Computer Logical Design Laboratory, Annual Review 2006, The University of Aizu

Annual Review 2006 > Computer Logical Design Laboratory

Computer Logical Design Laboratory


Tsuneo Tsukahara Professor	Robert H. Fujii Associate Professor

The following researches are progressing in Computer Logical Design Laboratory:

Tsuneo Tsukahara:

Software-Defined Radio Receivers
Related to this topic, the following work was done in 2006.
A Spectrum-Analyzing Method for Multi-band Wireless Receivers:
Recently the demand for wireless systems such as sensor networks has been rapidly growing. However, radio-wave resources are limited and invaluable especially in these days. Therefore, software-defied radios (SDRs) and cognitive radios, which is a principal application of SDR, can be the key to greatly improving frequency-spectrum efficiency. SDRs demand flexibility and reconfigurability in RF (Radio Frequency) circuits. Therefore, a frequency-efficient wireless transceiver architecture is indispensable. In 2006, we proposed a multi-band wireless receiver using a spectrum-analyzing method. In a receive mode, radio frequency (RF) signals are first amplified and converted to intermediate frequency (IF) signals less than about 10 MHz using mixers or frequency converters. In a spectrum-analyzing method, we use Fourier-series expansion of IF signals in the analog-signal domain because undesired strong signals should be attenuated before analog to digital (A/D) converters to save the number of resolution bits of A/D converters. Coefficients of the Fourier series are calculated using an analog multiplier and an integrator. Because IF frequencies of undesired strong signals are known, their coefficients are forced to zero, thereby being attenuated. Because Fourier-series coefficients of desired signals coming from different radio bands are simultaneously obtained, we can select suitable radio bands for stable communication.

Robert Fujii:

Neural Networks
Artificial Neural Networks emulate some of the functions carried out in the brain. A new model of a neuron called “spiking neuron” that can process temporal information was investigated. The input data into a spiking neuron consists of spike trains; each spike train consists of a number of spikes with real-valued inter-spike times. Each data spike consists of a unit amplitude short duration digital pulse. The inter-spike times can be used to encode information. A method to map a spike train into an output firing time with close to a one-to-one relationship between input and output was obtained. A learning algorithm for spiking neurons was developed. Computer simulations have shown that the proposed mapping and learning methods were effective in applications such as function approximation, sound recognition, and sequence recognition. For example, music melodies of various lengths were used as temporal and sequential data to test the effectiveness of spiking neurons in sequence detection. Results have shown that a spiking neural network can be made to learn long inter-spike sequences (e.g. 900) without suffering from the forgetting the earlier sequences problem of recurrent neural networks.
Future work in this area will include the mathematical modeling, simulation, and hardware implementation (analog/digital VLSI design) of neurons in areas of the brain such as the auditory nerve system.
Robotics
Research in the area of autonomous,intelligent,and interactive robotics for the home or in a work environment is gaining momentum. Research and development work in the areas of bipedal robot control,image processing for self-localization and heading angle determination, and physics based robot simulation were carried out.
For the bipedal robot walking motion control, the zero moment point (ZMP) control method was used to stabilize the robot; however, although the method was effective, the processing time needed to gather sensor data and control the robot servo motors were too long and the robot was not able to walk in a stable manner.
Image processing was carried out using the OpenCV computer vision library, HSV color data, and a camera mounted on the bipedal robot’s body. The margins of error for self-localization and heading angle determination were approximately 12 percent.
A four-legged robot simulator was developed using the Open Dynamics Engine (ODE) simulation libraries. Trot walk and gallop were simulated.
Future work in the robotics area will include real-time learning of some application such as soccer ball interception and reception for the RoboCup competition two-legged robot locomotion control using neural networks, and development of a robot moved by muscle-like devices.

Refereed Journal Papers

[tsuka-01:2006]

Kado Y., Douseki T., Matsuya Y., and Tsukahara T. Ultralowpower CMOS/SOI Circuit Technology. IEEJTransactions on Electronics, Information and Systems, 126(6):725.729, June 2006.

Invited paper: We have introduced an example of a system that embodies the concept of a ubiquitous communication services and explained the importance of low power consumption in the communicator that will serve as the bridge between the real world and the network. An effective solution of high energy-efficiency is the synergy of combining low-voltage analog/RF circuit technology and FD-SOI devices, thereby providing the operation power of less than 10 mW and the stand-by power of less than 10 μ W.

Refereed Proceeding Papers

Grants

[fujii-01:2006]

R.H. Fujii and K. Oozeki. Temporal Data Encoding and Sequence Learning with Spiking Neural Networks. In Artificial Neural Networks ICANN 2006, pages 780.789, Athens, Greece, Sept. 2006. ICANN, Springer.

Sequence Learning using a Spiking Neural Network (SNN) was performed. An SNN is a type of Artificial Neural Network (ANN) that uses input signal arrival time information to process temporal data. An SNN can learn not only combinational inputs but also sequential inputs over some limited amount of time without using a recurrent network. Music melodies were encoded using unit amplitude spikes having various inter-spike interval times. These spikes were then fed into an SNN learning system. The SNN learning system was able to recognize various melodies after learning. The SNN could identify the original and noise-added melody versions properly in most cases.

[tsuka-02:2006]

Tsukahara T. Grants-in aid for Scientific Research (KAKENHI) from JSPS [Young Researcher (Start-up)], 2006-2007.

Academic Activities

[tsuka-03:2006]

T. Tsukahara, 12 2006.
Tutorial lecture at Asia Pacific Microwave Conference (APMC):“ Introduction to CMOS RF Circuits ”

Ph.D, Master and Graduation Theses

[fujii-02:2006]	Kotaro Umi. Graduation Thesis: Autonomously Moving an Object to a Specified Location, University of Aizu, 2006. Thesis Advisor: Fujii, R.H.
[fujii-03:2006]	Taiki Ichishita. Graduation Thesis: Performance Evaluation of a Temporal Sequence Learning Spiking Neural Network, University of Aizu, 2006. Thesis Advisor: Fujii, R.H.
[fujii-04:2006]	Yuuki Kanou. Graduation Thesis: Self-Localization Effectiveness Comparison, University of Aizu, 2006. thesis Advisor: Fujii, R.H.
[fujii-05:2006]	Tamon Yoshida. Graduation Thesis: Extracting and Utilizing Chordal Information for Improved Structural Analysis of Music, University of Aizu, 2006. Thesis Advisor: Fujii, R.H.
[fujii-06:2006]	Masashi Takasaki. Graduation Thesis: Simulating the Dynamic Performance of Robotic Motions, University of Aizu, 2006. Thesis Advisor: Fujii, R.H.