Replies to This Discussion

Permalink Reply by Oyvind Dahl on July 11, 2012 at 11:36am

I've sent a friend request in order to send you a message.

• ▶ Reply

Permalink Reply by Jake Stew on July 11, 2012 at 9:59am

I think the cost would have to be pretty low ($100-200) to make this work.  

We already have sonar which is cheap and accurate, and works at the ranges we need.  A decent sonar unit can go 10 feet or so, which is good enough for altitude hold and takeoff/landing.  Once you're in the air GPS altitude is good enough since you don't need very precise altitude.

With GPS altitude, barometric altitude, and sonar I don't see that radar would be all that useful.

• ▶ Reply

Permalink Reply by Oyvind Dahl on July 11, 2012 at 11:39am

Interesting.

But if it is windy, does the sonar still work good enough for auto takeoff/landing?

• ▶ Reply

Permalink Reply by Jake Stew on July 11, 2012 at 12:04pm

I haven't had any problems with mine.  With good design they work pretty well.

There's no question that radar would be better, but with the added weight and expense it would really have to meet some sort of specific need for the mission.

One problem I also see with the radar is the minimum distance and/or accuracy at minimal distances.  It would be hard to use it for auto takeoff and landing unless it can work at very short ranges and be very precise.

• ▶ Reply

Permalink Reply by Oyvind Dahl on July 12, 2012 at 3:52am

The hardware we have now can work at distances down to 10 cm with an accuracy of 4 cm. That could be improved even further if that was what would make it interesting for this market.

• ▶ Reply

Permalink Reply by Brian A. on July 11, 2012 at 1:57pm

Hi Jake.
How does your sonar system work for airplane take off and landing... can it guarantee a perfect landing every time even for a rookie pilot?

• ▶ Reply

Permalink Reply by Jake Stew on July 11, 2012 at 3:06pm

I've just used it for telemetry so far.  I haven't gotten brave enough to use it for takeoff/landing and haven't really looked into how the APM is using the various sensors.

Stabilize and FBW modes help a lot for landings and I've been busy testing other things, so I just haven't tried autotakeoffs/landings yet.  The sonar readings look fine though, so there should be no problems with the sensor data.

• ▶ Reply

DeveloperPermalink Reply by Randy on July 11, 2012 at 6:33pm

I'm more on the copter side but my understanding is that the sonar isn't used in the arduplane code for auto take-off and landing although it's been discussed and possibly attempted by Doug Weibel.  My understanding is that it should be ok.

 

I think the thing the radar gives you that the existing sensors do not is a height above land even at high altitude.  Barometer is absolute altitude so no idea how close you are to crashing.  Sonar is good but only at short ranges.  It's this collision avoidance + not crash into the side of a hill functionality that is new.  I think the commercial jets use a radar for this as well and generate a "pull up!" voice command...at least I've heard of such systems.

• ▶ Reply

Permalink Reply by Jake Stew on July 11, 2012 at 11:59pm

Some planes have radar altimeters, but only the military uses airborne radar.

I'm just not sure what the real utility would be.  Sonar can keep you from hitting the ground, do obstacle avoidance at slow speeds, handle altitude holds at reasonable distances, is precise enough for landing, etc..

Sonar will do the low stuff just fine, and once you're at altitude does it really matter if you're at 300 feet or 25 feet off at 325 feet?  It's just not that critical.  For me the GPS+barometer deliver plenty accurate enough readings.  The entire surface of the earth is mapped with excellent terrain altitude detail, so it's easy enough to set your waypoints with an AGL altitude.

I'm just trying to accurately answer the OP's question about how much interest there would be and what a realistic price point is.  Current tech serves most needs AFAIK and is fairly cheap, so the competition is pretty tough.  For a radar to be successful it would have to be very good, or at least a lot better than the alternatives, and fairly cheap.  I don't think a $500-$1000 radar altimeter is going to catch on enough to be profitable.

There's also LIDAR and stereo machine vision in the works.  So there's also going to be competition in the high end of this area.  Personally, I'm betting that stereo vision is going to win out for high end applications because it has the potential to be cheap and has more applications in FPV and mapping.

• ▶ Reply

Permalink Reply by Oyvind Dahl on July 12, 2012 at 4:15am

I have gotten the feeling that people working with UAV's are quite technologically capable people (at least on this forum).

We have started playing around with the thought of offering an open-source hackable radar altimeter. It would come in the form of a radar module with software for providing general purpose altitude measurements with a certain specification.

For those who have special requirements like longer ranger, faster update speed, better accuracy, collision avoidance or things like that, they would be able (with a bit of programming knowledge) to customize the device to their own preference.

Do you think a more OEM radar module with customizable altimeter software would be a better fit for the UAV market than a finished closed box that just measures altitude and nothing else?

Oyvind

• ▶ Reply

Permalink Reply by Don LeRoi on July 12, 2012 at 7:30am

Oyvind,

Yes.  I'm salivating!

- Don

• ▶ Reply

Permalink Reply by Jake Stew on July 12, 2012 at 11:33am

That actually does sound pretty good.  Experimenters would probably get it doing some cool stuff and that would really add some value.

• ▶ Reply

• ‹ Previous
• 1
• 2
• 3
• 4
• 5
• Next ›
• Page