UST008

17 / 92

17
microwave, radar and UAVs, and if you
combine all these elements together
you can see dramatic things happening,
where you can use the embedded
systems technology to put radar and
sensing technology into UAVs,” he
says. “This combination could generate
something magical.
“I firmly believe that the MPU-FPGA
combination will be the future of all UAV
processors. The market is being misled
by the commercial UAV companies.
Flight controllers for consumer drones
run everything on one chip to reduce the
cost of the technology so that it can be
used by everyone, and this is cool, but
I still think the future of the UAV relies
on some serious processing and some
serious sensors with reliable and secure
processing, and almost zero chance of
failure.
“From that perspective you need real-
time processing to simplify the whole
design,” he says.
Radar-autopilot combo
The first step towards this future is a low-
cost, light radar system combined with a
UAV’s autopilot for sense-and-avoid and
collision-detection applications. “Now we
have the knowhow to miniaturise a radar
system that previously needed a lot of
power and electronics,” says Dr Wang.
The key to this is the FPGA – large
arrays of programmable logic, called
a fabric, that are particularly good for
handling signal processing. That means
elements of the autopilot and the radar
processing can be implemented in the
fabric. Instead of using thousands of
processor cycles to handle the signals,
it can be performed in a few hundred
cycles by the logic on the fabric.
That saves power and provides more
performance so that signals from a radar
sensor can be analysed quickly.
Unlike a microcontroller programmed
in a language such as C, the logic for
an FPGA is designed in a high-level
language such as VHDL or Verilog.
This logic is then synthesised into a
bit stream that is stored on an external
memory chip and then downloaded to
the FPGA to configure it.
“We focus on two technologies. One is
sensing, and we are focused expressly
on microwave sensing, as we believe
this could be something useful for UAVs,”
says Dr Wang. “We can compensate for
the disadvantages of other sensors and
provide something unique – it can’t be
blocked as an optical sensor and is not
as easy to interfere with as ultrasound.
“The second area is the processing
technology, and we are especially
interested in the combination of FPGA
and MPU. That’s a big challenge, as
we need to do the signal processing to
retrieve the signals returning from the
obstacles and the targets.
“It’s not about the sensor but the
combination of the sensor and the
processing algorithms. That makes
things a bit challenging to engineers as
you need a clear understanding of the
whole signal chain.”
One way to get the FPGA technology
small enough for use in UAVs is to
optimise the signal chain and reduce
the range of the sensor. “For most
commercial UAVs you don’t need
kilometres of range so you don’t need
as much power – a couple of hundred
metres of range is pretty good, and that’s
why we were able to shrink the size and
make the radar system smaller with less
power consumption in a relatively small
FPGA,” he says.
“We design the sensing and
processing systems ourselves, so
everything can be tuned for UAVs. It’s
not a case of, ‘Let’s make a small radar
that can be used for UAVs’ as you need
to tune the antenna pattern, the power
budget, the link budget, the processing
and interaction with flight controllers.”
Frequency spectrum
Aerotenna is working with a range of
chip makers on the radar transmitter
and receiver chain for various frequency
ranges. “We use a very wide spectrum
– we are not fixed though; we have a
bunch of products with different ranges,”
he says. Each of these uses a separate
logic design that is downloaded to the
most suitable type of FPGA.
However, it is the ‘system on chip’
(SoC) combination of both FPGA and
ARM processor cores on a Xilinx Zynq
SoC that opens up the opportunities for
UAV control systems. “This combination
is spot-on for our solution, using the
ARM cores for the flight control core
Dr Zongbo Wang | In conversation
Unmanned Systems Technology | June/July 2016
Dr Zongbo Wang has more than
15 years of experience in sensor
and electronics development. He
honed his expertise by leading
numerous engineering and
technology research projects at
universities and research institutes
in China, Spain, the Netherlands,
Singapore and the US. He received
his BSc and PhD in Electronic
Engineering from the Beijing
Institute of Technology in 2004 and
2009 respectively, and founded
Aerotenna in 2015.
Dr Zongbo Wang