Automated Precision Landing on Car, and at Night

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

@Benjamin

I am not confident that it will work with the APM hardware. It may be possible, but I think that another developer had some 'memory issues' when they tried integrating with APM. All of our testing is with Pixhawk.

Yes, the goal is to have standard integration (plug-n-play). We almost had the IR-LOCK code added to the next development cycle (ArduCopter 3.3), but I think it will take more time... We are also providing code to support advanced/customized development.

With the new Beacon, we have had good performance in sunlight. We had to increase the IR power, and the modulated signal prevents almost all false detections.... To give an example, if a 100Watt lightbulb was placed directly in front of the sensor, it may 'detect' it for 1-2 frames, then the algorithm will 'reject' it. We will continue to make improvements.

At night, the detection range is great (maybe ~75ft or more). In bright sunlight, we like to operate at 30-60ft. Of course, the range can be easily increased by adding an extra Beacon.

Your question about the GPS precision and field of view is a great one. We have already performed fully-automated missions, and the initial tests were successful. However, we need to fly at a certain altitude in order to ensure the Beacon will enter the field of view (due to GPS error). I don't have the exact numbers with me (and the specs might change), but our worst-case field of view at 45ft altitude is around 30ft (i.e., we can see a Beacon that is 15ft ahead or 15ft behind the quad). We can make some tweaks, but those numbers are comparable to GPS accuracy, so we will be ok.

You need to be a member of diydrones to add comments!

Join diydrones

Comments

Thomas Stone April 8, 2015 at 6:36pm

@Hasufel

Thanks! Great ideas and questions.

(1) Currently, we can handle multiple beacons, but NOT multiple distinct modulation IDs. It is difficult to explain why this is challenging. Basically, we need to sample the location of the Beacon at >5Hz. This is important for quadcopter controls. The sensor can only process Beacon locations at 50Hz (via machine vision). Somebody who is very clever may be able program a few Beacon IDs, but I think it would be MUCH easier just to turn the Beacons on/off as needed.

(2) That is a very cool idea (and I have heard it discussed in other settings :) ). But once again, you are limited by the machine vision processing time and camera sensor framerate. More expensive sensor hardware would probably be required. I have thought about it, but it gets expensive if you want to use non-standard high-framerate camera sensors.

(3) The primary limitation on range is the size and power of the IR target. When you get far away, the Beacon becomes smaller than one pixel.... I guess you could have infinite range if you put enough Beacons beside each other. Ha. :)

Best,

Thomas
Hasufel April 8, 2015 at 5:59pm

Hi Thomas, awesome work and greatly illustrated website: congrats. I have a few questions:

- will you implement modulation signatures? Say I have multiple beacons spaced by 1.5 meter with multiple UAS wanting to land. Could we be able to 1) assign a pattern modulation "id" to a beacon (bluetooth,usb?), and 2) Provide the UAS with the pattern modulation id signature of a beacon, through mission planner as a property of a waypoint.

- would it be possible to have programmable modulation signatures (the end game, with a simple API)? Beaming the id of the beacon and providing instructions would be quite valuable; like as a possible scenario: "this is beacon id:1, alt. desired: 5meters, at speed: 10cm/s, when at alt. desired: go forward with ir-lock sensor 90° up, listen to sequence beacon id#2". I'm thinking chaining beacons and add that 'smart' functionnality of a mesh of beacons.

- when you say "add more beacons to increase range" how would that work ? What max range could we have?

thanks!
Thomas Stone April 8, 2015 at 3:04pm

@Jiro

We are working on it. There are some issues that need to be sorted out, especially related to range, excessive power consumption, and other reflectors that may be present in the landing area.
Jiro Hattori April 8, 2015 at 2:45pm

I came across some thoughts.
As long as, the working distance is limited on IR-Lock.
Do we need active beacon only for this purpose ?
The retro-reflective material and on board IR projection will do the same thing.
Thomas Stone April 8, 2015 at 2:38pm

@Alex

With a simple setup, probably not. The moving car would probably escape the field of view of the sensor. But I can think of some solutions. :)
Thomas Stone April 8, 2015 at 2:35pm

@Benjamin

That is our goal (to target the next dev cycle), but I can't nail down any dates.

This industry is obviously going through a lot of rapid changes, so we will just 'roll with the punches'. At this rate of development, 9 months seems like an eternity. Who knows what things will look like at the end of the year. :)

Also, the official integration with ArduCopter does not necessarily depend on the 'dev kit'. It could move slower or faster, depending on some other factors... The 'dev kit' is mostly targeted at companies/researchers with customized UAS projects.
Benjamin Pelletier April 8, 2015 at 2:12pm

@Thomas, so you intend to target the development cycle after AC3.3 for inclusion of IR-LOCK code? When do you expect that release will be, and when do you expect I will be able to take delivery of a non-devkit IR-LOCK system for my Pixhawk? Sounds like maybe Q4 2015?
Alex Lee April 8, 2015 at 2:04pm

Can the car be moving?
Thomas Stone April 8, 2015 at 1:23pm

Thanks earthpatrol !

I liked this video.

https://www.youtube.com/watch?v=BNJzwu1keJI
Thomas Stone April 8, 2015 at 1:21pm

@Benjamin

I am not confident that it will work with the APM hardware. It may be possible, but I think that another developer had some 'memory issues' when they tried integrating with APM. All of our testing is with Pixhawk.

Yes, the goal is to have standard integration (plug-n-play). We almost had the IR-LOCK code added to the next development cycle (ArduCopter 3.3), but I think it will take more time... We are also providing code to support advanced/customized development.

With the new Beacon, we have had good performance in sunlight. We had to increase the IR power, and the modulated signal prevents almost all false detections.... To give an example, if a 100Watt lightbulb was placed directly in front of the sensor, it may 'detect' it for 1-2 frames, then the algorithm will 'reject' it. We will continue to make improvements.

At night, the detection range is great (maybe ~75ft or more). In bright sunlight, we like to operate at 30-60ft. Of course, the range can be easily increased by adding an extra Beacon.

Your question about the GPS precision and field of view is a great one. We have already performed fully-automated missions, and the initial tests were successful. However, we need to fly at a certain altitude in order to ensure the Beacon will enter the field of view (due to GPS error). I don't have the exact numbers with me (and the specs might change), but our worst-case field of view at 45ft altitude is around 30ft (i.e., we can see a Beacon that is 15ft ahead or 15ft behind the quad). We can make some tweaks, but those numbers are comparable to GPS accuracy, so we will be ok.

of 3

This reply was deleted.