Replies to This Discussion

Permalink Reply by Jelle on August 17, 2008 at 9:15am

You are right, until something changes. If you add some gear the sinkrate at a specific setting will be higher, or you are landing at a higher elevation, or...
Landing is more then setting for a constant sink-rate, you will also want to point into the wind (speed is not that important, the direction is) and flare before touchdown. With a light enough model a constant sinkrate might do, if you figure out a way to hold a high-alpha flight (use power to keep the nose and speed up) and how to sense that you have touched down. Next is to autonomously taxi back :)

• ▶ Reply

Permalink Reply by Krzysztof Bosak on April 21, 2009 at 3:17pm

"and how to sense that you have touched down"
that one is easy. main battery connector will usually unplug.

• ▶ Reply

Permalink Reply by Paul on August 10, 2008 at 11:55am

Our UAV's that we use in the Army don't even have Auto Land with the exception of the Raven, and that is even a controlled crash. Raven Pilots go through a week of training and most of it is emphasized on using the sensors. Most Raven Pilots don't get to fly too often, so when they do, they rely heavily on the Auto Pilot to do it, and Landing is pretty Push-Button. The Bigger UAV's whose Pilots have a dedicated MOS, or Job Title of UAV Pilot, Land their birds Manually. I've shown our UAV operators how to land the Raven manually, and have taught them to bring them in a lot smoother than the Auto Land Feature does. Just my 2 cents

• ▶ Reply

Permalink Reply by Wyliest on August 17, 2008 at 8:03pm

If you know your ground speed and true heading i.e. (GPS) and your air speed and the orientation of the aircraft with respect to true north i.e. (compass minus your local magnetic declination). The difference of these two vectors is the wind speed and direction. This is not so complex, the only, possibly, new sensors are airspeed and compass and some computing power. None of this is to hard to do.

• ▶ Reply

Permalink Reply by AludrA_55 on June 9, 2009 at 10:42am

i don't know what you have already set up in your plane...
but I think (newbie) that if you could get the speed of your aircraft with respect to the wind
and also get the speed of your aircraft using GPS, then (aircraft_speed - gps_speed) should
be approximately the wind_speed...

-aludra

• ▶ Reply

Permalink Reply by AludrA_55 on June 9, 2009 at 11:28am

i got another idea! =)
(not quite familiar with airplanes because my project is a helicopter uav)..

#1) figure out what is the minimum throttle and elevator settings when you land
your airplane manually...

#2) create a PID that controls the altitude of your plane (assuming you got accurate
altitude readings)...

#3) enter the first waypoint to orient your airplane to the landing point then
enter a second waypoint as your landing point... put enough distance between the
first waypoint and the landing point so that the airplane could turn to approach
the landing point...

#4) from air, the airplane would try to reach the first waypoint in any angle
once it reached that waypoint, it would go to the second waypoint... this will
automatically align your plan properly to land on the second waypoint...

#5) enable the PID that controls the altitude of your plane (altimeter must be accurate)...
this should slowly descend your plane to the ground...

#6) then as you approach the landing point, the PID for descent should also control to
approach the settings you took for #1 (throttle and elevator setting)...

of course this is probably a setting for a perfectly tuned PID with good altimeter...
but the whole idea is there...

i don't think that you should worry about windspeed because with normal flying, the autopilot
is still capable of reaching its waypoint without paying much attention to windspeed...
if you ask me, a regular aerial waypoint is basically similar to a landing waypoint, the
only difference is the altitude...

anyway, just an idea.. (don't blame me if your airplane crashed!, i'm still new!)...

-aludra

• ▶ Reply

Permalink Reply by Joseph Glanville on June 10, 2009 at 6:19am

Hmm, I was thinking a very similar idea.
Use some kind of preprogrammed landing approach.
Ideally accurate airspeed and altitude are required but it's quite possible.

• ▶ Reply

Permalink Reply by Tom in ON on June 10, 2009 at 7:59am

Newbie myself...so welcoming any suggestions.

What would worry me using a GPS waypoint approach is that the location error at the waypoints could be +/- 10 meters or more, but probably not less than +/- 2 meters. Not worried about the altitude, but the accuracy of the landing vector between connecting the two waypoints that supposed to line you up for landing.

You may need 3 waypoints. You would need to have the first two waypoints far apart that the maximum possible location error can be reduced to a vector that will still take you close enough to the start of the runway to a point where your next vector to waypoint 3 and the associated maximum error still allow you to land on the runway. Albeit, you may not land exactly on centerline of the runway. However, if the runway is wide and long enough and you estimate your rate of GPS error corectly, you should have enough space to land safely.

Another way may be is the line up the best you can using the above method and when you reach wayoint 2 and proceeding towards waypoint 3, use a compass to keep you straight on the landing vector (assuming you are lined already with a vector that will give you enouhg space to land on the runway surface).

As far as the descent. May be able to land by using an accurate altimeter (like on based on the SCP1000), or even the GPS altitude reading, if you designate a proper glide/landing angle (and the rate of descent either fed to the aircraft controls by a SCP1000 type altimeter or a single axis acclerometer) from waypoint 1 to the desired landing spot, assumig the runway is long enough to factor in the the GPS error induced vector possibilities. The autopilot should help you mantain the proper rate of descent for landing mitigating the wind gusts.

• ▶ Reply

Permalink Reply by Tom in ON on June 10, 2009 at 8:41am

• ▶ Reply

Permalink Reply by AludrA_55 on June 10, 2009 at 7:21pm

i've seen some people could pilot their airplane via transmitter and make their airplane hover by flying it vertically so that it looks flies like a helicopter... i know that it kinda sound retarded but that could be another way to safely recover your plane if your propeller has enough thrust to hover your plane vertically... =)

• ▶ Reply

Permalink Reply by Kim Skatun on June 24, 2009 at 3:31am

has anyone built a shield with airspeed and compass module and tested out the 4th method mentioned earlier in this thread? Its the most accurate way to do it. I think Papparazzi have done sth similiar to that.

• ▶ Reply

Permalink Reply by High Society on June 24, 2009 at 5:57am

The UAVs that I have built and fly use two waypoints. The first waypoint is designed to align the aircraft on the glideslope and to adjust the aircraft altitude to around 125 feet. It then proceeds to the second waypoint (touchdown).

There are cross track algorithims that keep the aircraft from drifting from the path so wind direction is not as big of an issue to me because it will just crab all the way down. There is also a better method that I am not too familiar with: Vector Field Path Following.

The bigger issue I always have is the altitude accuracy of the baro sensor ( better than gps though, ) so to compensate for that I have seen people integrate laser rangefinders sucessfully. This is also important for your flair timing and can help regulate your airspeed intelligently.

• ▶ Reply

• ‹ Previous
• 1
• 2
• Next ›
• Page