Maybe we can land on a moving car in our next demo. :)
IR-LOCK is a developing precision landing system for multi-rotors, which is expected to enable high levels of automation in UAS. Imagine autonomously landing on a small charging station, or autonomously performing inspections at very specific points on infrastructure or facilities. In our prototype testing, landing accuracy of 5-30cm has been achieved.
Since the system uses an IR Beacon, it is not hindered by low lighting conditions or even complete darkness. And the beacon modulation is designed to accommodate extremely bright conditions. Overall, the beacon detection has proven to be very reliable, which is obviously important in this application (we don't want to autonomously land on your dog/cat).
A Precision Landing Development Kit will be made available soon. The controls code will be provided so you can easily integrate the system into your UAS (based on ArduCopter 3.2.1). Preliminary details are here:
http://precisionlanding.irlock.com/?page_id=4310
Please do not hesitate to ask questions. We definitely appreciate the support and feedback from the DIYDRONES community.
Best,
Thomas
Comments
@Thomas, thanks for the responses! I think $150 for the sensor+beacon pair is quite reasonable (though on the edge of what I personally would pay), so maybe I should just ignore the fact that I feel like $50 for the beacon is a little high.
I'm a little surprised a rolling shutter is a benefit, but it's probably not worth dissecting design -- if it works, it works and it looks like it works :)
If I have a standard APM setup, how would the actual integration go? Would I need to compile my own firmware using your control code, or would you provide pre-compiled APM hexes with your control code, or would you add your control code into the main APM branch similar to what PX4Flow has done/will do, or something else? If I can just slap on your sensor to my existing quad, upload new firmware, and flip a switch to have it precision autoland, that's probably worth $150 to me. If it's just a sensor with example code that I have to integrate myself, that's probably not worth $150 to me.
I can see how the system would work excellently at night. Does it work equally as well on a sunny day? What is the approximate/typical acquisition altitude on a sunny day?
Combining the acquisition altitude and the camera's effective field of view, to what horizontal precision must the GPS fly the quad in order to be sure the beacon will be acquired from autonomous flight on a sunny day, assuming perfect altitude control?
Thanks again!
Cool, cool, cool, and yes, land on a moving car, or train or semi. :) Great work.
@Benjamin
... you made me think of a critical point (and perhaps, you were already thinking in this direction). It would be very cool if the PX4Flow could perform the optical flow algorithm AND IR beacon tracking. However, the IR-LOCK system actually 'cuts out' the visible wavelengths of light. We bandpass the incoming light such that it is an optimal match with our selected LEDs. (and obviously, optical flow needs the visible light spectrum to operate)
Thanks Daniel! I know you have a lot of experience with this!
Nick. Yes, some of the R&D is built-in, and the remainder is a labor of love. :)
@Benjamin
Great questions!
You are correct that the price could come down after mass-production, but we are focused on getting this out quickly to our developers and partners (before process optimization). I understand that $50 is expensive for many hobbyist, but it is extremely cheap for some companies who need this solution now.
We have previously worked with basic IR LEDs (link). It was a good start, but the problem we encountered was interference from other IR emissions (for example, strong sunlight reflections, or powerful incandescent lighting). So the new Beacon has a modulated signal that is recognized by our sensor firmware. ... Ultimately, we are targeting a >99% system reliability, especially for fully-automated UAS. We envision a future of automated drone systems that may run for multiple days without human intervention.
I will not claim to be an expert on the PX4Flow, but there is a fundamental difference between the sensors. Global shutter is excellent (and necessary) for optical flow algorithms, where you want to eliminate distortion within a frame. Rolling shutter allows us to quickly process rows of pixels 'as they come'. Then we have more processing power available for Beacon detection, etc.
Thanks again for the response.
Best,
Thomas
You're paying for more than hardware, that price likely includes the R&D effort and programming that go into the project.
Ah, ok. Is it communicating with the sensor? If not, $50 still seems too expensive as any arbitrary pattern can be generated on four LEDs pretty easily.
Nice work Thomas!
Benjamin, I believe the latest iteration of the IR-beacon is actually using a microcontroller to actively modulate the LEDs to send a specific signal to the Pixy in order to reduce false-positive IR sources.
~$100 for the sensor seems pretty good and I'd probably be interested at that price. But $50 seems like too much for the beacon -- isn't it just four passive IR LEDs? Could I make my own beacon using 4 passive IR LEDs and still get the benefit of the sensor?
Also, is there a reason the PX4Flow wouldn't work for this purpose if a lens without an IR-cut filter was installed?
Very impressive! Good work.
-Daniel
-
1
-
2
-
3
of 3 Next