporal composition of elementary phonetic sounds,
and its relation to the psycho-acoustics of hearing.
Extension:
Development of perceptual-state based methods for restoration
of the perceived quality of speech sounds.
3- Sub-band Acoustic Echo Cancellation for Hands-free and Mobile Phones
Basic Specifications:
A well-known problem with hands-free telephones, teleconference
systems, public address systems, mobile phones, and hearing aids, is the
acoustic feedback coupling of the sound waves between the loudspeakers
and the microphones. Acoustic echo can result from a combination of direct
acoustic coupling, and multi-path effect where the sound wave is reflected
from various surfaces and then picked up by the microphone. In its worst
case, an acoustic feedback can result in howling if a significant proportion
of the sound energy, transmitted by the speaker, is received back at the
microphone and circulated in the feedback loop. The overall round-gain
of an acoustic feedback loop depends on the frequency responses of the
electrical and the acoustic signal paths.
The main objectives of this project are the followings:
(a) Implemenation of a sub-band acoustic canceller.
(b) Comparative investigation of the use of RLS and LMS
adaptation methods.
(c) Investigation of the optimal number of sub-bands
and the effect of using linearly-spaced sub-bands, and critical auditory
sub-bands.
(d) Implementation and evaluation for a hands-free PC
telephony.
Extension
Application of the methods to mobile phone acoustic echo.
4- Automatic Director Enquiry Using Speech Recognition and Text to Speech Synthesis
Basic Specifications:
Name dialling is already used in many mobile handsets.
The aim of this project is to use the Text to speech synthesis and speech
recognition systems that we have developed in our lab to develop a simple
automatic director enquiry system.
The specific objectives are the followings:
1. Literature survey and study of state of the
art text to speech synthesis and speech recognition systems.
2. Integration of the TTS system with speech recognition
for name dialling.
3. Evaluation of speaker-dependant and speaker-independent
recognition systems.
Extension:
Implementation of a system for use in our laboratory.
5- Talking Head: Face and Voice Animation
Basic Specifications:
The next generation of multimedia communication systems
will have user friendly interface integrating voice and facial expressions.
The objective of this project is to design a talking-head integrating face
animation and text to speech synthesis. The applications of such a system
include user friendly voice based access to IT, broadcast studio technology,
animation in FILMS, entertainment etc. We already have a text to speech
synthesiser talking in various voices including that of Michael Palin’s,
and a voice mimic system that can imitate voices. We also have software
available for integration of voice and face.
The specific objectives are the followings:
1. Literature survey and study of state of the
art text to speech synthesis and face animation systems.
2. Integration of the TTS system with face animation.
3. Synchronisation of lip movements and speech
sounds.
Extension:
Synthesis of emotions such as happy face/voice, sad face/voice.
6- Separation of Music and Songs with Application to Hi-Fi Karaoke System
Basic Specifications:
The aim of this project is to develop a digital system
for separation of song and music with application to Hi Fi Karaoke. The
system will decompose musical signals into a number of subbands and then
estimate the short-time powers of the music content and the song content
in each subband.
The specific objectives are the followings:
1. Literature survey and study of the state-based
statistical classification and estimation methods.
2. Decomposition of signals into a number of subbands.
3. A study of separability via experimentation
on digitally mixed music and song.
4. Estimation of the time-varying powers of the
music and the song in each subband.
5. Development of a signal separation method.
Extensions:
1. Mean opinion score analysis of the system output.
2. Cost analysis of an affordable mass-market
implementation.
7- WaterMarking of Digital Images
Basic Specifications:
The objective of this project is to develop watermarking
techniques to prevent unauthorised copying of digital image materials.
The strategy is to mix an ‘invisible’ copyright signal with the image signal.
Digital correlation can be used to prove a given image comes from a certain
source.
The specific objectives are the followings:
1. Literature survey and study of image processing
and digital watermarking .
2. Study of the visual masking phenomenon.
3. Development of a spectrally shaped copyright
signal for mixing with an image signal.
4. Development of watermarking matching methods.
Extensions:
1. Examine the system reliability.
2. The effect of coding/transmission on the watermarking
signal.
8 - Dialogue System for the Millennium Homes Project Using Speech Recognition and Text to Speech Synthesis
Basic Specifications:
“Millennium Homes” is aiming to develop technology to support frail older people who wish to live independently in their own homes. Researchers at Brunel are developing technology to monitor activity within a person’s home and raise warnings if problems are detected (e.g falls, hazards). When problems are detected a dialogue system will be activated to communicate with the home occupant (to check whether they are OK or to issue instructions). One mode of communication will be via voice (speech output and input), which could be implemented either via the telephone, or to work in the general home environment via far field microphones and loudspeakers. The aim of this project is to use the Text to speech synthesis and speech recognition systems that we have developed in our lab to develop a dialogue system for the Millennium Home.
The specific objectives are the
followings:
1. Literature survey and
study of state of the art text to speech synthesis and speech recognition
systems.
2. Integration of the TTS
system with speech recognition for the Millennium Homes application, to
work via a telephone connection.
3. Evaluation of the success
of the speech recognition system with alternative vocabulary sets / grammars.
Extension:
1. Implementation of a system
for use in the Millennium Home Pilot site on campus.
2. Investigate the viability
of implementing the dialogue system via far field microphones and
loudspeakers.
9 - Stereo Music Coding for Mobile Applications.
Basic Specifications:
The aim of this project is to model and utilize the correlations
in the left
and right channels of a stereo music recording for music
compression. In
particular the focus is on the adaptation and the use
of methods such as
(code excited linear prediction coding) which have been
successfully applied
in enhanced low bit rate speech coding.
The specific objectives are the followings:
1. Literature survey and study of state of the
art music coding methods.
2. A critical study of Transform coders versus
model based coders for music
coding.
3. The modeling of interchannel entropy and correlation.
4. Development and implementation of a coder including
correlation models,
probability models and perceptual models.
5. Perceptual evaluations of the coded music.
Extension:
1. Interchannel coding for efficient error control
coding.