Possible use of R/C Multicopters (Drones) to search for and retrieve models

[georgegassaway]

This is an offshoot from the inflatable plane thread: https://www.rocketryforum.com/showthread.php?138676-goodyear-inflatable-plane

In message #6, I mentioned how a human-carrying drone could be used for rescuing a downed pilot rather than the inflatable plane idea of so long ago. Well, that began another change to the thread with these:

I know this kind of strays away, but I was thinking for a while that drones could be used for model rocket search and recovery operations. some kind of hook to snag the shock cord, shroud lines, parachute. Camera and Gps. Someone is surly using them to search/retrieve their rockets. This might be a subject for a new thread.

Right now at least, a typical multicopter isn't going to be able to lift more than a few ounces. So, it doesn't make sense to risk a multi-copter costing hundreds of dollars to rescue a rocket costing a few dollars.

I think the most practical use is for the search, not so much the physical retrieval.

A drone with a good live video feed costs at least $500, the better ones closer to $1000. Such as DJI Phantoms. You need a good camera with a good video feed so you can visually search for a model. In the last couple of years, I have tried to use aerial video onboard R/C models to look for lost rockets. Originally, putting a keychain camera on an electric sailplane and flying back and forth. Then landing, checking the video on computer, and finding nothing.

Last year, test flew some G Streamer Duration rockets that worked, but landed far off. Apparently in a cornfield, BUT the corn was not very high yet, a model could have been seen in between the rows (and the mylar streamer was 1 foot by 10 feet). I used a borrowed Blade 350 Quad, carrying a borrowed GoPro, to fly it up to look for the lost streamer model (BTW - I did put a radio beacon tracker onto the Quad, in case it went down, i'd be able to search for the quad more easily. Reason for not putting the tracker into the rockets being tested was in case the models went down in woods and could not be retrieved out of a tall tree,wasn't worth losing the tracker. And one of those models may indeed have gone down in woods).

Big part of the problem was not having any live video. I have wanted to do FPV (First Person View) for years, using an R/C plane, and even more now with Quadcopters. Flying that Blade 350 around to look for the streamer models, was just blindly "point and hope", not even sure where it was pointed or what it was seeing at the time. And a lot of the time it was just too high, the wide-angle view of the camera made the cornfield look way farther away than it really was. I'd have flown lower if i'd seen it real-time.

Another issue was th limited visual range I had. I had the WiFi on the GoPro running, not realizing the WiFi interfered with the Blade 350's GPS receiver (found out later on RC Groups that was a known problem. I'f I'd known at the time, I'd have turned off the WiFi on the GoPro). So, on those search flights the GPS was not working. The main impact of that was that the "Return to Home" feature would not work, as it needs a good GPS lock (When first turned on, it sets its current GPS coordinates on the ground as "home"). So, I did not dare fly the Blade 350 as far as I really wanted to go look, in case I lost sight of it and therefore would not be able to command it to Return To Home (BTW - if it loses transmitter signal, it automatically would Return To Home if it has a god GPS lock).

So, despite at least three flying sessions over two days, I never spotted a streamer. Here is one of those videos:
[video=youtube;wxC54UbmW14]https://www.youtube.com/watch?v=wxC54UbmW14[/video]

At about 25 seconds into the video, off in the distance, you can see some standing water (lot of recent rain), and then a curved line of trees. The line of sight was a bit to the right of the water, and two models were likely near those trees. One probably short of them by a few hundred feet, the one I was trying to find, and one possibly past them. But visually I could not see the Blade 350 well enough to fly it out that far. And even if I had, it would have been blind flying. I wished I could have walked over to that area and flown the Blade 350 to look around closer. But I had injured my foot days before so my walking distance was limited (especially for a farm field).

I realized later that if only I had live video, FPV, then I could have flown the Blade 350 where I really wanted to do the search, and fly a pattern to look better than the haphazard blind pointing.

And FPV is getting to be less and less expensive, at least the low-end. I can get more into that later.

So, we can discuss more about doing a search to find a model. IIRC, there has been at least one known case of a FPV Drone finding a lost HPR rocket, which IIRC was in an area of uneven terrain with small trees here and there and other random stuff, was not an simple open field.

As for using a drone to actively pick up a model..... there are several youtube videos of drones rescuing other drones. I have also seen one where a small R/C plane accidentally landed on a tiny island (like 50 feet across) in a big lake, and a drone flew over and got it.

Here is the first rescue I ever saw, several years ago, that got reposted. A drone ended up on the top of an old castle ruin, Hadley Castle. A lot of neat footage, but it is long. Third and final try begins at about 15:20, FPV view. For some reason there is no onboard FPV footage of the actual moment of rescue, there is only ground-based camera footage.

[video=youtube;U5BflGSX34Q]https://www.youtube.com/watch?v=U5BflGSX34Q[/video]

Here is a short one, rescue from a house roof:
[video=youtube;32IPBmcwplQ]https://www.youtube.com/watch?v=32IPBmcwplQ[/video]

I also found a few that involved rescues from trees. But that runs a big risk of getting the rescue drone stuck in the tree too. So, active rocket retrieval would be more practical for situations where the rocket is on a roof, or laying on the ground in some place that is difficult to simply walk over and get it

Of course it matters a lot, the relative mass and thrust of the drones, as to how much weight they could lift. My Lunar Module Quadcopter, weighing three pounds, has a max thrust to weight ratio of over 2X, IIRC. It could easily lift a pound, probably two pounds. Now, I would never risk that scale Lunar Module Quad on a rescue mission. I just mention the lift capability.

 

[Ravenex]

I successfully used a drone to locate the crash site of a rocket at bong recreation park.

https://www.rocketryforum.com/showthread.php?t=126973

 

[TangoJuliet]

Recently a few of us (myself and Andy Woerner) have been launching rockets at our R/C field and a couple drifted out of line of sight. One of the R/C'ers had a small drone with FPV and helped locate one of the models so it could be recovered.

 

[neil_w]

Any possibility of using IR? I'd think that the downed rocket would put off a decent heat signature for a while after landing. Dunno what the cameras on the drones are capable of though.

 

[georgegassaway]

Brian, your video was the one I was thinking of! More trees and different terrain than I had recalled. Definitely the kind of place where airborne video can be useful.

You were doing that using live FPV, right, not blind "point and hope"?

In the thread you linked to, someone mentioned some clubs do not allow R/C. They need to reconsider that. For the reasons discussed here. And also the positive P.R. when drones can obtain some great video or photos of rockets being launched. Best thing is that the NAR's National Sport Launch, NARAM, TARC, and TRA's LDRS have allowed drone photography for years now. So they are leading by example.

I am finally getting into FPV and hope to get some good rocket flight video/pics this year using a Quadcopter. Talked about it at the club meeting last night, about doing so at club launches. Everyone's on board with it. They haven't had any issues with R/C models to begin with.

Any possibility of using IR? I'd think that the downed rocket would put off a decent heat signature for a while after landing. Dunno what the cameras on the drones are capable of though.

Probably not. Even if there was any good daytime-capable IR camera, the rocket would probably cool off too quick, unless the search copter went looking immediately. IIRC, IR cameras do not work in daylight (well, not tiny hobbyist-affordable ones). Now, if a rocket had a flasher onboard, a copter's camera might be able to see that pretty well from a good distance, at night. Tricky thing about that would be what to do once the quad gets to where the flashing light is....... when everything else is black, no point of reference as to WHERE. Well, some drones do have the ability to track their own flight path (either stored onboard or for ones like DJI Phantoms, their FPV tablet/phone software interface would show live where it was on a map). So, at night, find where it is, then the next day follow the map and perhaps fly the drone again at the indicated area to help find it if the model is not easily visible (say in tall weeds, that would have allowed the flashing light to be seen from above at night). BTW - some FPV cameras sold for drone use have different options, one optimized for daytime (IR block), and one optimized for night (no block). They both work day and night, but the image quality is different. The one I was flying was IR Block, optimized for daytime (I had mostly done my 2nd FPV flight, at night, because..... I didn't want to wait for the next day to try it again.

I almost tested something like that. I was making my 2nd ever FPV flight, and it was night by then. I could see enough to be able to make things out so as to fly without getting lost, and I wasn't flying very far anyway. Biggest challenge was flying back to me.... as at first I was not visible. But I had it flying to the general location where I was, and finally saw myself. But what I planned to do next was to take a bicycle rear flasher light, and set it on the ground next to me. So I'd probably be able to easily see that flashing light. The next time I flew at night, I was going to do exactly that and see how well it worked. And then I broke my left arm and all that got put on hold.

 

[ksaves2]

Landowner preferences have to be followed. Our local site the prefect issued an edict on no drones, no R/C, period. I didn't question. The last MWP had a drone flying doing some really nice photo work and the drone was expertly flown and
didn't seem to interfere with anything. I was impressed. The video links are on TRF. Kurt

 

[TangoJuliet]

FYI - When doing Helicopter Search and Rescue (SAR), we use a search pattern to cover the area being searched in a methodical fashion, either concentric squares, or an out and back grid like reading a printed page vertically, left to right. It keeps you from covering ground you've already been over and keeps your eyes focused to a more specific zone. You'd be surprised how difficult it actually is to see an object from the air.

 

[Flyfalcons]

I was with a friend of mine when he used his quad to locate a downed plane in some very tall grass. It worked well.

[video=youtube;LD1RPZAwhxI]https://www.youtube.com/watch?v=LD1RPZAwhxI[/video]

https://www.youtube.com/watch?v=LD1RPZAwhxI

 

[Steven]

I used to have two Kavan Jet Rangers which could yank several pounds of payload in the air with no sweat.

 

[Peartree]

I've lost a couple rockets and thought that it might've been nice to have a copter (or even a balloon with a camera) to help look for it. In cases where rockets are hung up in trees, rather that trying to lift the rockets, it would often be nice to simply be able to drop a light line over the branch that could be used to pull up a heaver line, etc. Alternatively, could a quadcopter drop a fishing line (with a hook on the end) over a shroud line?

 

[dhbarr]

Dropping some braided fluoro would be the way to go, use that to pull real cable up.

 

[Peter Olivola]

This is standard SAR practice. I would add that the separation between tracks must be determined by the viewing angle of the video. There will be a trade off between increasing the viewing angle by flying higher and the ability to see something. One possible way to overcome this is to use a wide angle lens on the video. Permits wider coverage at a lower altitude. Video has some advantages over the human eye (low light vision primarily,) but it can't easily duplicate the ability to look broadly and then focus narrowly or simultaneously scan during the track.

FYI - When doing Helicopter Search and Rescue (SAR), we use a search pattern to cover the area being searched in a methodical fashion, either concentric squares, or an out and back grid like reading a printed page vertically, left to right. It keeps you from covering ground you've already been over and keeps your eyes focused to a more specific zone. You'd be surprised how difficult it actually is to see an object from the air.

 

[Kelley]

FYI - When doing Helicopter Search and Rescue (SAR), we use a search pattern to cover the area being searched in a methodical fashion, either concentric squares, or an out and back grid like reading a printed page vertically, left to right. It keeps you from covering ground you've already been over and keeps your eyes focused to a more specific zone. You'd be surprised how difficult it actually is to see an object from the air.

That is when you need the CAP guys with SAR training around

 

[Winston]

A drone with a good live video feed costs at least $500, the better ones closer to $1000. Such as DJI Phantoms

Second best and much cheaper - put a cheap Chinese HD action cam pointed downward and fly a search pattern at 400 foot altitude (the FAA's limit from what I understand - may have changed for drones as I know it has for RC sailplanes) or whatever altitude proves to be best for spotting visibility and ground search coverage vs flight time and review the video on a laptop in the field. Flight duration is a potential issue, but battery size mods are not uncommon. RCGroups.com is the pace to go for most anything RC related.

Full autonomous flight mission with the Cheerson CX-20 auto-pathfinder

https://vimeo.com/97056189\

Cheerson CX20 CX-20 Open-source Version Auto-Pathfinder Quadcopter RTF ($216 shipped)

https://www.banggood.com/Cheerson-C...der-Quadcopter-RTF-p-932145.html?rmmds=search

On some TV program, I saw a segment about a guy who uses a lightweight foam flying wing to do this by flying programmed autonomous search patterns. He then reviews the HD camera video in the field on a laptop. He found a child lost in the woods by doing exactly this after massive ground search efforts had been unsuccessful.

For launch sites with difficult terrain where rockets are typically lost, even much cheaper entirely manually controlled drones could be used to accomplish the same thing by just making a pass over those areas.

This one has been around for a long time. There may be something even better now since I haven't been paying much attention lately (I believe there are some equally large brushed motor (i.e cheap) quads with a very useful altitude hold function now), but it is/was extremely popular, as a result having a huge user base for support, and it can carry an action cam although for flight duration improvement a Mobius or more modern equivalent would be better:

JJRC H16 YiZhan Tarantula X6 IOC RC Quadcopter Without Camera ($55 shipped)

https://www.banggood.com/YiZhan-Tar...ter-Without-Camera-p-964161.html?rmmds=search

https://www.youtube.com/results?q=YiZhan+Tarantula+action+cam

 

[TangoJuliet]

"... at 400 foot altitude (the FAA's limit from what I understand - may have changed for drones as I know it has for RC sailplanes)"

There is no altitude restriction any longer for R/C sailplanes that I'm aware of.

 

[jadebox]

You'd be surprised how difficult it actually is to see an object from the air.

It's even more difficult when all you are seeing is live video from a small camera viewed on a tablet. My multicopter has a nice FPV system with a 720P display on a decent size tablet. But, I haven't had much luck helping others look for rockets.

-- Roger

 

[jadebox]

"... at 400 foot altitude (the FAA's limit from what I understand ...)"

The 400' limit is an FAA guideline, not a requirement. The AMA safety code, however, generally restricts flights to 400' or less. And, of course, flying higher just makes it more difficult to see something on the ground.

-- Roger

 

[Winston]

"... at 400 foot altitude (the FAA's limit from what I understand - may have changed for drones as I know it has for RC sailplanes)"

There is no altitude restriction any longer for R/C sailplanes that I'm aware of.

Yeah, I know it was waived for sailplanes since there wouldn't be much point to the sport if they didn't (other than slope/ridge soaring), but at least there used to be a pledge on the FAA's amateur RC pilot registration (which is NOT "drone" registration as it's too often called since it is the pilot that is registered, not the aircraft and not just specifically "drones") page to not fly over 400 feet. Still there, but they've now just limited that rule to commercial RC flying although they've noted even that as being possible to get waived:

https://www.faa.gov/uas/getting_started/

 

[Winston]

For sites where look-down recorded searches might not be productive due to a lack of adequate distinguishing features to provide visual cues of the aircraft's approximate position and since airborne GPS and the need to record that location on the HD video being recorded can raise the overall cost significantly, the following are potential, inexpensive fixes.

Use a fisheye lens on the downward pointed camera and fly at a lower altitude. The fisheye image will then provide relative bearings from objects seen on the horizon (if there aren't any at your site, this fixes nothing). Or use a second forward pointing camera with a DIY first surface mirror image splitter (mirrors at 90 degrees to each other) in the field of view. Crude viewing sync could be achieved in the field by viewing both videos simultaneously in two viewers.

Or use as your search camera one of the HD dashcams with built-in GPS which record those coordinates onto the HD video. The location and orientation of their internal GPS antenna would need to be taken into consideration. It would be best to mount the camera forward of the drone electronics to give it a clear sky view. The imbalance would, depending upon the multi-rotor craft used, be adequately compensated for by the flight controller by the automatic throttling up of the forward motors to level the craft but, especially with large BRUSHED motor aircraft, it would be best to instead shift the aircraft's battery location to shift the CG back to its original location.

Amazon search for HD dashcams with GPS:

https://www.amazon.com/s/ref=nb_sb_...nics&field-keywords=dashboard+camera+with+gps

 

[jadebox]

DJI's quadcopters record GPS-based location data in an SRT (captioning) file associated with the video. So, you can easily determine the location at anytime during playback. Another way to record a location would be to take a still photo. The EXIF data in the jpg image file will contain the location.

-- Roger

 

[Zauskycop]

I bought myself a DJI Phantom 3 Standard (live video feed, GPS) and paid $399 at Best Buy. Also had a $125 gift certificate...so $275. $1000???? Hardly

 

[Winston]

Search and Rescue (SAR) Drones site:

https://sardrones.org/

UAVs For Search and Rescue Operations (several theory/concept PDF links)

https://rcadvisor.com/uav-search-rescue-operations

A masters degree thesis - "Examples of both hardware and software approaches in the design process for an aerial imaging system are shown for small UAV applications.":

https://scholarsmine.mst.edu/cgi/viewcontent.cgi?article=8536&context=masters_theses