Hi guys!
Want to share what we have been working on lately, it is an autopilot shield for Raspberry PI. It is called Navio and has everything what you will usually find on an autopilot platform. A powerful platform like Raspberry gives many opportunities like streaming video, 3g,wi-fi, ethernet connectivity and possibility to run computation intensive algorithms like Kalman filtering or RTK GPS. RTK or real time kinematics is one of the main features of the board, it is equipped with a GPS capable of providing raw data output: carrier phase, pseudo-ranges and ephemeris. Processing this data against a stationary receiver increases GPS positioning accuracy to centimeters. A radio link between the two is required, but you anyway have it to GCS. If network RTK is available in your area, you can use corrections from the internet over 3G thus eliminating the need in second receiver.
Our plans include porting APM to Raspberry Pi + Navio.
Features:
MS5611 barometric pressure sensor
MPU9150 (MPU9250 on future models) 3-axis gyro, accelerometer and magnetometer.
ADS1115 16-bit ADC
PCA9685 PWM extender to control servos
u-blox NEO6T GPS module with raw data on SPI, we saved the only UART on raspberry for Xbee or different telemetry.
13 servo connectors
Pixhawk compatible UART, I2C and SPI connectors
RGB LED – just because we love them
What we wanted to know if there is any interest in platform like this, because we have everything ready for manufacturing and to keep the cost down we need to manufacture as many as possible. We have written tutorials how to use the board and are now preparing them for publishing, all code for the board will be released under open source license. Some more pictures and details are available on our website.
Comments
I assume the MPU9150 is on I2C? (I see no mention of SPI in the datasheet). That could be a real problem for the copter code, as the 400kHz limit of I2C and the contention on the bus from other sensors will make getting high rate gyro/accel values very difficult. I think it means you will have to use the on-sensor filtering only, and pull data out of the FIFO at the main loop rate. That is likely to be a significant limiting factor.
You could probably fly plane and drive a rover OK with that, as both are happy with a 50Hz main loop. The current APM:Copter code runs the main loop at 400Hz, which I don't think will be doable with I2C gyro.If you make the main loop run at 100Hz instead then I think you may be able to get it to work OK, although it will still take some very careful coding!
Good luck with the project!
Cheers, Tridge
@MD
Well, the hardware prototypes are ready, and all we need now is to get enough preorders to make a large batch.Probably trough Indiegogo campaign. The tutorials are on their way, but there are almost endless applications for the board, we will keep them coming at a good rate. If you want to get one of the prototypes, you can drop me a PM with a project you have in mind.
@MD
Thank you! We will keep you updated!
@James Cartwright
Thank you for you support! We are targeting not only autopilot applications, but also any projects requiring IMU, servo control and RTK GPS.
I am interested for sure. Would like to see some tutorials on this.
@Guy McCaldin
We understand your concern. Our vision is that there is no need for hybrid systems that were so popular lately - they highly overcomplicate the setup. Actually the simpler the system the less place for error exists. Computational power of Raspberry could make more complicated navigation algorithms possible. And also developing new features would be much easier, with multiple available libraries that are not available on MCU platforms, plenty of CPU and RAM.
We know several commercial autopilots for planes that run Linux with real-time kernel.We’ve been able to tune Raspberry Pi with Linux to run control loops at more than 50Hz without a problem. We will release a tutorial on this topic soon.
@Thomas J Coyle III
Thank you for you comment! We are familiar with this problem from flying platforms,that we usually build, but have not got any problems with this particular setup. In case there is some interference it is still possible to mount external compass.
@Kabir
Thank you!
@Rana
Thank you!
@Graham Dyer
-Considering the price of parts and devices with same characteristics this condition would be hard to met, but we are targeting something like 135-170 depending on the version + raspberry.
-They are not that bad and give compatibility with existing peripherals. We've got most important sensors onboard so you probably won't even need to use DF13. Servos, PPM and BEC connect trough standart 2.54 servo headers. We will also supply df13 crimped wires with each board for prototyping!
-We truly understand the needs of modelers. I personally do F3K for almost 5 years.
-We will be releasing many tutorials soon, stay tuned!
-This is the hardest part, but a lot of work is already done by ArduPlane community and this is our main goal.
-Depends on running APM. From the hardware point of view everything is ready.
Thank you for your reply!
Personally, I would seriously consider buying if the following conditions were met:
- Priced around US$100 or less,
- No or very very few DF13 connectors! (deal breaker!),
- Easy setup and connection, aimed at normal modellers,
- Decent documentation,
- Runs ArduPlane natively,
- Connects to Mission Planner natively.
-
3
-
4
-
5
-
6
-
7
of 7 Next