Keywords: parallel processing, HPC, robotics, real time, fuzzy logic, genetic algorithms, neural networks
Start Date: 1 February 1995 / Duration: 30 months
[ participants / contact]
The research project is to develop, analyse and implement compute-intensive algorithms for robotic applications, and, in particular, advanced algorithms for real time control of robots. Several software modules that represent core routines in neural networks, genetic algorithms and fuzzy logic systems will be developed. Parallelism inherent in these algorithms will be exploited.
Although theoretical work on these algorithms has long been started, there is no commercially available robotic system that employs state of the art algorithms for the aforementioned applications. The main reasons are the computational complexity of the algorithms and some theoretical limitations of the methods used. With the increase in performance of computing hardware and the emerging paradigm shift to superscalar multiprocessor implementations, new generation machines can be used to exploit the parallel nature of these algorithms. The feasibility of advanced robotic systems with intelligent sitting on top of parallel computing hardware will be assessed.
The National University of Singapore
Dep. of Mechanical and
Production Engineering
10 Kent Ridge Crescent
Singapore 119260, SG
EU Partners
Parsytec GmbH, D
Pallas GmbH, D
Non-EU Partners
The National University of Singapore, SG
Dr. Hong Geok Soon
Tel: +65 779 2210
Fax: +65 779
1459
E-mail: mpehga@nus.sg
SMIARA - ITDC-131, May 1997
please address enquiries to the ESPRIT Information Desk
html version of synopsis by Nick Cook