Author Topic: Suitability for DIY model airplane remote control (RC) system?  (Read 889 times)

jholster

  • Newbie
  • *
  • Posts: 1
Suitability for DIY model airplane remote control (RC) system?
« on: December 13, 2014, 08:32:29 AM »
I'm thinking of building an RC system based on Moteino. Range is not a problem as tested by other people, but what about other limitations?

The RC system is aimed for aircraft controlling with the following requirements:

- The system consists of two Moteino-based units: handheld ground station unit ("transmitter" in traditional RC hobby slang) and on-board unit ("receiver", although they both are of course transceivers). The ground unit sends control signal (joystick input) and receives telemetry data from on-board unit (speed, altitude, gps coordinates, etc).

- Typical RC servo control frequency is at least 50Hz, so the total delay from ground station control input to servo output signal must be less than 20ms, including radio transmission itself and various CPU processing in both ends (sampling potentiometers etc). Is that feasible?

- The transmission timing in other direction (telemetry data from air to ground) is less critical, e.g. 25Hz or even 10Hz should be okay.

- The data packets in both directions are not big and can be fixed length, e.g. 10 bytes. CRC is mandatory, as misinterpreted control data could easily crash the plane.

Thoughts & questions about practical implementation:

- The system would use 868 Mhz band as I live in Europe (434Mhz is legal as well but allows less transmission power). Because no sub-channels are available (AFAIK), the system would be half-duplex, e.g. the same radio band must be time-shared between control data and telemetry data.

- The simplest protocol would be "send a packet, then listen for a packet for short time, repeat". ACK are not used because loss of individual packets is not critical (data is highly repetative), minimum lag is prioritized. Can you do all that in 20ms? Will there be timing/syncing issues leading to both ends transmitting same time? (Is carrier detection enough?)

Of course only real-world testing will tell if the system works good enough, but before that, can you think any obvious flaws in my plans?

Note: I'm aware that this kind of system cannot be as reliable as commercial RC systems using high-freq channel hopping etc. anti-interference techniques, and I wouldn't use this for controlling multi-kilogram high-speed craft, but only for slow-speed park-style flyer weighting no more than few hundred grams.
« Last Edit: December 13, 2014, 10:32:15 AM by jholster »

Felix

  • Administrator
  • Hero Member
  • *****
  • Posts: 5992
  • Country: us
    • LowPowerLab
Re: Suitability for DIY model airplane remote control (RC) system?
« Reply #1 on: December 16, 2014, 12:09:21 AM »
Well when you get around 30 packet roundtrips per second it gets crowded, this is with the default settings of 55kbps baud rate. You could increase the rate to buy some more time, but I'm wondering what are you doing with all that data. With that much traffic you will inadvertently run into some collisions and lost packets, will your RC unit crash or get derailed?

I think you're really pushing it at 50hz. I would stay below 20 ACKd packets per second if at all possible.
The library takes advantage of a CSMA technique - listen before talk provided by signal strength indicator (RSSI) and it will only transmit when the channel is "clear".