Biomedical Engineering Reference
In-Depth Information
2
Defect-Tolerant and
Routing-Aw
are Synthesis
In this chapter, we present a unified synthesis method that combines
defect-tolerant architectural synthesis with droplet-routing-aware physi-
cal design [50,51].
Droplet routability
, defined as the ease with which droplet
pathways can be determined, is estimated and integrated in the synthesis
flow. The proposed approach allows architectural-level design choices and
droplet-routing-aware physical design decisions to be made simultaneously.
Presynthesis and postsynthesis defect tolerance are also incorporated in the
synthesis tool. We use the dilution steps of a protein assay as a case study to
evaluate the proposed synthesis method.
2 . 1 B a c k g r o u n d
Next-generation biochips are likely to be multifunctional and adaptive
“biochemical processing” devices. For example, inexpensive biochips for
clinical diagnostics offer high throughput with low sample volumes, and they
integrate hematology, pathology, molecular diagnostics, cytology, microbiol-
ogy, and serology onto the same platform. The emergence of such integrated
and multifunctional platforms provides the electronic design automation
community with a new application driver and market for research into new
algorithms and design tools.
Over the past few years, several automated synthesis methods have
recently been proposed for digital microfluidic biochips. These design auto-
mation methods address operation scheduling and module placement for
digital microfluidics [14-18]. In Chapter 1, we reviewed these methods and
described a unified synthesis algorithm for microfluidic biochips based on
parallel recombinative simulated annealing (PRSA) [15]. The top-down syn-
thesis flow described in Chapter 1 unifies architecture-level design with
physical-level module placement. This method allows users to describe bio-
assays at a high level of abstraction, and it automatically maps behavioral
descriptions to the underlying microfluidic array.
However, the synthesis flow described in Chapter 1 suffers from two
drawbacks. For operation scheduling, it is assumed that the time cost for
19
