|
Research
Grants
Principal
Investigator: Cooperative hybrid intelligent agents for real-time
management of traffic networks, S$ 38,600, (from 4/2002 - )
This project
endeavors to develop an innovative distributed approach based on
Intelligent Agents for real-time management of traffic networks. This
real-time decision support system is being designed to interact with
central control center, receive real-time traffic and control data, and
provide recommendations for control and management of the system.
Several intelligent agents, with self-learning capabilities, support
incident management operations. A multi-decision making approach reflecting
the spatial and administrative organization of traffic management agencies
is being developed to provide a coordinated solution.
Recently
Concluded Projects
Principal
Investigator: Power quality improvement using artificial intelligence
techniques, S$73,786.
This
research aims at applying AI techniques for power quality improvement,
either by putting an “intelligent” element in the lines to improve
the power quality (for example a device to study and compensate the
harmonics in power lines) or by developing power converters that do
not 'pollute' the lines. While such work is not new, the main focus of this
work is in the use of AI in controlling such power converter applications.
Power converter systems used in power quality applications tend to be
complex and difficult to model. Thus, by applying AI techniques, the need
for accurate modeling of the system may be obviated.
Intelligent
On-line Decision Support for Control and Operation of Power Distribution
System,
S$181,650.
This collaborative project
with the Indian Institute of Science, Bangalore,
India, and
Singapore Power as the industrial partner, was successfully completed in
1999.
Development
of New Incident Detection Methodologies for Expressways, S$131,000
This multi-disciplinary
industrial collaboration project combined the expertise of NUS in traffic
engineering and artificial intelligence, with CET’s experience in
communication and vehicle
navigation. The programme was supported by NUS and NSTB.
AI
Techniques for Fault Detection and Network Re-routing, S$23,500.
Several hybrid
incident detection models using neural network, evolutionary computation
and fuzzy systems were developed in this project. A shell or layer-like
structure was developed for systematic modularization for practical
implementation in large, complex systems, and tested on incident detection
problems in power distribution systems and traffic networks.
Fuzzy Neural
Network-Based Short-term Electrical Load Forecaster, S$ 17,000
The work on this
research project led to the development of a comprehensive load forecasting
system applicable to local and regional conditions. The software
incorporates hybrid artificial intelligence-based techniques for accurate
load forecasting on all days of the year. A prototype system
developed was evaluated at the power system control center in Singapore
and its performance was found to be far superior compared to the currently
used load forecaster in the control center.
|
