Robert H. Fujii

Associate Professor, Department of Computer Hardware

Ph.D. 1987 (University of Illinois at Champaign-Urbana)

Background Data

He has worked in the semiconductor industry in California's Silicon Valley as an integrated circuit design and process engineer after receiving a bachelor's degree in Electrical Engineering. He then earned the master's and Ph. D. degrees in electrical engineering from the University of Illinois at Urbana-Champaign. Subsequently, he joined the electrical engineering department at Purdue University in W. Lafayette, Indiana as an assistant professor. His research interests are in the areas of neural networks, robotics, analog/digital circuits, and computer-aided music.

Research/Educational Interests

Research Description

  • In the area of neural networks, methods to use neural networks for image processing, sequence detection, sound recognition, and are being investigated. In particular, research of neural networks for sequence detection/generation and incremental learning capable neural networks have produced good results.

    Spiking neuron networks were used to encode spike train data into spacio-temporal data. A new spiking neuron learning algorithm was developed. The spiking neuron network could do classification of data as well as detection of a sequence of data. The new area of research will be in the neuroscience area in which we will model parts of the brain and develop new neural networks that emulate the behavior of their biological counterparts. This is an exciting research area which will explore how a neuro-computer as capable as our brains can be built.

  • In the area of robotics, various methods which allow multi-agent autonomous robots to cooperate with each other to perform tasks are being investigated. We took part in various AIBO robot competitions including the AIBO soccer and AIBO challenge competitions in 2004 and 2005. Our team placed 4th in the May 2005 AIBO soccer competition held in Fukuoka and 3rd in the AIBO challenge competion held in 2003 in Niigata. We are now starting to work with bi-pedal robots.

  • In the area of analog/digital circuit design, sub-threshold MOS analog/digital neural network circuit research was carried out. Spiking neural networks for encoding spike trains and for learning are being designed.

  • In the area of computer-aided music work has been done in: a) implementing Generative Theory of Tonal Music (GTTM) rules by computer using the MAX software tool; and b) developing a basic tool for melody recognition from a partial MIDI music score. Research on computer-aided music analysis and music arrangement is being carried out.

    Selected Publications

    e-mail: fujii@u-aizu.ac.jp