Geology Reference
In-Depth Information
Figure 3. PID-BELBIC control algorithm
is integrated with the neuromorphic-PID control
algorithm. The converting algorithm can be ei-
ther an inverse MR damper model or a clipped
algorithm (Kim et al. 2009). In this article, to
accomplish this, a clipped algorithm is utilized.
A clipped algorithm for a MR damper application
is (Yoshida and Dyke 2004)
value relating the MR damper force to the voltage;
and
V
max
is the maximum voltage to be applied
to the MR damper. A schematic diagram of the
proposed hybrid neuromorphic-PID-Inv control
algorithm is shown in Figure 4. The structural
responses (e.g., interstory drift) and the BELBIC
control force signals are used to generate pre-
defined sensory input (
SI
) and emotional signal
(
ES
) functions. In addition, the
SI
function is
constructed as a linear combination of the struc-
tural responses and BELBIC control signals using
the weighting factors in Equation (15). The
ES
function signal is developed through the integra-
tion of the weighted integral BELBIC control
forces with the PID control signals as shown in
Equation (16). Based on the
SI
and
ES
functions,
the BELBIC controller is operated as previously
stated: in brief, unlike classical learning or adap-
tive controllers, the BELBIC controller is di-
rectly inspired by biology, wherein the amygdala
constantly learns the associations between the
SI
and the
ES
and functions based on these learned
associations. On the other hand, the orbitofrontal
(
)
{
}
v V
=
H
f
−
f
f
,
(10)
a
Neuromorphic-PID
m
m
where
µ
⋅
f
for
for
f
≤
f
c
Neuromorphic-PID
Neuromorphic-PID
max,
V
=
V
f
f
max,
a
euromorphic-PID
>
max
N
(11)
where
v
is the voltage level, H is a Heaviside step
function,
f
m
is a measured MR damper force which
is calculated from Equations (14) to (20), and
f
Neuromorphic-PID
is a control force signal generated
by the neuromorphic-PID controller;
µ
c
is a
Search WWH ::
Custom Search