auto altitude descent?

[wsmalley]

How could one fly a plane down to a point, or specific altitude, automatically. Stopping short of an automatic landing system, being able to use throttle but automatically guide to a point on, or near, the ground? First thought was a laser beam, but that's so small, one would have to be dead on the beam. Another option might be using waypoint(s) to line up the approach and somehow 'stepping' the altitude-like a staircase with some type of pre-programmed altitudes. Any thoughts on if, or how, this might be done?

[tnorris]

The easiest way is to have waypoints. Each waypoint will have a pre-determined altitude. That of course will only work if you have GPS or your plane has a transponder and altitude encoder that can feed info to your autopilot.

[Medve]

well, the altitude could be controlled with a reduction in power at specific waypoints. Most likely while you circle over your landing area, but a linear approach would work as well. If you have a controller that is capable of triggering a mission servo at certain waypoints, it's doable. The other idea I'd had was having the mission servo kill the motor completely, and trigger a parachute to land the plane. Not pretty, but works without too much extra equipment and cost.

[Steve McBride]

I would not rely on GPS for the altitude though - even the best can be off enough to cause a very bad 'landing'. If you are going to have to build somethign custom, you could use the GPS method and also use ultrasonic rangefinding to measure the distance to the ground. When the ultrasonics pick up the ground, they would take over the reduction of throttle and flair.

Have any of you seen the 'easy lander' motor glider? Used a sort of 'de-thermalizer' where the elevator pivots up and the model comes down flat - I suppose you could adapt something like that if the direction could still be controlled.

Just some thoughts.

Steve

[wsmalley]

I think this is like trying to devise an 'automatic' localizer system for an ILS. You would, I guess, have to have a 'window' altitude, that is, plus or minus so many feet, with the horizontal position being the last waypoint or two. A contoller circuit would have to then read an 'X' and 'Y' axis to fly to a point on the ground. Guess it could be a small radio Tx, light beam, or other electromagnetic 'target'. To be practical, I would think you'd need at least a 1000' range. Unless one had miniature DGPS receivers, I doubt GPS altitude data is a help. Can't seem to figure this out..........! Bill

[clolson]

I haven't tried anything like this, but I was wondering about getting some airplane that would survive or even enjoy flying the glideslope right into the ground. (I think my Kadet Sr. could do it.) Combination of slow enough decent rate and the right type of plane and I think it could be worked out. Then setup for a gentle glide slope and aim maybe 1/3 down the field to account for gps position errors. You might need an autopilot unit with some programability though to pull it off, and a big enough field to make up for the inevitable gps errors. I've seen some cheap sonar sensors and have wondered if you could start picking up a "radar altitude" when you get down to the last few feet in order to program a flare?

[Sparky Paul]

Polaroid makes (or made) a range-finder module for their cameras that has been used to display altitude in a close-to-the-ground situation. The output of this could be used to flare the plane.
"Dethermalizing" when near the landing spot would also work. Wouldn't work so well if deployed at an altitude.
http://www.dfrc.nasa.gov/gallery/Photo/Schweizer-1-36/HTML/index.html

[Steve McBride]

Polaroid makes (or made) a range-finder module for their cameras that has been used to display altitude in a close-to-the-ground situation. The output of this could be used to flare the plane.
"Dethermalizing" when near the landing spot would also work. Wouldn't work so well if deployed at an altitude.
http://www.dfrc.nasa.gov/gallery/Photo/Schweizer-1-36/HTML/index.html

That is the system that I was referring to. The modules are readily available and very accurate.

The easy lander type of determalizer would be a good system to look at. I guess they are all pretty similar, but bring the model down like a pancake if I'm not mistaken. No real control though.

After playign with my first FMA co-pilot yesterday, I'm fairly certain that with the right throttle control, a control program could be written that uses the co-pilot as a sort of glide slope modulator. You'd have to have a lot of space, but would be cool to see.

Steve