On Mon, 24 Sep 2018 15:18:07 -0700, OHM Ω http://aviation.derosaweb.net
wrote:

https://www.sciencedaily.com/release...0920101605.htm

With this technology could anyone win a soaring contest?

Expiring minds want to know. ;-)

This was done way back in 2007/8 by the ALOFT Project, and the autonomous
soarer that was developed was able to place well when flown against human
RC soaring pilots in an XC contest.

ALOFT used surprisingly simple sensors and off-the-shelf equipment: a GPS
receiver, standard RC autopilot (presumably fitted with MEMS solid state
attitude and rate sensors) and software running on a laptop with a
bidirectional radio link to the model. The model was a carbon 5m ARTF
thermal soarer. The code in the laptop handled navigation and thermal
utilisation strategies.

ALOFT found that the GPS was a less noisy data source than the sort of TE
varios that were commonly used in RC soaring, so all vertical velocities
were derived from it. I'm unclear why that was - possibly it was
something to do with the RC soarer being more affected by small-scale
turbulence than being used to flying a larger, heavier sailplane might
lead you to expect.

ALOFT was a successful Phd project for an aeronautical engineering
student, with the ultimate test being to fly ALOFT autonomously in one of
the Californian RC XC soaring competitions, where they fly tasks of up to
100km with the pilots in convertibles or on the back of pickup trucks.

Links:
======

http://www.xcsoaring.com/contests/mccc/2008/report.html

https://arc.aiaa.org/doi/10.2514/1.C000287

http://www.dtic.mil/dtic/tr/fulltext/u2/a614555.pdf

The first is the contest report, the second is the Phd thesis transcript
and the third is a detailed description of the control system and its
strategies for finding and using thermals. Derivatives of this control
system might be suitable for inclusion in our nav systems, though some
might say that using a nav system to predict thermal locations from
recent flight data is cheating. I couldn't possibly comment about that.


--
Martin | martin at
Gregorie | gregorie dot org