Hardware Reference
In-Depth Information
By considering the physical constraints in cyberphysical microfluidic system,
we have developed a layout optimization technique for cyberphysical PCR biochips.
A statistical model for DNA amplification has been introduced. The module is
used to predict whether the droplet has enough DNA strands to carry out the PCR,
and it helps us to improve the efficiency of PCR on cyberphysical biochips. We
have developed a layout algorithm that considers interference between the on-chip
elements such as reservoirs, sensors and thermal units. An optimization algorithm
has been developed to minimize the area and electrode count of the biochip and
satisfy proximity constraints. Several real-life bioassays have been used to demon-
strate that the proposed design method reduces the overall mixing/dilution/detection
time for a given bioassay.
This topic has also presented a synthesis algorithm for bioassays that is capable
of taking into account uncertainties inherent in biochemical reactions. The proposed
time-uncertainty aware synthesis algorithm can overcome the inherent variability
and randomness of biological/chemical processes and improve the quality of
product droplets. An “operation-interdependence-aware” synthesis algorithm has
been developed—the first on-chip biochemistry synthesis procedure that does not
use the module library as a design guideline. Using this algorithm, the dependence
on characterization can be reduced. The algorithm leads to a design approach that
considers completion-time uncertainties for fluidic operations, hence the accuracy of
fluidic operations is also improved. The proposed approach explicitly considers the
transportation of droplets. Hence it guarantees the feasibility of droplet-routing, and
determines the transportation path for each droplet. In order to reduce the execution
time of the bioassay without additional degradation of electrodes or hardware cost,
the design of microfluidic biochips using multiple clock frequencies has also been
considered. An on-line droplet-routing method with low computational complexity
has been developed. The response time of the resulting cyberphysical system is
negligible. Therefore, the degeneration of intermediate products in the bioassay can
be avoided.
To minimize the number of control pins and simultaneously maximize the
flexibility of biochips, we have developed design techniques for general-purpose
pin-limited biochips. A graph model for analyzing pin-assignment configurations
of a biochip has been developed. Based on the graph model, we have developed
a heuristic algorithm for the design of pin-assignment configuration, devised an
acceptance test, and developed a scheduling algorithm for fluid-handling operations.
The manipulation of droplets with various volumes has also been considered, and
new design constraints associated with the processing of large droplets have been
discussed. The metal wire routing method for pin-limited has been investigated.
Finally, by combining the design algorithms for general-purpose pin-limited biochip
and cyberphysical microfluidic biochips, a design flow for cyberphysical pin-limited
biochips has been developed.
In summary, this topic has presented a comprehensive algorithmic infrastructure
for the design and optimization of a cyberphysical biochip system. A series
of related design problems for cyberphysical microfluidic systems have been
