Page 122 - Control Theory in Biomedical Engineering

108 Control theory in biomedical engineering

Fig. 1 Fuzzy arrhythmia classification methodology.

signal from the artifacts and to extract the features. Then, the obtained fea-
tures are used as inputs for the FLC in order to classify the MIT-BIH record-
ings into five arrhythmias (NSR, RBBB, LBBB, PVC and P). This classifier
demands two major steps including its configuration and optimization.
The proposed FLC imitates the process of a standard controller. Indeed, a
standard controller, as it is described in Fig. 2A, uses the error between the
output and the reference input to activate the control action. This action
tries to reduce the error. If the error is reduced to zero, the output is equal
to the reference input. By doing an analogy with the standard controller, the
proposed FLC, as it is illustrated in Fig. 2B, uses the error between the pre-
dicted output and its corresponding target to activate the membership
parameters’ optimization by using a GA. Then, the FLC is updated with
the optimized parameters and new outputs are generated. This process is
repeated until the error reaches zero.

Input

Reference Error Output
– Controller System
input
+

(A)
Input

Target Error Output
– GA FLC
+

(B)
Fig. 2 (A) Standard controller (B) Genetic FLC.