Need to figure out what the best form of Infrared Receiver would be and what the best project setup for me would be.
Concept - Use IR Beacon tracking to direct a drone to a certain area - The "Landing Zone". Need a range of around 3m
Material Layout - Have an array of 5 IR sensors. 4 pointing in each direction - slightly angled downward, and one pointing downward; Have 1 IR Beacon sending out signal to be detected by the sensors. The drone will be controlled by emitting either a 38KHz or 57KHz signal.
I had seen this IR receiver module and I liked it because of the distance, however I believe IR receiver modules are strictly digital output (either HIGH or LOW). If I were to use these instead, I guess I would have to use a sensor array of 9 sensors - 8 pointing in various directions.
For my beacon I would like to put a Fresnel Lens on it in order to increase the beacon spread. Although it is a bit pricey - are there any alternatives? Do you think it would be better to just have multiple IR Led's pointing in various directions for the beacon?
Considerations -
Would prefer to use IR Phototransistor's because of their analog output - this can better help me determine which direction the drone is facing
I had seen this phototransistor post where reply #7 suggested using a Goertzel Filter to extend the range of an IR phototransistor and measuring relative amplitudes, but not really too sure how I could apply this for my needs
Do I put the sensor array on the drone or the "Landing Zone"?
My thoughts, put the sensor array on the LZ this way I don't need to have a processor on board the drone. However, I do think I will be including an IMU in this setup to more accurately track the drone.
What do you guys think would be the best beacon-receiver setup?
I'd think that you need a pulsed IR sender signal, so that you can distinguish it from background (sunlight...) radiation. Then a simple RC highpass filter may be sufficient (not sure) to separate the pulses from a somewhat constant background signal.
For the direction I'd try to restrict the sensors in a way, that only one sensor will "see" the beacon at a time. When two sensors see the beacon at the same time, turn the drone until only one signal remains, and follow that direction. Another vertical sensor will be required to guide the touch down, when no other sensor can see the beacon any more.
Placing the sensor in the LZ means a high transmitter power, because the transmitter doesn't know of a direction where to send a more focused signal. In either case the transmitter power must be sufficient for bridging the maximum distance between the drone and the LZ; I'm not sure which devices can provide a sufficiently strong signal at all (depending on the maximum distance to bridge). I'd think that a practical solution will require multiple transmitters, that send their signals into different directions. Then you can scale the power consumption by using as many transmitters as required.
We did up a set of prototypes with 8 receivers and 10 transmitters(beacon) for a similar idea.
We only got to test indoors and it worked quite well, as good as a RADAR
(Unfortunately, no drones available, to play with)
Below is a photo of an early prototype (The hoods are important, we have since got some 3D printed)
The beacon is similar with 10x IR leds(8 poles+ up&Down). Putting heatshink tubing over the emitters to narrow the angle helps a lot, but if you are just looking for a single homing point to fine tune after GPS then a single emitter might do.
You need to use a modulated/pulsed carrier at 38kHz. The reason for 38 (vs 56) is that the only readily available receiver that will work is the TSSP4038 (family).
If you are mainly interested in distance vs direction, then proximity sensors are the way to go, but may not be great outdoors.
However, the bigest challenge is the Sun, which reflects IR off leaves/grass blowing in the wind. It can be mitigated with good coding particularly at relatively close ranges like 3m.
No need for a fresnel lens, just drive the emiiter at higher current (good example on our blog). There are also very high power LEDs available. I would use TSAL6100.
regarding location of sensors...it is possible to design it on the drone or on the LZ. However, based on what you have posted I would lean strongly towards the drone for the receivers.
Likely to have more interference if the sensors are looking up (towards the sun).
You only need to activate the sensor circuit during landing and it could easily be handled by an Attiny with sufficient IO pins. (~20mA when active & almost no current when disabled)
FYI the weight of the populated board as shown in the attached photo is 29g (~1oz), excluding external wires/power. You could reduce this further by using thinner PCB, lighter hoods & SMD components for the receivers.
using TH coponents for IR, means you can fine tune directions with a bit of bending (vs SMD, which are fixed position & much lighter)
In case you are in any doubt, it is not an easy build/project. But I guess you know that.
