Open-source drone developer kit

[itsmadmax420]

I am a part of the engineering team of a startup company called Zeius.

We are currently in development of a drone with gaming capabilities. As we've made progress on our internal developer tools we thought the drone community could benefit from these tools as well. We are developing a completely open-source super compact single board computer designed specifically for drones (think of a Raspberry Pi specifically made for drones).

In short, we're looking to reach out to the drone community to see if the tools we are working with would interest anyone else.

The key functionalities that we are proud to support are:

- Single circuit board for entire functionality

- Completely open-source software based on Debian Linux

- Fully compatible with ROS and other software platforms for custom flight controllers, computer vision, mapping, etc.

- Modular design for adding extra peripherals via UART, I2C, and USB

- All Under $100

I would love to hear some thoughts on this from the community! We've never really reached out like this before. Please feel free to ask any questions but also filling out this poll we just made will help immensely.

 

[wafflejock]

oooo I should have looked at this post sooner these are a few of my favorite things

So is the idea to replace the normal FC based on ARM processor and open source betaflight with your own flight control software and your single board computer will have IMU of some sort? Or is idea to still have FC and this is an auxiliary thing that lets you add more compute before sending control signals over to the existing FCs?

Also curious why not start with something betaflight compatible and add what you need (there is already a fair amount of peripheral support and it is also open source).

I've been doing a lot of aws iot work recently and shares the pub sub core architecture of ROS with MQTT messages between devices so I can see how that could be useful but not sure why start from scratch or raspberry pi vs using STM32F7 or similar MCUs being used on modern flight controllers and branch off betaflight which added some smart filtering to the gyro over the last few years.

 

[itsmadmax420]

Thank you for your interest!

For a little background we wanted to create a very simplistic (from a systems perspective) single board solution for drones or other remote control applications. The primary goal was that real-time image processing must be possible on the drone's hardware itself using reasonably inexpensive and accessible SoCs. We accomplish this via a quad-core Cortex-A7 ARM CPU running around 1.2GHz with 512 MB of RAM.

We are creating our own homebrew flight controller in Python. This is not the greatest solution for real-time or sensitive applications, such as flight controllers, but hear me out. We wanted to keep the entire system exposed and easily modifiable for our team but also future enthusiasts. Having our flight controller set up this way will make it easier for the community to make additions. If someone wants to change the PID control to LQR, or they want to add hovering stability for position control, it's all easy to edit the flight control script and execute it on the hardware of the entire system.

The rest of the system incorporates an IMU and various other communication interfaces for adding extra modules. We have an extreme focus on simplicity. In our current setup if the drone crashes and needs to be replaced, simply remove the SD card, put it in another board, mount the motors and board to a frame, plug in the motors and battery and you're done. No soldering or anything tedious required. A big focus for us has been to allow for 3D printed frames, allowing users to fully customize and get creative with their designs.

Once again thank you for your interest! The more feedback we get the more it'll be possible to provide a useful and open solution for the community.

 

[wafflejock]

Thanks for the detail and sounds like a pretty great project. Occasionally someone will pop through the forum here trying to figure out how to do this themselves from scratch or using arduino based flight controllers and have had a conversation with a teacher here looking for almost exactly what you described (simpler access to get to the point of testing functionality or variations on 3D printed designs). I'm still of mixed feelings on if this should be an add on to existing FCs though just because they do their job pretty well as is and want to keep the PID loop as "realtime" as possible so would be a bit worried about any non real time kernel or other process scheduling that could slow things down, it might be best to still keep that on a separate co-processor or something to guarantee the loop timing. Plus this way leaves the higher level computer to do higher level operations like image processing or route planning etc. I guess main thing I'd suggest for now is explore the betaflight, iNav and ardupilot/apm code bases and issues a bit too and see if worth standing on shoulders of any of those projects or if they are all just too complicated. The solder less setup would definitely be nice for schools but would want it to be an open standard connector so as to not be locked to a single vendor (amass mr-60 is a nice option and can alternatively just use any old bullet connectors if they disappear or prices get crazy)

 

[itsmadmax420]

Yeah, we've been going back and forth on the FC in a microcontroller instead of on the processor. RT Linux microkernel is a possibility to achieve real-time for the control loop but that's an open area of research for us. Luckily our control loops run slow enough that we don't see too much variation.

I appreciate all the advice. I'll have to admit, I hadn't heard about betaflight or iNav until you mentioned it. I'm going to look into those today. We'd like to be able to allow for existing platforms to be compatible as well so the more we know about the existing platforms the better.