Close
0%
0%

YAUVC - Yet Another Unmanned Vehicle Controller

A modular, distributed processing, unmanned vehicle
controller.

Similar projects worth following
Unmanned vehicles have common needs: energy to think and move around, sensors to interact with their environment, communication skills and servo control. It really doesn't matter if the vehicle is going to fly, float, dive or move over hard surfaces. In most cases all of the above requirements are present in one form or another.

This project is about a modular flight controller. It was made modular so I could work on one module without worrying about breaking down the rest of the system. I wanted to be able to upgrade a module with a new sensor, component or capability as much as I wanted the whole system to configure itself every time a new module was plugged. Finally, I wanted the system to be able to handle redundancy, just in case the user plugged more than one of the same module, like two pilots, or three sensor modules.

To guide the system development I followed some guidelines:

  • A generic micro controller module should be developed.
  • The connection between modules should follow some kind of standard, permitting them to communicate easily.
  • The addition of extra hardware to a module should be easy. For that to be true, the generic module should make most of the popular interfaces available to a 'sub module' where the new hardware should be installed. This submodule should be plugged to the back of the generic module and that should be the only requirement for it to work.
  • No module should have the same function as any other module on the system, unless redundancy was required.

I also wanted:

  • Generic modules to be cheap and easily replaceable. I had to select a micro controller to use in them. Out of my ignorance I chose ATMega328p.
  • Generic modules to be able to use different micro controllers, as long as they followed the same communication protocols and made available the same services to the sub modules, in the same pin order.
  • Any hardware installed in the system to be easily replaceable by newer, better versions in a way that the rest of the system would not notice the difference between the new and the old one. Backward compatibility.

This sounds much like an Arduino, but with some basic differences:

  • The ATMega328p micro controller has two SPI interfaces, but Arduino blocks one of them with the serial connection. I wanted both interfaces working: one to use as the main connection to other modules and the second one made available to any hardware installed in the submodule.
  • I wanted every byte of flash and RAM available to the firmware, so the Arduino bootloader was not an option and was removed.
  • I wanted full control of all the firmware code. Hidden code was also not an option. All the C++ code used in the system was (re)written and placed in open source classes.

To make all this work I had to develop and test some standards:

  • The physical generic module size should be standard, be as small as possible, and should include a connection to the main bus and a connection to the submodule.
  • The connection to the main bus should include power lines, communication lines and the possibility to reset a misbehaving module.
  • The connection between a module and a submodule should have as many interfaces as possible, following a standard pin order.
  • The submodule physical size should be as small as possible, but big enough to hold the necessary electronic components required for a vehicle control system, like sensors, GPSs etc.
  • The communication between generic modules should be done using a simple, popular interface. I had to decide between SPI and I2C. The second one seemed to be the obvious pick, but after reading horror stories about locked I2C buses, I decided to use the SPI interface.
  • As both interfaces transmit only a byte at a time, I needed a protocol to be able to transfer floats, longs or strings between modules. This protocol was written in C++ and is also a system standard.

The electronic standards were translated to components in EDA software. In the beginning this software was Eagle, but after hitting its limits I decided to migrate everything to KiCAD. It took me some time to reach the current project status, which is partially described below, module by module.

1 - MCU module: an ATMega328P in a little PCB, connected to the other modules by a main SPI bus. This module is surrounded by pads much like an Arduino stamp. These pads form a standard connection to a smaller module which I call a submodule. They deliver a second SPI, an I2C, a serial connection, two analog inputs and some GPIOs to the submodule. To enable the second SPI I had to remove the boot loader from the ATMega328P and strip all Arduino and wiring traces, rewriting what I needed from the beginning and redefining many things.

2 - AMGP sensor submodule: a small board that connects to the generic MCU module. This submodule has an accelerometer, a gyro, a magnetometer and a barometer. They...

Read more »

Adobe Portable Document Format - 1.06 MB - 03/05/2017 at 02:16

Preview Download

Adobe Portable Document Format - 492.45 kB - 02/19/2017 at 21:07

Preview Download

View all 2 files

  • FindChips lists!

    E. N. Heringa day ago 0 comments

    Hi! I'm happy to inform that, due to a nice partnership between Hackaday and FindChips, I can now share the BOMs (bill of materials) of two of the project modules. More lists are coming in the next logs, but the first two ones are here:

    MCU module BOM: https://www.findchips.com/u/list/28455-mcu-module

    PS module BOM: https://www.findchips.com/u/list/28469-ps-module

    Thanks, Hackaday and FindChips for this amazing new feature!

  • New classes. New submodules.

    E. N. Hering03/10/2017 at 06:08 0 comments

    Hi. I'm working on the flight control module code. I added a few new classes: Navigation, PID, FCM (Flight Control Module), SCM (servo control module), GPS (GPS module) and a very abstract GNC (guidance, navigation and control) class. My main objective is to make GNC the main class for vehicle stabilization. All other vehicle specific attitude control classes should derive from it. Some of the new classes have only the very basic structures needed to integrate them with the system, but FCM inherited code I have already tested against FlightGear in the past, and is able to take a simulated airplane from waypoint a to waypoint b without crashing it.

    I regenerated the documentation and uploaded it to labvant.com website. The absence of comments is proportional to the absence of my free time, but doxygen does a great work crossing, formatting and displaying code. Generating documentation with Doxygen was a suggestion from an IRC collaborator: NoHitWonder.

    About the new submodules, a brushless motor control module has already been drafted with the help of Erlend^SE, another project IRC channel collaborator. He is also inspiring and helping with the development of a new power supply module and a ESP8266 module to substitute the RN131G. You can join the project IRC channel by connecting your IRC client to a free node IRC server and joining #labvant.com.

    If you like how this project is evolving, please consider supporting it by helping to develop the electronics, the code or the firmware. All the project info is publicly available on the project repo, including schematics, PCB drawings and firmware code in C++. You can also develop your own modules using the project board/connection standards, or propose new standards for future systems. Supporting this project financially by sending a $1 PayPal donation is also welcome. That will help covering the PCB and equipment costs. Thanks a lot for your help and support!


  • Module recognition implemented.

    E. N. Hering03/04/2017 at 15:31 0 comments

    Today the module recognition code was successfully debugged. Now it is possible to plug a module on *any* slot of the backbone PCB and it will be recognized by the main communications module. After recognition the module is integrated into the system workflow automatically. If the new module is a sensor one, for example, it is read in sequence with other data source modules, making data available for decision making by the flight control module and further servo positioning.

    This is a major step in the system development, giving it a plug-n-play like capability.

    And here is a snapshot of the serial connection to the communications (COM) module. Serial commands are one character long. In the example below, 's' lists the slot occupation, showing the AMGP module (identified by number 1) on slot 4. The other commands return sensor data: a for accelerometer, g for gyroscope, b for barometer (altitude, pressure and temperature) and m for magnetometer. All values are multiplied by 10 or 100, depending on the case, to avoid using floating point numbers. The values shown are averages over a few samples, followed by standard deviation.

    The bad news is that the power supplies are failing. I had two buck-boost converters killed by an unknown (to me) factor. I hope to find the mysterious cause when my scope arrives.

    If you like how this project is evolving, please consider supporting it by helping to develop the electronics or the firmware. All the project info is publicly available on the project repo, including schematics, PCB drawings and firmware code in C++. You can also develop your own modules using the project board and connection standards, or propose new standards for future systems. Or you can help to support it financially by sending a $1 PayPal donation to help covering the PCB and equipment costs. Thanks a lot for your help and support!

  • Power supplies working!

    E. N. Hering02/19/2017 at 13:48 0 comments

    Both power supplies have been tested and seem to be working as expected. I have not tested them with LiPo batteries installed, because those batteries have not arrived yet.

    Below are pictures of the 3.3V and the 5V modules. The power supplies are, in theory, capable of sourcing 1.4A at 3.3V and 0.8A at 5V. They can be monitored and controlled by the MCU modules that are attached to them.

    All the project data, including code, references, datasheets, PCB schematics and design are available as open source and open hardware. You can download everything from the repository and use as you wish. You can even create modules for your own purposes.

    If you find this project interesting and believe you can help in some way, please write a message to me (enhering @ gmail . com). Your contribution is very welcome!

  • Power supplies assembled! With LiPo support!

    E. N. Hering02/13/2017 at 11:14 0 comments

    Hi! I've just finished assembling two power supply submodules, one to run at 3.3V and another at 5V. Each can hold around 1A of current and manage a LiPo cell, charging it automagically or drawing power from it, or from the input line, depending on which options are available at the moment. The 5V power supply will work mostly to feed the servos, while the 3.3V one will feed the logic. These submodules will be attached to a MCU module each. The MCU will monitor and control the power supply, sending and receiving information about available power to the system via the main SPI bus.

    I have also assembled three more MCU modules. Two to drive the power supplies and another to read and process GPS data:

    Soon I'll publish the test results of these modules.

    If you like how this project is evolving, please consider supporting it by helping to develop the electronics or the firmware. All the project info is publicly available on the project repo, including schematics, PCB drawings and firmware code in C++. You can also develop your own modules using the project board and connection standards, or propose new standards for future systems. You can also support it financially by sending a $1 PayPal donation to help covering the PCB and equipment costs. Thanks a lot for your help and support!

  • Sensor data via WiFi and serial connection!

    E. N. Hering02/09/2017 at 04:49 0 comments

    Hi! I have good news!

    After a long debugging session, today I could read the sensor data via UART and via Wifi. The data was produced in the AMGP sensor module and delivered via SPI to the COM communications module, where it was routed to my computer via UART and via WiFi. This is a major milestone in the system development. I took a screenshot of the serial communications and of the TELNET windows:

    (AA = Accelerometer averaged data, AS = Accelerometer std dev data, same for magnetometer (M), gyroscope (G) and barometer (P), which delivers pressure, temperature and altitude instead of a three dimensional vector like the other sensors)

    And here is the crazy setup to program and debug the modules:

    I'd like to thank the people from #avr IRC channel on freenode servers for all the help in debugging the code running on the modules, specially carabia, Lambda_aurigae, specing, CapnJ, rue_house and all the others I forgot to mention because my memory does not work well.

    I'll be back soon with more good news. Thanks a lot for your support!

  • 9 dof submodule assembled!

    E. N. Hering01/20/2017 at 20:09 0 comments

  • *HELLO* !

    E. N. Hering01/19/2017 at 03:27 2 comments

    Hi!

    Today the RN131G WiFi module had its firmware successfully upgraded to version 4.41 and I could connect to it via Telnet.

    This module is already mounted in his horseshoe PCB breakout and connected to the communications MCU module via the backbone PCB.

    This means that now I can connect to the MCU module via WiFi!

    My next step is to start to assemble the sensor submodule. Soon I'll publish the sensor readings via WiFi here.

    Thank you for your support!

  • First signs of backbone activity

    E. N. Hering01/17/2017 at 00:01 0 comments

  • MCU Module working!

    E. N. Hering01/14/2017 at 19:44 0 comments

    MCU module is alive!

View all 29 project logs

Enjoy this project?

Share

Discussions

Douglas Miller wrote 05/23/2016 at 00:40 point

'Fly Hard, with a Vengeance'? FHWAV. :)

  Are you sure? yes | no

E. N. Hering wrote 03/13/2017 at 23:32 point

Sorry for taking so long to reply, Mr Miller. I believe we need a better name. But I still have no idea on how to find it.

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates