I may be barking up the wrong tree, but this technology seems pretty interesting. I need a drone with the following attributes and I'm not sure if one exists or if someone is able to create one.
1. Ability to fly an HD video recorder (or several). The recorder itself needs to broadcast a real-time image to a monitor. There must be remote controls to allow for in-flight camera adjustments such as zoom and panning.
2. The drone itself needs to be able to fly/hover accurately outdoors for at least 20 minutes at speeds of up to 30 mph at between 1 and 20 meters altitude.
3. It must be nimble. It has to be able to quickly change direction.
4. It must be able to follow a pre-determined path while following a target. I'm not sure if there is software that can be incorporated with the camera, but the target needs to stay in the picture. The more this is fully automated by the drone, the better. The target can either carry a homing device or the drone must be able to "recognize" the target.
5. It needs to be able to avoid unforeseen obstacles such as limbs and saplings for flying through the woods.
6. the drone must be able to communicate power level.
Let me know what you guys think.
Replies
Sounds like the limitations of follow me is sensor related, and can't cope with the "fine detail" of wooded trails.
However...how about road cycling? I have a specific requirement to track and follow a rider down an alpine road descent. Would the "resolution" issue still be a problem? Does the sonar and optical help with positioning at all?
In an ideal world, the camera would be gimballed to track the riders sensor, allowing the 'copter to wander a little. But as I'm finding out, the sensors for the "follow me" are the tricky bit...(I'm a total newb btw).
Sorry I'm late to the discussion...others have already alluded to the point I'm making...
It sounds like you want Garrett Brown's Skycam, except without the cables.
It would take a lot of UAV horsepower to replace the lifting and controlling
capabilities that comes with motors and cables and pulleys.
-peter
With the exception of
4) Pre determined course (see alternate method below); and
5) obstacle avoidance this can be pretty easily done.
Zero UAV's YS-X6 multi rotor controller has a WiFi based "Follow Me" feature that will follow someone with an iphone in their pocket, so if it's an outdoor course and you can set the altitude to be say 5 meters above the tallest tree, then your goals can be achieved now.
http://www.zerouav.net/newsmore.aspx?id=77
Heres a test we did with an iPhone strapped to my Wookong Octocopter whilst a YS-X6 follows behind:
https://www.youtube.com/watch?v=mNPDD8Q1fHc
Cheers
Matt
Sounds a bit risky for MTB races where there are crowds next to the course.
FPV is more fun.
I'm wondering,
R_Lefebvre's comment made me think...
Pre-planning/prepping the flight path to double check for small obstacles *twigs etc* if it is a known path and clearing,
then setting up a pre-defined tight geo-fence on each side of the path for the length of the course, and using alt-hold? Like R said then maybe all that is required is a forward/back speed control and a decent gimbal, anyway just thinking out loud, haven't done any work on geo-fencing so not sure how possible this is.
The ability to fly through the woods automatically is not going to be possible this year. Next year, maybe. 2 years, quite possibly. That's my prediction.
What you need would be some sort of LIDAR based environmental dead-reckoning. Not gonna happen this summer.
Why can you not fly the machine yourself? it would save a load of hassle...
1. Yes.
2. Yes.
3. Sort of. Bigger means less nimble and that also depends on your definition of nimble.
4. Sort of. Most likely it would have to be piloted manually to get good enough shots to use in production.
5. No. Maybe check back in 20 years?
6. Yes.
I think this is going to be a bit of a tall order, so you may need to scrap some requirements unless you have a very large budget, like 10k USD+. Most multicopters struggle to hover more than 30 minutes with no payload.. add a payload (and remove some extra batteries) and then add in the 30mph flying and flight time will decrease significantly.
As far as I know there are no commercially available multicopters that have onboard obstacle detection and avoidance, so that will have to be done manually by an operator watching the live video stream.
Most multicopters have the ability to transmit power level (roughly), are nimble and several can follow a target based on a beacon including, I believe, Arducopter.. I think it needs some hacked version of ardustation to work though.
So you're definitely asking a lot; everything you listed is technically possible but doesn't exist all as one device yet.