I want to locate the position of a beacon mounted on things like an RC Helicopter or perhaps a person playing sports, surfing etc.
A Range of up too 200 meters or so would be nice.
This is so a bunch of servos can aim / point a camera at the beacon.
However, I would like to point out the following which I think should be excluded.
GPS might not be a good solution, the coordinates would have to be transmitted, this would require alot of electronics, which may not be robust for a sports man (especially when its a rugby player and 10 guys tackle him), and also it might be too bulky.
IR sensing, this has limited range, and might not be good outdoors under bright sunlight.
Ultrasonic sensing, once again, limited range.
As new technologies come out, what are the means of locating an object outdoors at ranges of up to 200 meters. I am guessing a radio beacon perhaps?
Are there any modules which would make this project easier?
Has anyone else come across this one? or conducted a project like this?
I think it will be very difficult to track something like a rugby player because of the fast and unpredictable movements and the likelihood of him being hidden by other players. To say nothing of making an electronic device that can survive in that environment.
My sense is that wireless is not sufficiently directional. It will be essential for the transmitted signal to be visible in any direction as the object carrying it (person, model plane etc) can turn through 360 degrees. What about a light - a bright LED, perhaps flashing, perhaps infra red? (A simple 555 chip might be all that's needed - cheap and robust).
I don't know about the range but I can imagine a "receiver" with a narrow field of view that scans from side to side (and possible up and down) with the camera always following it and when it sees the light it could trigger the camera shutter. This puts all the complicated stuff on the ground where it can be managed easily.
Unless you know how to make a very directional receiver antenna (I certainly don't) I can't imagine that RSSI would be able to determine direction. I wouldn't expect the RFM22 or XBee transceivers to have any significant directional ability as that would be contrary to what most users want from them.
Lights have the added attraction that visible ones make debugging easy at the initial stages.
I think some radar systems locate transmissions by receiving on several different antennae at the same time and measuring the timing differences of the received signals.
I would go with something like the flashing IR beacon Robin 2 has proposed, but make it with five receivers. One central receiver and four receivers that look in the directions around it. The more a side receiver receives the signal the more it moves the sensor in the opposite direction. This is actually how early IR tracking missiles worked, and it makes sure that you don't have to actively scan for a target in order to follow them.
The pulsing should help with outside disturbances as you can notch filter out just the frequency you set, and most sources of IR are close to pure DC. I have wanted to do a project like this for spotlight tracking for a few years now as a senior design project for school, so I have put some thought into it. I was even hoping to encode data in the IR flashes, but that would be a stretch goal for sure.
I like @Jroorda's idea of extra receivers to provide some peripheral vision.
I think it will be next to impossible to see the beacon when the receiver is directly facing the sun. But then you won't get a good photo in that situation either. You could, of course locate your camera/sensor system with the sun behind you. The sun doesn't move fast or erratically.
As far as the encoded signal is concerned I think you should do some calculations to figure out how long it takes to detect and verify the signal compared to how long you want to wait, after the camera aligns with the beacon, before taking the picture. It costs very little to take 200 rubbish digital photos (I do it all the time).
By the way you may need some fancy mechanics to make the camera pan/tilt system move quickly and accurately without shake, vibration or backlash. There would be little point designing a detection system that required better mechanical performance than you can provide.
Thought,
You are only concerned with X & Y coordinates ... A miniature gyroscope-accelerometer sending via RF could provide short-term movement differentials for the camera servos. When the subject is no longer moving (camera has subject centered), the receiver resets the XY coordinate differences to 0, 0 in preparation for the move/play thus ensuring continuous synchronization. In such a scenario, simple trig defines all movements.
If you imagine a device attached to a footballer what starts of as the x-axis could quickly become the z-axis when he is tripped up. And model airplanes move in 3 dimensions.
what are the means of locating an object outdoors at ranges of up to 200 meters.
he x-axis could quickly become the z-axis when he is tripped up.
The key word is "quickly" which represents only a small displacement (6 feet or so) toward the ground. From 600 feet away, using a telephoto lens, the "X" displacement is truly small as a percentage of total frame, X * Y (ground area captured in video frame.)
This is so a bunch of servos can aim / point a camera at the beacon.
I believe the Ops intent is to keep the player in the cameras frame at all times. At play stop, the official play clock is stopped and this signal can force resync and return of the 0, 0 coordinate for the player at the (dustant) camera location. The helmet-beacon has no concept if 0, 0 except when a signal is sent to it. Manual intervention may be required of the remote camera to ensure that the player is centered before play; that is, 0, 0 beacon IS NOT 0, 0 remote.
I have not run a model, but the ground movements should appear 2-dimensional, from a distance... The aerial component, is definitely 3-dimensional but one can believe that more computing resources could be placed there to keep the camera pointing correctly toward the playing field (than in a helmet.)
The other scenario, camera on the ground tracking a RC helicopter involves taking into account the 3-dimensional movement if the flying platform... This would be critically important with optical zoom. Resync may not be possible unless the craft hoovered often (or manual intervention.)
However, ground tracking of an aerial object could rely on multiple technologies to increase the chance of success. I would think that spewing metallic coated mylar chad would be frowned upon by most national authorities. XD
Ray
My trusty HP67 still functions (sans card reader) from 1976 (received in Jan. '77 as a bDay present) still works great. If it doesn't compute on an HP, it does not need to be computed.
Although in the "just for fun" category, this will work with Arduino:
It would be hard to do with an Arduino, but the simplest method for keeping a target in a camera frame is to use optical recognition to recognize the object you want to track. For instance, if you used a computer vision algorithm to seek out a football it would theoretically always follow the person with the ball. Tracking people would be a bit harder, but there are plenty of folks out there doing it. That said, a Pi or other device with more computing power would be in order.
For my hypothetical stage light system I was going to use the IR beacons to acquire a target in the dark, but once they were in the light I am guessing it would be much easier to identify and track them with a visible light camera.
As for a inertial solution, it would be very hard because you would need to keep track of the two bodies (copter and target) for a good deal of time. Commercial sensors are OK, but nothing that would be considered small or light can yet give you the kind of accuracy you would need for more than a handful of minutes. Maybe you could live with this, but it would take a lot of hardware to even get it to that level.