32 DESIGN007 MAGAZINE I JANUARY 2018
selection and guidance. The results are far
more amenable to professional PCB designers
and engineers (Figures 7 and 8).
This current generation of routing automa-
tion is a great step forward, providing speed
for designers while allowing them to maintain
quality and control of routing. But in my opin-
ion, it won't be enough to carry us into the
future of IoT.
Neural net-based topological
routing engines need to be way
more powerful. But the only
way we can improve them to
the point of useful results for
every desktop, is through deep
learning.
Again, coming from a com-
plete layman, deep learning
is a fancy term for expanding
the data set that teaches the
artificial intelligence engine
called a deep neural network,
and applying that expansive
set of stimuli and responses
to develop the software (or
machine's) behavior.
The real question here is,
what if the neural net routing
automation had a practically
unlimited set of PCB designs
to train it? What if, each time
you design a new PCB in your EDA tool, the
EDA tool could learn your moves and begin to
help you by anticipating how the board could
be routed, or even how moving some parts on
the board (within your constraints, of course)
could drastically reduce routing time and layer
count? What if you could leverage and contrib-
ute (voluntarily) to a hive-mind of engineers
and PCB designers anonymously to train the
world's greatest routing engine?
There'd be no excuse left for not designing
the next imperative device!
DESIGN007
References
1. Ericsson
2. IBM
3. Cisco
4. IEEE Tech Talk
Ben Jordan is senior manager of
product and persona marketing.
He is a computer systems and PCB
engineer with over 20 years of
experience in embedded systems,
FPGA, and PCB design.
Figure 7: User-guided routing automation (Source: Altium LLC 2017).
Figure 8: User-guided automation results
(Source: Altium LLC 2017).
