"Do It Yourself" airborne proximity warning device

Maybe someone should develop a device like the MRX PCAS which detects
transponders and includes azimuth in addition to range and elevation. Most
of the algorithms have already been developed. There are well established
methods for very accurately locating a transponder. Look up ASDE-X, for
example (LAT/LON/ALT derived from transponder replies). Alas, I suspect
development cost would far outweigh expected return on investment.

Steve makes excellent points, especially the environmental issues with using
consumer electronics in an aviation envrionment. Having said that, I must
ask if Flarm and PowerFlarm have FAA certification, as does a certified
transponder. Steve's post implies that it is but, according to the FAC
posted at gliderpilot.org, it is not, and does not require certification.
As to Steve's question of the veracity of the system under discussion in an
environment including 50 gliders in close formation, consider that this
display has only 64 pixels and 3 of them display the own ship! That would
be a pretty busy display...

According to the NTSB accident database, since 1994 in the USA there have
been exactly 6 midair collisions involving a glider as listed below (9 if
you consider two gliders running into each other):

1999 - Gllider hit tug which was towing another glider
2003 - Piper Cub flew into aerobatic box and collided with glider using the
box
2006 - Glider and corporate jet collided at 16,000' MSL near Reno, NV
2008 - 2 gliders collide while thermalling
2012 - 2 gliders collide head on
2012 - 2 gliders collide while thermalling at the Worlds Championships

I'm going to guess that neither Flarm nor PowerFlarm were available in 2008
or earlier and, if that is the case, then this technology might have
prevented exactly two accidents in the US. I'll bet that both gliders in
the World's were so equipped and the technology failed, so I'm still waiting
for difinitive proof that it's worthwhile. In most cases, a good outside
scan would also have prevented an accident. In response to the anticipated
statement that we'll never know how many accidents were actually averted by
Flarm, I can only say that a good traffic scan is usually all that's
required, or rejecting or leaving a crowded thermal (competition excepted).

I used the keywords "midair" and "glider" in my search but there may well be
others which I missed. My point is that, considering the number of glider
flights conducted in the US, the risk of a midair is extremely low and, in
my opinion, does not warrant the expense, complexity, or distraction of a
collision warning device for most of the glider flying done in the US.
Competition flying is different, of course, as it concentrates so many
gliders in the same airspace. Europe is much more congested and has far
more glider flights than we do and I can see more of a benefit for them.

And, finally, for a good many of us glider pilots, we cannot simply lay down
for an ASG-29, full panel, and Cobra trailer. For us, the sport is somewhat
cost driven.

"Steve Koerner" wrote in message
...
This is interesting and clever.

But it does not work with Flarm! Flarm/PowerFlarm is seeing rapid adoption.
A system that competes with Flarm has only the possibility of reducing
safety during the time frame that I expect to remain an active glider pilot.
Though competition is usually good, it is not a good thing to have
competition in this case. What makes it worse is the possibility that any
pilot might consider waiting for this instead of installing Flarm/PowerFlarm
right now.

Just like Flarm, this system requires that both gliders be like equipped.
Having a contingent of Flarm users and a contingent of WiFi users at a
contest means that we cannot get to the significant level of safety
improvement that would be otherwise achievable with fully adopted Flarm or
PowerFlarm.

As I read through the material I couldn't find a single element of technical
superiority over PowerFlarm. It seems to me that for one technical standard
to replace another established standard it needs to be distinctly better
than the first. Being equivalent (if it were) is not near good enough, even
if the cost is lower. Soaring is not so strongly cost driven as consumer
products for example and in general having an avionic component supported by
a manufacturer is a very important benefit.

Starting from where the developers are now, they are very far behind
PowerFlarm. Part of the goodness of PowerFlarm is the years of evolution in
the algorithms for the collision risk analysis. In a fast changing
environment of side by side cruising and close thermalling, PowerFlarm makes
good determinations. Even the most brilliant programmer on earth cannot
just sit down and write that code. It takes years of observation and
feedback to make it work really well in the real world. The electrical
components are not the major part of the problem; the magic is really in the
software.

And for close proximate flight, I'm led to wonder how the designers might
have come to the conclusion that 2-3 second latency would be acceptable for
good warnings? Having flown with PowerFlarm, I have to believe that the
latency is lower than that.

On the hardware side, I think there are things the developers are not
considering well. The use of a high gain (5 dBi) antenna is not advisable.
It's important to use a low gain dipole pattern antenna in order to couple
well with turning gliders. With a low gain antenna at both TX and RX, the
link analysis will be significantly impacted and you will not have the range
that has been speculated. PowerFlarm uses a simple dipole for this reason
and yet has greater range than is contemplated with the high gain antennas
suggested here.

Also related to the coupling matter is the choice of frequency. 2.4 GHz
will be significantly more impacted by the nearby human body and other items
of near wavelength dimension in the environs of the antenna. This can be
overcome to a certain extent with power margin but there isn't power margin.
PowerFlarm provides an auxillary receive channel to partially address this
issue. An auxillary channel is needed in spades at 2.4 GHz.

There is no mention in the article as to the level of degradation that might
be expected in a contest environment with say 50 gliders all within radio
range. What is the duty cycle of the waveform? How much would 50 gliders
be expected to further reduce functional range?

In this self assembly scenario, who does the testing? One of the things
about electronics in general and avionics in particular is the need for
sophisticated testing. Having a manufacturer behind an avionics product
means that the items have been tested. There is the production testing of
each article as it leaves the assembly area. Even more important is that
all of the components that go into the design have been technically
qualified as suitable. That means that they are tested for operation over a
wide temperature range as well as shock and vibration and humidity and
pressure. They are tested for having a suitably small degree of parameter
variance over the environmental range. All of the USB consumer items that
are identify for this project are items that are generally made in China and
are intended only for use at room temperature in benign environments. It
would be almost remarkably if they all happen to also work over aviation
temperature range. I'd be particularly suspicious about the radio module
power output and the radio sensitivity over temperature; especially for a
device that was never actually intended for operation over temperature.

In fairness the original poster, he did not describe the system as intended
to be a replacement for Flarm/PowerFlarm. Yet as described, that would be
the obvious thing that many readers might be considering here. For that
reason it is worthwhile to point up these considerations and limitations.

Even as I hope that it eventually works well for OP's club, I'm also hoping
that no US pilots in particular might be looking at this as a suitable
substitute for PowerFlarm.

On Tuesday, December 10, 2013 10:19:07 AM UTC-8, Dan Marotta wrote:

I used the keywords "midair" and "glider" in my search but there may well be
others which I missed. My point is that, considering the number of glider
flights conducted in the US, the risk of a midair is extremely low and, in
my opinion, does not warrant the expense, complexity, or distraction of a
collision warning device for most of the glider flying done in the US.
Competition flying is different, of course, as it concentrates so many
gliders in the same airspace. Europe is much more congested and has far
more glider flights than we do and I can see more of a benefit for them.

And, finally, for a good many of us glider pilots, we cannot simply lay down
for an ASG-29, full panel, and Cobra trailer. For us, the sport is somewhat
cost driven.


For some reason it is hard to get the NTSB database to cough up all the incidents. You missed several midairs I know of in the past 5 years or so including one requiring a bailout and one where one pilot tried to bailout but was unable to and thankfully was able to land without injury. That doesn't include a number of scary near misses.

I believe the data shows that midair is the second leading cause of fatality next to stall-spin/collision with terrain. Glider-glider collision is at least ten times likely as glider-GA collision and (by the data) infinitely more likely than glider-air transport collision. If we ever got one of those it would be ugly and bring the stats up to making glider-glider 100 times more likely than glider-air transport.

Yes, contests gather gliders and concentrate traffic but if you look at the some of the work that has been done to accumulate OLC traces into glider flight path "heat maps" you discover that the combination of topography, airports, airspace and (especially) lift sources puts gliders in much closer proximity to each other than you might otherwise think. Gliders tend to occupy a small, common proportion of the available airspace, even though we think we are flying just anywhere. This explains why we see more glider-to-glider collisions than any other kind of glider involved collision. It raises the question as to whether if forced to trade off transponder vs Flarm for cost reasons the most bang for the buck really might be Flarm, even for non-contest flying near terminal areas. The midair collision data suggests this might well be true since the penetration of Flarm and transponders in gliders are both low. The equation would only flip for very small numbers of gliders (5) flying right up against a busy international airport - though there aren't many of these. I carry both Flarm and a Mode S, but I realize others feel they can't afford both, just the way some feel a parachute isn't worth the cost (I believe there also are fewer successful bailouts than glider-glider midairs - so selling one's parachute to buy a Flarm may also be a statistically superior solution - though emotionally I can't imagine anyone making the switch).

In any case, having two, incompatible, Flarm-like technologies is a terrible idea for the reasons already articulated.

9B

Excellent reply!

I tried several keywords and the midairs I found in the NTSB database were
those that I listed. I will accept your assertion that there are more - I
just couldn't find them.

While reading your response regarding Flarm being better than a transponder,
it occurred to me that, where I fly that is just not the case. Due to the
altitudes that we fly, pretty much all powered aircraft have to have
transponders (above 10,000' MSL), and there are probably less than a dozen
of us that venture far from the airport. Our airport is also very lightly
used by power traffic and the cross country pilots usually return late in
the day after all hangars are closed. Our only major concerns are the IFR
arrival and departure routes which are near the airport. So, speaking
purely from my flying situation, a transponder is a far better solution than
a Flarm.

Your situation is, of course, different.

wrote in message
...
On Tuesday, December 10, 2013 10:19:07 AM UTC-8, Dan Marotta wrote:

I used the keywords "midair" and "glider" in my search but there may well
be
others which I missed. My point is that, considering the number of glider
flights conducted in the US, the risk of a midair is extremely low and, in
my opinion, does not warrant the expense, complexity, or distraction of a
collision warning device for most of the glider flying done in the US.
Competition flying is different, of course, as it concentrates so many
gliders in the same airspace. Europe is much more congested and has far
more glider flights than we do and I can see more of a benefit for them.

And, finally, for a good many of us glider pilots, we cannot simply lay
down
for an ASG-29, full panel, and Cobra trailer. For us, the sport is
somewhat
cost driven.


For some reason it is hard to get the NTSB database to cough up all the
incidents. You missed several midairs I know of in the past 5 years or so
including one requiring a bailout and one where one pilot tried to bailout
but was unable to and thankfully was able to land without injury. That
doesn't include a number of scary near misses.

I believe the data shows that midair is the second leading cause of fatality
next to stall-spin/collision with terrain. Glider-glider collision is at
least ten times likely as glider-GA collision and (by the data) infinitely
more likely than glider-air transport collision. If we ever got one of those
it would be ugly and bring the stats up to making glider-glider 100 times
more likely than glider-air transport.

Yes, contests gather gliders and concentrate traffic but if you look at the
some of the work that has been done to accumulate OLC traces into glider
flight path "heat maps" you discover that the combination of topography,
airports, airspace and (especially) lift sources puts gliders in much closer
proximity to each other than you might otherwise think. Gliders tend to
occupy a small, common proportion of the available airspace, even though we
think we are flying just anywhere. This explains why we see more
glider-to-glider collisions than any other kind of glider involved
collision. It raises the question as to whether if forced to trade off
transponder vs Flarm for cost reasons the most bang for the buck really
might be Flarm, even for non-contest flying near terminal areas. The midair
collision data suggests this might well be true since the penetration of
Flarm and transponders in gliders are both low. The equation would only flip
for very small numbers of gliders (5) flying right up against a busy
international airport - though there aren't many of these. I carry both
Flarm and a Mode S, but I realize others feel they can't afford both, just
the way some feel a parachute isn't worth the cost (I believe there also are
fewer successful bailouts than glider-glider midairs - so selling one's
parachute to buy a Flarm may also be a statistically superior solution -
though emotionally I can't imagine anyone making the switch).

In any case, having two, incompatible, Flarm-like technologies is a terrible
idea for the reasons already articulated.

9B

Hey Dan,

There is actually nothing in my posts that implies that PowerFlarm is FAA certified. I know that it's not and I have no concern that it's not.

Moreover, though your collision statistics are incomplete, your data would in itself drive me to an entirely different conclusion than the one that you prefer.

GW

Sorry, Steve.

I misinterpreted your use of the term "avionics" to mean certified. Please
see my recent post concerning those statistics and flight environment and
how they led me to my conclusion.

Dan


"Steve Koerner" wrote in message
...
Hey Dan,

There is actually nothing in my posts that implies that PowerFlarm is FAA
certified. I know that it's not and I have no concern that it's not.

Moreover, though your collision statistics are incomplete, your data would
in itself drive me to an entirely different conclusion than the one that
you prefer.

GW

On Tuesday, December 10, 2013 10:19:07 AM UTC-8, Dan Marotta wrote:

Maybe someone should develop a device like the MRX PCAS which detects
transponders and includes azimuth in addition to range and elevation. Most
of the algorithms have already been developed. There are well established
methods for very accurately locating a transponder. Look up ASDE-X, for
example (LAT/LON/ALT derived from transponder replies). Alas, I suspect
development cost would far outweigh expected return on investment.

I looked it up - ADSE-X is an active radar system that uses either a rotating or phased array antenna (apparently normally mounted on top of the control tower). It's not the sort of thing you'd find you could fit in a glider - even if it were legal. Here's a long to a schematic of the Raytheon version:

http://avstop.com/stories/asde.html

I am not aware of any system you could even adapt to put in a glider that would allow you to get accurate azimuth information off of transponder returns - even in theory.

9B

Well, define "accurate", and "could put in a glider".

I've never seen one, but the rather boxy Zaon "XRX" was supposed to give azimuth information. I believe it was crude ( quadrant or octant ), and I have no information about how well it worked other than reviews. I have a Zaon "MRX", which is a small altitude-only reporting receiver, and find it useful. Too bad Zaon closed operations recently.

Review: http://www.flyingmag.com/avionics-ge...oidance-system

On Tuesday, December 10, 2013 2:34:45 PM UTC-6, wrote:
...


I am not aware of any system you could even adapt to put in a glider that would allow you to get accurate azimuth information off of transponder returns - even in theory.



9B

On Tuesday, December 10, 2013 1:41:26 PM UTC-8, Sarah wrote:

Well, define "accurate", and "could put in a glider".

Fair enough.

If you were able to put TCAS in a glider that would do it, but trying to get azimuth information off of passive monitoring of radar returns (like PCAS does) has to be a hit-or-miss proposition (pun intended) since you don't have the ability to actively interrogate other transponder-equipped aircraft to string together enough bits of information to get good direction. You are dependent on ground radar or TCAS-equipped aircraft to do the interrogating for you which is no always reliable. Some sort of directional antenna added to a PCAS might help in the way you describe (showing quadrants), but I have to believe it's not the sort of thing you could really count on and would totally suck for glider-glider scenarios.

I'd also add that the research shows that no matter how diligent the scan, see-and-avoid detects not more than half the targets that are collision threats. Non-threats are much easier to pick up because of the angular movement of non-collision targets. So, the fact that you see other aircraft when you are flying to some extent generates a false sense of security - your are much less likely to see the one that's going to actually hit you. There are scenarios in the research where successful detection in time to act is on the order of 10-20%. That gave me some pause.

9B

You don't need directional antennae. Without getting too long winded (I
know, I know...) you monitor the arrival times and frequencies of the
interrogation signals and the replies and combining that with your known
position, you can mathematically determine the positions of all the
emitters. Multiple samples enable the system to determine velocity (a
vector of direction and speed).


wrote in message
...
On Tuesday, December 10, 2013 1:41:26 PM UTC-8, Sarah wrote:

Well, define "accurate", and "could put in a glider".

Fair enough.

If you were able to put TCAS in a glider that would do it, but trying to get
azimuth information off of passive monitoring of radar returns (like PCAS
does) has to be a hit-or-miss proposition (pun intended) since you don't
have the ability to actively interrogate other transponder-equipped aircraft
to string together enough bits of information to get good direction. You are
dependent on ground radar or TCAS-equipped aircraft to do the interrogating
for you which is no always reliable. Some sort of directional antenna added
to a PCAS might help in the way you describe (showing quadrants), but I have
to believe it's not the sort of thing you could really count on and would
totally suck for glider-glider scenarios.

I'd also add that the research shows that no matter how diligent the scan,
see-and-avoid detects not more than half the targets that are collision
threats. Non-threats are much easier to pick up because of the angular
movement of non-collision targets. So, the fact that you see other aircraft
when you are flying to some extent generates a false sense of security -
your are much less likely to see the one that's going to actually hit you.
There are scenarios in the research where successful detection in time to
act is on the order of 10-20%. That gave me some pause.

9B

On Wednesday, December 11, 2013 9:12:45 AM UTC-8, Dan Marotta wrote:
You don't need directional antennae. Without getting too long winded (I

know, I know...) you monitor the arrival times and frequencies of the

interrogation signals and the replies and combining that with your known

position, you can mathematically determine the positions of all the

emitters. Multiple samples enable the system to determine velocity (a

vector of direction and speed).

But so what, these ground based multilateration systems have been around for years and are fairly widely used worldwide to supplement SSR radar. The way you get this traffic data to an aircraft today is via TIS-B. The position accuracy is not great (not compared to what GPS/ADS-B data-out can provide). And to receive that TIS-B traffic data your aircraft/glider needs to have ADS-B data-out, which requires an expensive IFR rated GPS, and an ADS-B data-out capable transponder or UAT (but please use a Transponder in a glider for PowerFLARM 1090ES compatibility) and a 337 field approval (for certified aircraft) that is supposed to be based on a previous STC in a similar aircraft (none of which were actually developed for gliders). So while it may be possible, good luck having that conversation with you local FSDO. And to receive that TIS-B signal requires you to be in range of the ground based ADS-B service, YMMV in some popular glider areas. And worrying about this stuff now just seems pointless given that ADS-B data-out carriage will be required in the USA in many aircraft by 2020. So you might as well just read that much higher resolution ADS-B data-out position data straight over the air now and usage will just continue to increase in future.