UAV needs

I'd like some feedback on Unmanned Autonomous Vehicles, particularly of the aviation kind. Are RC Servos the ultimate, or would digital PID
control of gear motors be a desireable thing?
--
Randy M. Dumse
www.newmicros.com
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Real UAV's or toys?
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Martin Euredjian
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

I hadn't given the distinction much thought. We have a number of customers who've bought IsoPod(TM) for their UAVs. Most of them are working on hush hush projects and we can't get much information on what they're working on, just that it flies, so they must be military applications. Since some asked questions about RC Servos, I figure they're using them in real UAV's. But even more so for toys, I'd think the RC Servos were a pretty weight-efficient solution. I was just wondering if RC Servos were the way to go for all UAV controls, or if other servo loops were in common use.
--
Randy M. Dumse
www.newmicros.com
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Randy,
Years ago I had the pleasure of being involved with UAV's. I'd say you can roughly classify them into two groups. Keep in mind, this is rough and very general, there are exceptions, of course.
The first group could be the "proof of concept" stage. These might be thought of as sophisticated model airplanes designed and built to test whatever concept is being studied. For example, there was one that tested the idea of a flying plank (different from a flying wing) ... the wing/plank had, if I remember correctly, some 18 servos and a very sophisticated control system. There was another model I remember that was built to show the feasability of sending a folded autonomous aircraft to Mars. The plane would be folded-up in a canister that would pop open on descent. The plane would then unfold itself and go about it's mission.
These "proof of concept" models can, and usually use, whatever is available off the shelf, like your IsoPod, RC Servos and the motor controls I was manufacturing at the time.
The second category of UAV's is more like a real aircraft. They have to have quantifiable characteristics, reliability and duty cylcle (just to mention a few criteria). This family can span a wide range, from weather research planes to target drones and (as we've recently seen) unmanned recon and attack planes. These can't generally use such things as RC servos because they are not designed to industrial standards. You have to remember that, as good as they are, they are designed for toys. Having flown RC model aircraft most of my life, I can tell you that, as an engineer, I wouldn't think about using RC servos for anything serious. They simply lack the robustness, speed, accuracy, protection, monitoring, feedback, reliability and survivability that you might want (or need) in a real UAV. Also, remember that most of what I'm calling "real" UAV's are large enough that RC servos couldn't possibly be attached to their control surfaces.
Back when I did some work on UAV's I saw a distinct transition between the "proof of concept" stage and the "real" stage. I'd supply a lot of cheap off-the-shelf motor controls for the first phase and, later, when they built the real thing, I might be hired to design a much more robust control with very specific requirements.
I don't know much about you IsoPod, I've never seen or used one, but I would suspect that it does not have things like multiple redundant processors, lighning and static damage protection, over/under current/voltage protection, RFI or EMP protection, etc. I had to deal with some of these issues back then, the standard of performance when you have a real aircraft flying over people/property/battlefields is definetly different from that of $200 model airplanes with $15 RC servos that can crash for no reason (just saw three crash this weekend within ten feet of takeoff for no apparent reason).
Just to be clear, a ton of research is done every year in aeronautics using every-day off-the-shelf RC hardware and wonderful products like your IsoPod.
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Martin Euredjian
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

The one's I've opened all have a single pot for positional feedback, I gather 5K and linear. I was surprised how well they do seem to work.
--
Randy M. Dumse
www.newmicros.com
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

I
Yes, and I believe that the motors are supplied with an almost constat voltage until they reach the desired position, which is sensed by the pot.
I suppose a digital feedback is necessary whenever one wants to controlle also the velocity of the servos not just their position.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
FYI Dept. stuff...the motors are pulsed, basically in a PWM fashion, no matter what the speed. The frequency of the pulsing is quite low, and is derived from the controller or receiver operating the motor (usually 40-50 Hz), via the signal line. This is readily seen if your controller sends say, just 10-15 Hz pulses. The servo operates in a herky-jerky fashion.
Even the digital servos work this way, but the pulse rate is internally regenerated. Even though the servo may get a 40-50 Hz pulse streem, the motor is being pulsed at 200-300 Hz. The higher PWM rate is is partly what gives the digital servo more power. (Not the only reason of course.)
Digital servos still use analog speed control techniques. The term "digital" is really a misnomer. The MCU in a digital servo is basically supporting a programmable analog circuit (as opposed to operation set by the values of discrete components).
The one "disadvantage" to digital servos is that they cannot be turned off simply by removing the pulses on the signal line. This is often done with servos that have been modified for continuous rotation. Turning the pulses off stops the motor; you don't have to send exact centering pulses (the exct value is really dependent on temperature drift and other factors anyway). With a digital servo, if you remove the pulses to the signal line, the last good pulse duration is just repeated.
-- Gordon Robots for Less at Budget Robotics: http://www.budgetrobotics.com/ Author: Robot Builder's Sourcebook & Robot Builder's Bonanza
Dan Oprisan wrote:

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Digital servos for R/C are now available.
wrote:

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Yes, I googled and found this: http://www.futaba-rc.com/servos/digitalservos.pdf So, they use a microcontroller in the feedback loop for optimizing the dynamical response of the servo (e.g. the motor gets more power at startup, to overcome the inertial forces, etc.).
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
You can make them yourself out of cheap analog servos. Colin MacKenzie over at http://www.colinmackenzie.net/techlib/robots/Symapod.html even built a hexapod walker using his own digital servo controller design. If you e-mail him about it, he'll gladly send you code and schematics to build your own. His design uses a PIC microcontroller, the I2C bus to minimize wiring, plus he found a way to measure force feedback. He says you could upgrade to digital for around $20 if you construct them in high enough quantities.
wrote:

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.