2007 Fiscal Year Final Research Report Summary
self-tuning and stochastic resonance-theory and application of information processing with use of noise
| Project/Area Number |
18560060
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Engineering fundamentals
|
|---|
| Research Institution | Kyoto University |
|---|
Principal Investigator |
MUNAKATA Toyonori Kyoto University, Dep. Applied Math, And Physics, Professor (40026357)
|
|---|
| Project Period (FY) |
2006 – 2007
|
| Keywords | self-tuning / stochastic resonance / threshold system / noise intensity / double-well system / two-state system |
|---|
| Research Abstract |
(A) selftuning and stochastic resonance-mathematical aspects of threshold systems
We choose a threshold value and the noise intensity as important parameters to characterize a simple threshold system.
By applying the self-tuning mechanism to these parameters, separately or simultaneously, we estimate perforniance of the system based on mutual information and signal to noise ratio (SNR). We also studied how the response of the threshold value to slowly varying external input signals.. Due to the feedback we found that the response become slower and there occurs an ergodic-nonergodic transition in the threshold system.
(B) self-tuning and stochastic resonance-mathematical aspects of two-state system.
As an approximate description of the double-well system,дwo-state system has been studied intensively. We introduced the mechanism of self-tuning to the two-state model and studied its effects and discussed the relation between the self-tuning and stochastic resonance.
(C) self-tuning and information processing-filtering and its application to picture processing High sensitivity of our auditory system is explained based on the idea of selftuning. The environment where the auditory system is put may be very noisy or very quiet. So in order to achieve good performance irrespective of the environment it is put in, it is desirable to have some mechanism by which we can adjust the threshold value to the noise intensity of the environment. We applied this idea to the picture reproduction in conjunction with the filter theory.
(D) Stochastic resonance in the FitzHugh-Nagumo model.
The FitzHugh-Nagumo model is usually described by a differential equation. We introduce and define the mutual information (M) to this Continuous time system and discussed stochastic resonance based on the MI
|
