Evolutive Neural Fuzzy Filtering: Real Time Constrains
J. J. Medel Juárez, J. C. García Infante, J. C. Sánchez García
Microelectronics Research Department
Professional School of Mechanical and Electrical Engineering
Av. Santa Ana No. 1000, Col. San Francisco Culhuacan
Del. Coyoacan, México D.F.
Centre of Computing Research, National Polytechnic Institute
Vallejo D. F. 07738, México
jjmedelj@yahoo.com.mx, jcnet21@yahoo.com, jcsanchezgarcia@gmail.com
Abstract: - In this paper, we describe the evolutive neural fuzzy filtering properties with real time conditions,
giving the principles of its operation based on a back propagation fuzzy neural net, which adaptively choose and
emit a decision according with the reference signal changes in order to loop the correct new conditions for a
process. This work is an approach about the evolutive neural fuzzy digital filters (ENFDF). This filter using the
neural fuzzy mechanism select the best parameter values into the knowledge base (KB), actualizing the filter
weights to give a good enough answers with respect to the reference signal in natural linguistic sense.
Additionally, the filtering architecture includes a decision making stage using an inference into its structure to
deduce the filter decisions in accordance with the previous and actual filter answer in order to updates the new
decision with respect to the new reference system conditions. This process requires that all of its states bound
into ENFDF time limit as a real time system, considering the Nyquist and Shannon criteria. In this paper, we
characterize the membership functions building the knowledge base in a probabilistic way with respect to the
rules set inference to describe the reference system and the inference to deduce the new filter decision,
performing the ENFDF. Moreover, the paper describes in schematic sense the neural net architecture with the
decision-making stages in order to integrate the filtering stages as an evolutive system. The results expressed in
formal sense using the concepts into the paper references. Finally, we present the simulation of the ENFDF
operation using the Matlab
©
software. The paper has eight sections conformed as follows: 1. Introduction, 2.
Filtering conditions, 3. Neural net description, 4. Rule base strategy, 5. Real time scheme, 6. Simulations,
7.Conclusion and References.
Key-Words: - Digital filters, neural net, evolutive systems, fuzzy logic, inference mechanism, real time.
1 Introduction
There are some problems in the development of a
system as restrictions into its capacity to model high
evolutive changing processes. The new systems
should use into its architecture evolutive tools in order
to get its own perception and give the best decision
answers, using this to solve complex problems,
actualizing and adapting its perceptions and answers
in accordance with a reference dynamical system.
The use of learning techniques based on searching the
optimal solutions represents a classical alternative
manner to obtain knowledge. The evolutive systems
into the neural fuzzy digital filtering is an option for to
obtain different decision answer levels, in accordance
with a system dynamics or reference model, adapting
its answers to the possible changes by selecting the
best values in order to get the necessary convergence
conditions, which should has the best operation at
each time.
An evolutive neural fuzzy digital filter has a solution
searching process based on the reference model
conditions having at each moment a regulation
inference mechanism that describes the desire
conditions, selecting the best answer into the
knowledge base in accordance with it, emitting the
new answer condition for the best adaptation.
Proceedings of the 8th WSEAS International Conference on SIGNAL PROCESSING, ROBOTICS and AUTOMATION
ISSN: 1790-5117 221 ISBN: 978-960-474-054-3