Comment by Tj Bordelon on September 20, 2010 at 5:45pm

How long are your "runways"? I used to try coming in low like you, but a few things happened. First, my GPS sometimes would be off by 10 meters on cloudy or snowy days and I would hit the ground. Second, my fields are usually surrounded by trees so coming in low isn't an option with 120 foot pine trees around.

I am very curious though on how long you are taking to land from your 10 meter height, and if you are using anything aside from the standard Easystar control surfaces. I'm sure you saw my thread ... I am making some death metal landings these days, from 180 feet up and in (if I'm lucky) 500 feet.

Developer Comment by William Premerlani on September 20, 2010 at 6:40pm

Hi Tj,

Nice to hear from you.

I saw your thread but I have not been following recent comments. I sympathize with your situation, and I admire the strategy that you worked out. Cool.

I am using a stock EasyStar with rudder and elevator only.

We have plenty of room at our flying field, so I was not trying for a short landing. From the turn to final to landing was about 600 feet. However, the technique I described could work for a much shorter runway, because it provides a "ceiling" that forces the plane to come down before it reaches the end of the runway. All you have to do is define a path that intersects the ground where you want to land.

With a 10 meter height, it would be rather easy to land in 50 meters. It would be coming in hot, but it would be level, it would be a smooth landing. Next time I get a chance, I will try it out and see what happen.

Has anyone suggested that you might try coming down in a helical spiral? It shouldn't be too hard stay within a reasonable diameter circle, say 300 feet would be easy, you could probably stay within 200 feet without too much trouble. In windy conditions you could use wind estimation to account for the wind. You could even use a rather shallow glide angle, to keep the airspeed low. You would not be sure where the plane would land, but what would that matter, as long as you placed the circle over a suitable landing area?

I have in fact used that technique manually once in a while to land in a tight spot. For example, one time I was flying very high over a soccer field when a team showed up and started to practice, so I could not land on the field. I found a small patch of grass to land on, I put my plane into a helical spiral, and landed it that way. Of course, this was manual control, but I don't think it would be too hard to program up autonomous controls to do the same thing. The idea intrigues me, I will put it on my list of things to try out some time.

Best regards,
Bill

Comment by Tj Bordelon on September 20, 2010 at 7:20pm

Hey Bill - Yep, I've tried the spiral. I can do it manually too but I discovered flying below the tree line is just bad in general for my AP. First, tight circles pull plenty of G's and very tight circles for >30 seconds start to confuse my IMU. I haven't debugged this fully, but it seems to be a combination of things. And GPS starts to get a bit inaccurate below the trees, so I figured I had better just stay above the trees, dive bomb and land.

Maybe "folding" the landing pattern, where you just do a single turn at one side of the runway. That way you get double the distance. Anyway, it's nice to see someone else getting their EZ STAR landings perfected :)

Comment by passunca on September 21, 2010 at 3:37am

maybe we need to come up with some sort of DIY ILS system.
The ILS helps an airplan to ajust there landing vector by creating some sort of RF landing strip.
So... I think something similar can be achieved by using a webcam at the edge of the landing strip mounted with the desired landing angle.
The webcam (via the groundstation) would track the drone and provide feedback. The drone would be able to adjust its path according to the feedback from the goundstation.

Moderator Comment by Ritchie on September 21, 2010 at 4:55am

Excellent

Comment by Tj Bordelon on September 21, 2010 at 6:20am

I have often wondered if this could be done easily with a 2 big IR LEDs at the foot of the runway. See both = lined up. See more of one than the other = off to the side. I bet it could be done very simply. The webcam sounds workable as well, providing you want to land by the laptop. Which in my case is usually where I want.

Comment by passunca on September 21, 2010 at 7:11am

well you can put those IRLEDs on the drone and you only need to remove the IR shield from the webcam and puff that damn thing will light up on screen like nothing else.

Comment by Tj Bordelon on September 21, 2010 at 7:42am

I like it! That would get your landing SPOT ON. My only regret is that I have made my ground station so simple. I have no laptop, only a very low power micro and 256x128 green screen display. I'd have to do a bit of work. But that kind of system would probably allow me to land in the 100 feet of space in my backyard, only I'd have to have a net at the end :)

Comment by Tj Bordelon on September 21, 2010 at 7:48am

OH- but remember latency. Ideally you want the IR camera on the plane so you can keep the RF modem latency out of the loop.

Comment by Rana on September 21, 2010 at 7:59am

Billu Bhaiya, I do agree with you, it is really nice approach for autonomous landing without ultrasonic range finder.
But you would also be having similar views as mine and that is, it only a temporary concoction if ultrasonic range finder is not used.

Best Regards

Rana

Comment

• ‹ Previous
• 1
• 2
• Next ›
• Page