IR-LOCK Sensor for Precision Landing, etc

We finally have a progress report on the integration of the IR-LOCK (Infrared Pixy) vision sensor with Pixhawk. A group of students at the Georgia Institute of Technology is using the sensor for their 'Package Delivery Drone', which requires precision landings. They plan to use GPS to travel to a waypoint, and then use the IR-LOCK sensor to land precisely on an infrared beacon. Recently, they were able to get an IRIS+IR-LOCK to auto-hover over an IR beacon (see videos).

Other applications for IR-LOCK/Pixhawk are in the works (or at least in mind):

Search&Rescue (auto-search for emergency IR beacons)
Moving target following
Aerial surveying (I will be able to post a relevant video soon)
... your thoughts/ideas are welcome :)

We are in the process of making the IR-LOCK/Pixhawk interface more developer-friendly. The current developments (by the student group) are based on an IR-LOCK sensor 'driver' and codebase discussed on this FAQ page (link). A 'more official' implementation should be under development soon. The relevant discussion is here (link).

Small IR 'Pods' can be detected by the IR-LOCK sensor at about ~30-60ft. The IR-LOCK sensor reports the (xy) position of IR targets to Pixhawk at 50Hz.

@Dylan

Thanks for the enthusiasm! We are happy about the results, too.

1) Yes, we are actually approaching the problem with that same mindset. We intend to give a GPS waypoint, and then the 'IR-LOCK mode' will take over when the IR beacon is in range.

2) If you are using an M12 mount, then you could probably use the IR-LOCK Filter kit (link) on the camera that you linked to. Actually, the 'hole tabs' on the filter might need to be filed down to fit in between the components on that camera board... However, in the end, I don't think you will be able to achieve a good fps on the blob detection algorithm with Raspberry Pi (but I could be wrong). Pixy is optimized for blob detection (at 50Hz).

3) I agree that it is probably feasible. Our main limitation is simply labor/time. I wish we had more hours in the day. :) .... But I am curious. What do you think is the main advantage of putting the sensor on a gimbal? Extending the field of view?

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

Join diydrones

Comments

Thomas Stone March 1, 2015 at 12:29am

@Randy

Great! Your FF suggestion fixed virtually all of our problems.

We are very happy with the results. Actually, we are getting results faster than we can produce videos. :)
Dylan Eave February 28, 2015 at 11:50pm

Nobody is to distract Randy for any reason! :)
Developer

Randy February 28, 2015 at 11:44pm

@Thomas, that is looking really really good. Congrats!

P.S. as we type I'm working on some changes to get the irlock driver (written way back in Nov) into PX4Firmware branch (new pull request is here).

Driver for the IR Lock vision sensor (derivative of the Pixy vision sensor) by rmackay9 · Pull Requ…

This is a driver to allow communicating at 50hz with the IR-Lock sensor. It's perhaps slightly premature to submit this because it hasn't be…
Thomas Stone February 28, 2015 at 10:04pm

Here are the latest test videos (after freezing FF). The results are great. Landing accuracy is around ~1/2 foot. ... We will have more videos with full Precision RTL soon.

[Yaw command]

[Pilot commands throttle down]
Thomas Stone February 23, 2015 at 6:22pm

@Crispin

Thanks. We are not currently releasing hardware files (but software is all open and available). I see that you already made a purchase, so I can PM you a 'deal'.

Alternatively, you could use your own sourced LEDs. The performance-level depends on the 'wavelength match' between the IR-LOCK filter and LEDs. Most 940nm LEDs will work 'ok', but we sell the optimally matched LEDs on our site. Here is an example of some LEDs that were an 'ok' match with our IR-LOCK filter. (link)

Ultimately, we are trying to get distribution in Europe and other locations around the world.
Thomas Stone February 23, 2015 at 6:09pm

@Dylan

Sounds good. :)

The RPi has a lot of processing power, but I think the issue how the camera data is transferred around before it can be processed. We would have no problem supporting RPi cameras (w/ IR-LOCK filters/adapters), but I haven't seen any good blob detection framerates so far.
Dylan Eave February 23, 2015 at 12:24pm

@ Thomas... WRT point #2.... If the Pi could manage it would likely not be able to perform any other additonal functions well... Pixy it is :)

Point #3 could be donsidered point #2a... Ok... I did not think this one through very well -- I was envisaging a dual purpose camera. FPV and IR-Lock capable....

Scratch 2 and 3.... :)

Glad there will be failsafe to GPS.
Crispin February 23, 2015 at 11:43am

That is very cool!

Will put it onto my test rig (held together with plastic screws so crashes cost nothing).

Do you perhaps have an STL of the pod? I can print one myself and source the LEDs locally (UK). Shipping from US is a bit stinky :(
Thomas Stone February 23, 2015 at 11:34am

forgot this link: http://irlock.com/products/ir-lock-filter-for-pixy ... Notice that the 'tabs' are wide.
Thomas Stone February 23, 2015 at 11:32am

@Dylan

Thanks for the enthusiasm! We are happy about the results, too.

1) Yes, we are actually approaching the problem with that same mindset. We intend to give a GPS waypoint, and then the 'IR-LOCK mode' will take over when the IR beacon is in range.

2) If you are using an M12 mount, then you could probably use the IR-LOCK Filter kit (link) on the camera that you linked to. Actually, the 'hole tabs' on the filter might need to be filed down to fit in between the components on that camera board... However, in the end, I don't think you will be able to achieve a good fps on the blob detection algorithm with Raspberry Pi (but I could be wrong). Pixy is optimized for blob detection (at 50Hz).

3) I agree that it is probably feasible. Our main limitation is simply labor/time. I wish we had more hours in the day. :) .... But I am curious. What do you think is the main advantage of putting the sensor on a gimbal? Extending the field of view?

of 6

This reply was deleted.