Drones and Slung Loads?

Recommended Posts

I'll ask out in public what I have been asking a few people in private: Anyone using slung loads with UAVs? For background, I am asking because we work in improving the accuracy of predicting slung load aerodynamics (all sorts of shapes!) speed and safety. One advantage is immediately apparent: You put a large pizza box (quite light) under a small quadrotor and you block the rotor wakes. You sling it 10 feet below or 20 feet below, and the drone does not have to come down to whether people/pets and drones won't hurt each other, and you can deliver the pizza without even hovering. But obviously, that's not the only application. The work originally came from DoD interests and also from flood evacuation interests for large and precious payloads but is now aiming for a much broader range of civilian applications.  I am getting some good feedback (positive and negative both deeply appreciated!!) on problems and opportunities.

Link to post
Share on other sites

It’s a bit of a challenge to think of how using a sling on a multirotor would improve its mission capability.  That’s not to say that it can’t but if you don’t have any experience with it, it’s hard to imagine.  

A couple of things to keep in mind that might effect performance a bit differently than using a sling with large drone or manned aircraft.  Because multirotors use fixed pitch, low mass props they rely on changes in RPM and to a lessor degree gyroscopic progression to maintain stability, the center of gravity is important because most flight controllers rely on multiple sensors and EKF filtering to keep them level.  

If the sling is too long or too heavy the CG is lowered to a point where the delta between normal CG and the sling is too great and the controller can’t deal with it and you get this undesirable oscillation or “toilet bowl” effect.  That can probably be tuned out but that depends on the PID’s and the how good the filtering is.  DJI’s have a very small envelope and are susceptible to control issues with anything out of CG.  PX4 provides a lot more tools for dealing with this but you’d better be good at PID tuning and have a PhD in MissionPlanner...

In the early days of flight controller develepment there were lots of online competitions where people would see how long and how fast they could sling load a cinderblock from one point to another.  The crashes were spectacular ?

This could be a fun topic, you should consider starting your own thread.   It doesn’t have much to do with thermography.

Link to post
Share on other sites

Thanks, all very helpful!  FYI, we have flight-tested a small (<$200) quadrotor with a 4lb box some 15 to 30 feet beneath it in a casual "let's try this at the park b4 anyone wakes up) and gone at fair speeds (est. 30mph-plus). Perfectly smooth without bringing in any of our exotic technology. Large pizza box is harder, but doable. Also, now over the terror of OMG! A Tree! Nice sound the rotors make as they cut through foliage in front of our (organization's) President's Office. Perfectly legal, mind you, flight only above Green Space, no humans underneath.

One observation: Current prohibition of slung loads over populated areas is based on human-piloted vehicle rules that require a sling that can be "pickled". Sensible, but with drones that is not essential. Drone crash? Load crash? all the same. You can probably use load aerodynamics to set the drone gently down if still tied to it.

Yes, I will be delighted to start a new thread (hope anyone else also shows interest). Relevance to thermography? I wonder about thermography from 10 feet above the ground versus 60 feet above. Unlike those spectacular images people post, one does not need pixel-level precision. Some work I see, depends on catching plumes rising, but from closer to the ground one may be able to sense much better. Yes, the motion of the sensor is an issue but we know how to deal with that for much worse motions. These motions are clean periodic, I believe we can get essentially perfect recovery of the images even with the long exposure times of thermography. Hope this makes sense.

The other reason to post here is that our actual interest is in the business aspects of drones with loads. Tired of writing the tech stuff in Peer-Reviewed Papers etc, got loads of that. Greatly hoping for any guidance on the business lingo such as Value Propositions, hypotheses, Channels, Customer Segments, reaching customers in this business, conventions and trade shows, advertising, pricing, customer service. All related to the drone business. Thanks!

  • Like 1
Link to post
Share on other sites

Thanks! Alan, hoping to get some comments from you.. :) From what we see, slung loads are typically associated with

a) big military cargos (CONEX, HumVees, all sorts and shapes of other things) that have to be moved fast over places where you don't want to dilly-dally

b) Moving a groggy rhinoceros (Google that one, I am not kidding) or skycrane applications where slow and steady is the order of the day as you lower the top half of a tower onto the bottom half for an Alaskan WiFi network.

c) Logging ops where pilots have learned to do dynamic, swinging loading onto a truck without pausing (also Google that, it's eye-opening)

d) fire-fighting in Idaho/Oregon

We are aware of all these, but think that a vast number of other applications arise. A dreaded scenario is another Hurricane Katrina: 1000 people standing on roofs, water rising, a poor guy or gal having the job of Signing Off on the Safe Speed for the Empty People Carrier, or, even worse, the loaded one. Do you stay safe and get blamed for 200 people drowning? Or be brave and lose a vehicle and crew and maybe a whole family?

I think we can handle that.

But now with drones, all sorts of other scenarios come into play, which is what we are exploring. ("We" by the way is a small team of researchers/maybe entrepreneurs).

Plus, the whole idea of deriving accurate data as your sensor/camera swings in large arcs or cones. Fire away please, we're used to being called crazy (that's very nice compared to other things we're called), and we learn from all such comments.

Edited by Slinger2
Link to post
Share on other sites

What’s the advantage and how does swinging a camera in a large arc make for accurate data?  How do you control the speed of the arc?  

Most of the cameras people are using for photogrammetry are CMOS and suffer from rolling shutter and if you want the camera to travel on a large arc I’d think motion blur might be a factor?  

Link to post
Share on other sites

Great questions, thanks.

1. The speed of the arc can be controlled by UAV motion, just like the logger helo pilots have learned to deposit their logs accurately into trucks. Takes a bit of practice, but we have tried this in the lab with a simple kinematic arrangement, and within days students learned to execute some amazing maneuvers, very precisely : like sticking a cellotape piece to a doll at the end of the swing and thus grabbing it, without it falling off or being pushed away. One would presumably add sensors, then again, cellphones today already have those sensors! 

2. Post-processing will be required, but even with the visual inspection business, I see that you have to do a good deal of that (BTW, yes, the resolution on those is mind-boggling, thanks). The large arc (conical) allows you to do the job of sweeping a much larger distance than by flying over every bit of it in lawnmower pattern. May not be an advantage for things like checking every connector on a power substation, but for sweeping over vegetation in a field this may be highly productive. Does one really care whether one is going over Cabbage Martha or her neighbor Cabbage Agnes?  :) You will get a good deal of redundant images, true, but that also helps accuracy. Averaging does wonders to remove noise... 

On the whole, the precise motion history of the vehicle and the end of the sling are both predictable and measurable with good accuracy and precision, not much doubt about that. We are learning to do the image correction, but I don't believe there is anything original there. Must be established lore in the spyplane/recon communities.

3. Motion blur is a good question, I have to think more about that since I frankly do not know why CMOS would be a disadvantage. Since most operations are in bright sunlight, I thought it would be possible to use short shutters, but I may be ignorant there. Is it any more difficult than sports videography? I was watching the SuperBowl on my MacPowerBook, and the imagery was so amazingly sharp!! In the thermography application, I think motion blur may not be an issue because one is usually not trying to get sharp outlines (ignorance bell going off?)

Link to post
Share on other sites

A couple more things to consider, the missions we fly are autonomous, we also don’t scan NADIR, we scan Obliques.  Our shutter speeds are never below 1/800, we fly 7-10m/sec-I’d imagine a camera on the end of sling would be traveling at a pretty good clip.       You need to shoot at least f./5.6 so unless it’s a bright day your shutter is too slow and not all missions are flown on clear bright days.  For photogrammetry overcast is better. Also, shutter speed has nothing to do with rolling shutter.  

The accuracy of the scan has little to do with the precision of the flight of the drone and more with the relationship between the sensor and object being scanned.  The better you can account for the variables the more accurate the model.

Link to post
Share on other sites

Thanks. Actually the sling tip need not be traveling particularly fast, it is basically on the same order as the UAV's speed, plus or minus a bit.  I think 1/800 shutter speed is plenty fast enough.  Think oscillation periods on the order of many seconds. One neat application demonstrated in technical papers is that you can use a sling to actually come to rest at a point - and fly a fixed wing UAV in circles around it. Whatever can be done from a fixed wing, we should be able to better with a rotary wing, except fly at high Mach number.

Link to post
Share on other sites
  • 2 weeks later...

Chiming in a bit late here...I'm new to drones, but I consider myself an expert on slinging loads. As a Navy H-46 crewchief, I have slung thousands of loads and millions of pounds of cargo. We called it Vertrep. The advantage was that we could move lots of cargo fast, as opposed landing each time and loading cargo into the helicopter. Aerodynamically our biggest worries wear the load swinging into the rotor arc, swinging loads changing the center of gravity,  and how the power inputs affect the flight path of the load, The aerodynamics of the load was also affected by where we were in our flight path, ie wind shear across the flight deck/mountain LZ, take off/landings, and especially during Emergencies were the emergency dictated how we flew-sometimes causing us to fly in ways that were not "standard".

In order to optimize the speed of slinging loads, I think drones would have to utilize an auto-loading/auto-releasing technology. Quick auto-loading should be easy to make happen. The tricky part would be a release mechanism that is remotely operated. An erroneous signal could cause the load to be prematurely jettisoned. This would be a safety concern at the least.

Well, that's my experience...hope that gives you some ideas.


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.