Hello Navio community! I’m working on a Navio+ powered quadcopter that will be autonomously roaming the halls of a building and tracking its location via image processing (room numbers above doors) and possibly also using ultrasonic range finders. Safety is my number 1 concern here, followed by stability for better image processing, and I am looking for any tips or critiques on my current build before I order all my parts.
This will be my first time building a quad from scratch so I’m not entirely sure I’ve picked the right parts since I don’t have a good understanding of it all yet. Things I’m most unsure about are if I picked the right ESCs. Motors are only 12A, and ESCs are 30A, so should I look for a 20A ESC? If so, what do you recommend?
Transmitter/Radio combo seems to use SBUS which I believe is not compatible with the Navio, so I threw a SBUS to CPPM decoder in there. Is this correct and will I need any other adapters? I don’t actually plan to use the Transmitter much, but I do want it for calibration purposes and to do some initial testing.
I’m also open to any other questions or critiques you have about my build. I’m trying to clear up my own confusion here too, since this is a lot to learn about at once, so any pointers you guys may have are highly appreciated! Thanks for reading.
My intentions for the sonar are to use it for fine-tuning positioning, such as getting closer to a wall for a clearer image of the room number, or to determine how far away a wall is. I intend for most of the automation to be done through DSP , so when the camera recognizes a wall, I will have an estimate of how close the wall is. The building itself is very simple. It’s like a square enclosing a center square of rooms, where the hall essentially runs along the inner perimeter of the building. There are no pillars or strange curves, or anything that obstructs the walkway. Just 4 straight lines making a square.
In your research on Sonar vs Laser, did you find any other suggested alternatives for obstacle avoidance?
For Sonar, I experimented with having 6-8 sonars around the quad. You will need to poll each one sequentially to avoid multi-path interference. Even then, I did not find the results particularly useful in a real-life fully autonomous environment. Anything that is not a flat surface will give inaccurate results due to reflections, or the quad missing the object. If you are dealing with just walls with nothing else around (ideal condition, or in a lab), then you will be ok (although you need to watch out for the corners of the walls, where the sonar will bounce around). Once you have other objects around (chairs, tables, people), then you will find it more challenging.
I found laser giving much better readings, but of course the trade off is the higher price (eg Lidar Lite).
You can also look at a true Lidar solution (cheapest I found was > $6000-$7000). Even then, they are really built for mapping applications, it will be pretty easy to trip up the drone if you are wanting to do autonomous obstacle avoidance.
Yikes. Those lasers really do cost a pretty penny. There are a few benches lying around where I am testing (though not many), but the environment is decently to an ideal setting because of the simplicity of the building. I was thinking about testing the dependability of the sonar by taking samples of 10 and throwing out outliers that just aren’t physically possible with current speeds, but I realize that I’m not going to be terribly accurate, either way. Thanks for your input.
As an alternative to using SBUS-to-CPPM converter you can use an ACCST-compatible receiver with PPM output like FrSky D4R-II, the build would be cleaner with one less electronic unit.
Thanks @mikhail.avkhimenia. Thanks for the tip. So would you recommend something like a modded Turnigy 9XR PRO Radio Transmitter + FrSky D4R-II Combo rather than the SBUS-to-CPPM converter? I figured that going with the converter might save some time, because I’m not sure how difficult it would be to mod the Turnigy 9XR PRO.
@dgrat, By 4 in 1 ESC, do you mean something like this? If so, do you ever have any problems with a single ESC burning out, requiring you to replace the entire component?
As for the motors, my frame only supports 2208 1200KV or 2212 1000KV brushless motors. I didn’t find many motors with these specs, but maybe I was just looking in the wrong place.
The frame specs list support for 3S LiPos, but this is something I will have to test out on the field to determine if I will need something smaller. The smallest discharge on a 4000mAh battery would be 15C for my total draw of about 50A, but that discharge is hard to find, so I picked a 20C battery. And thanks for the tips on the propellers - I’ll look for something a bit sturdier. I appreciate all the input
I meant that if you’re going to use Taranis then you could just swap the receiver for D4R-II.
In case of 9XR you would need to swap both transmitter module (for FrSky DJT 2.4Ghz for example) and the receiver. Installing\Replacing the transmitter module should be easy, no need to hack anything.
Also there’s nothing wrong with using the converter:)
FrSky has a new firmware for the X receivers, that enables CPPM output. So you can use a modern X4R-SB receiver without a SBUS to CPPM cable.
Even cooler, you can flash the receiver with the Taranis. No extra hardware needed.
I will look up the exact procedure and post it here.
The firmware is for X4R and X4RSB receivers.
With X4R you get 8ch ppm on output 1 and pwm for ch9, 10 and 11 on output 2, 3 and 4
With X4RSB you get the same, except output 4 is still SBUS so only two pwm channel.
Awesome! That’s so great to know.This is the firmware on the Frsky site for the X8R receiver. Is this the version you are referring to? I didn’t see any detailed notes for this firmware other than bug fixed for SBUS
Thanks for reading.