Now that the LidarLite is off the market (until Garmin brings it back), there are not many options for accurate altitude sensing.
I have a optical flow sensor that I would like to mount up, but I don't want to spend $250 for the Lightware SF11, or $230+ for a TeraRanger.
So, my question to any developers: will ArduPilot be adding firmware support for this sensor?
The specs and the price seem very promising:
http://leddartech.com/en/leddar-sensor-products/leddarone
http://www.robotshop.com/en/leddartech-leddar-one-sensing-module-rs-485.html
Replies
Actually, I was limited in time to finish my previous work. But it's great if the problem is solved and even if somebody used any part of my work. Thank you for the support!
Randy said:
In an interesting turn, a pull-request has come in for the LeddarOne. I think it's a new effort, separate from what Anton has done above. Not sure if I'm being cheeky or not but maybe Anton could look over this PR? It's perhaps a bit of a shame to have two separate developments of the driver but on the other hand, two is better than none.
Hi Anton,
Good to lesson back from you. I'm trying to make x8 multi copter(8 coaxial propellers) size of my drone is very big each propellers TM-CF 34x11 inches and T-Motor U13 the total self pay lode including Bt is around 22Kg using Pixhawk i flying it. i did some stabilization for it through PID. i plan to add 18+ Kg of payload with it so interested in going with coaxial propellers.
Once i was testing my quad with your leddar one integrated frame were i got good stability in altitude. My plan is to use low altitude flying with high payload. thus add 4 more moter(U13) in coaxial conf and converted my quad to x8 multi copter. i can lift up to 25 KG of payload now, but still lag in stability.
I need your help in developing the .px4 for my X8. If you could share me some documents, how you integrated .cpp & .h of leddar one to .px4 quad, i can follow the same for developing .px4 frame-ware for my X8. I've also brought leddar sensor evolution kit for obstacle avoidance which need to be integrated with pixhawk.
I request your assistance in compiling custom sensor integration frame-ware with PX4 environment.
Thank You, Happy week end.....
Anton Horpynich said:
image1.JPG
Anton,
Also just to let you know, you can find the daily chatter of the ardupilot developer here: https://gitter.im/ArduPilot/ardupilot
If you're hacking code, it's a good place to get quick feedback from the other devs.
Thank you for the advice! Actually I'm not familiar with subversion system, but I think there were some changes after I made mine. So I'll use the newest sources and then pull request when I have more time for it.
Randy said:
Anton,
Maybe you could raise a pull request to get it into master? I've had a quick look and it generally looks ok. The large array for the CRC stuff eats a chunk of memory so it might be nice to make a function to do that instead if possible. Anyway, if you do a PR there will be feedback from a bunch of people, undoubtedly some changes will be requested but it's the best way in any case, thanks for this!
Hello, Sivashankar,
Do you want to build a firmware for Arduplane? Did you mean skywalker X8? Firmwares for copters and planes are completely different and I haven't tried to build arduplane yet. First I'm going to submit the changes for arducopter as it already worked for me. But I want to make extensive testing before. Could you help me with the more detailed information about your tests?
Sivashankar said:
Hi I've worked with the above .px4 firmware for quad, which was successful, I wish to work with leddar one with my x8 and tried downloading .cpp and .h file and setups px4 toll chain for building the px4 file for my x8 but i'm failing as i'm new to that. If possible can you share me the procedure for doing the same, are a .px4 file for pixhawk x8. I hope if i learned the way to do this will help me future work so any doc r video relater to this quire, will be happy.
Anton Horpynich said:
Hi Anto,
This is a great accomplishment and such excellent results from the Leddartech system are surprisingly good.
However, there is, in my opinion a basic problem for this use with Leddartechs implementation especially for outdoor use.
They use multiple LEDs as a pusled IR illumination source, in my opinion for outdoor use this will under many circumstances not yield a reliable return value.
An actual focused laser concentrates much more energy in a much smaller spot and the likelihood that you will get a satisfactory return is greatly improved.
Many outdoor surfaces are very IR absorbent and there are also surfaces that may be illuminated by direct sunlight.
In either case, the IR illumination provided by the IR LED array may not be enough to permit a distinguishable return.
Generally the much more tightly focused laser will punch through.
The divergence quoted by Leddartech is 3 degrees, for a lidar altimeter that is a lot and it also represents a very rapid reduction in relative brightness.
Laser divergence will typically be less than 1 degree.
I think the Leddartech sensors might possibly work OK indoors for limited object avoidance, but a genuine TOF laser based solution like those offered by Lightware is going to be much more reliable as an outdoors altitude detection system.
This is not at all related to which company is better, just that in this case, a tightly focused laser is a lot better than broader area LED illumination.
I realize that your initial outdoor results have been excellent, but I strongly believe that you will find that under other ground cover conditions and other direct sunlight conditions you may well find results degrade considerably.
The Kinect which also uses IR LED illumination becomes nearly useless outdoors in direct sunlight and although it is a much wider illumination area the problem is much the same.
For outdoor aerial altitude determination a laser based system is going to be generally more accurate and, I believe, much more reliable over a greater range.
Best Regards,
Gary
Finally I had a chance to test the sensor outdoors using modified Arducopter-3.4rc-1. And it works great! No problem with sunlight, the altitude data is quite stable even with the grass of up to 0.5 m in height. During the test the heights of up to ~12 meters had been measured without problem. Different modes (althold, loiter, auto) also were tested with success. See the graph of auto missions showing terrain following with the different speeds (5 and 10 m/s) and target height of 10 meters.
The height difference between the upper and lower poins is about 5 meters as shown in absolute height graph.
Also see the 3D view of the flight path:
-
1
-
2
-
3
of 3 Next