Hardware Reference
In-Depth Information
1:
Phase1
:
2:
for
each permutation of
n
devices (written as
{d
k
1
,
d
k
2
, ...,
d
kn
}
) to be placed on the layout
do
3: Place
d
k
1
at the origin of the layout (usually,
d
k
1
represents the heater));
4:
for
each
d
ki
ā{d
k
2
,
d
k
3
, ...,
d
kn
}
do
5: Draw the forbidden circular regions constrained by each of the devices
{d
k
1
,
d
k
2
,
...,
d
k
iā
1
}
with respect to
d
ki
. Construct the union of digital objects enclosing the
forbidden regions;
6: Find the grid point
g
i
on the boundary of the union of digital objects, such that the
distance between the origin grid and device
d
ki
is minimum.
7:
end for
8: Output the placement geometry and calculate the cost function;
9:
end for
10: Choose the device placement that can minimize the value of the cost function; // At this
step, the locations of all the devices are determined. All the reservoirs are considered iden-
tical.
11:
Phase 2
:
12: Find out the basic mixing clusters in the sequencing graph;
13: Determine the best reservoir allocation of reagents based on the clustering result of the
sequencing graph;
14: Establish routing paths to connect all the devices following the sequencing tree;
15: Schedule operations included in basic mixing clusters;
16: Schedule operations not included in basic mixing clusters;
Fig. 5.9
Pseudocode for layout design for PCR biochips
concurrently. The scheduling of operations in the same cluster can be determined
based on the input-output interdependency [
29
].
Operations included in the basic mixing clusters are initially performed
when executing the bioassay. Other operations are executed after the operations
in mixing clusters are completed. For example, in Fig.
5.4
, operations M
1
and M
3
are in the basic mixing cluster
cluster
.R
1
;R
4
/. After operations in
cluster
.R
1
;R
4
/ are completed, one of the output droplets of
cluster
.R
1
;R
4
/
will be further mixed with droplet R
2
in operation M
8
; the output droplets
of
cluster
.R
1
;R
4
/ and
cluster
.R
2
;R
4
/ will be mixed in operation M
9
.Inthis
approach, the scheduling of all the fluid-handling operation in the sequencing
graph can be derived.
The pseudocode for the entire design cycle is shown in Fig.
5.9
(here the selection
of optimal solution is based on Approach 1).
5.5
Visibility-Aware Droplet Detection
In a cyberphysical digital microfluidic system, the monitoring of droplets is vitally
important. Researchers often use cameras and fluorescent microscopes, which are
fixed on the hardware platform to simultaneously monitor multiple operations on
