gr-ccsds

this is part a GNU Radio module for processing data which is encoded with the CCSDS Recommended Standard for TM Synchronization and Channel Coding using Reed Solomon.

it was done as part of my master thesis at NTNU in the spring of 2016.

install instructions

to use the blocks, you need to install GNU Radio (I recommend using pybombs)

then clone this repo and follow these instructions

mkdir build/
cd build/
cmake .. -DCMAKE_INSTALL_PREFIX=/path/to/your/prefix
make
make install

GNURadio Conference Talk

I presented my additions to this OOT module. My contribution was adding convolutional coding to create a nearly CCSDS-compliant full-duplex modem. I cheat a little by using my own toy encapsulation and link layer protocols.

See the presentation here: https://drive.google.com/open?id=0B-CV_07uSBIuVXBBdzRZYWw1cG8

And the video here: https://www.youtube.com/watch?v=gVCCQ6rX3nk

IP Encapsulation Instructions

You need to run a flowgraph on two separate systems, each with their own USRP, with TX/RX and RX2 connectors cross-connected, with approriate attenuation (at least 40 dB). The flowgraphs set up a TUN device, so you have to run them as root (use sudo).

On system 1:

./concatenated_qpsk_modem_txrx.py --mtu 1115 --rx-freq=400M --tx-freq=415M --gain 45 --tx-user-data-rate=500000 --rx-user-data-rate=500000

On system 2:

./concatenated_qpsk_modem_txrx.py --mtu 1115 --rx-freq=415 --rx-freq=400M --gain 45 --rx-user-data-rate=500000 --tx-user-data-rate=500000

Now you need to setup the TUN devices' properties. This sets up a point-to-point network between the systems. You can adjust the IP addresses to avoid conflicts with any network you are using.

On system 1:

sudo ifconfig tun1 10.11.10.1 pointopoint 10.11.10.2

On system 2:

sudo ifconfig tun1 10.11.10.2 pointopoint 10.11.10.1

Now you can network between the systems, over the RF link. For example:

From system 1:

ping -A 10.11.10.2

ssh -vvv user@10.11.10.2

scp -rv file user@10.11.10.2

Note that file transfer may stall. This may be due to the TCP/IP stack congestion window being reduced when it mistakenly identifies a packet loss or long RTT time as congestion on a normal network. We should be able to fix this with a PEP. Another possibility is that we're sending a packet that exceeds the MTU of the TUN, but I don't think this should be possible.

I usually use the following custom settings for OTA SSH (just add them to your ssh config file for the link you want to use:

Host usrp
Hostname 10.11.10.2
User hawk
Port 22
StrictHostKeyChecking no
#PreferredAuthentications password
#PubkeyAuthentication no
#PasswordAuthentication yes
ChallengeResponseAuthentication no
CheckHostIP no
ConnectTimeout 300
ConnectionAttempts 10
ServerAliveInterval 15
ServerAliveCountMax 4
TCPKeepAlive no
Cipher blowfish-cbc

More fun:

rsync -rzhPit --bwlimit=45k --stats --sockopts tcp_frto=1,tcp_frto_response=3,tcp_slow_start_after_idle=0,tcp_keepalive_probes=100,tcp_keepalive_intvl=3,tcp_low_latency=1,TCP_CONGESTION=vegas,TCP_MAXSEG=1113 Downloads/AIS.SampleData.zip usrp:~/Downloads/

It's often a good idea to keep some backround pinging going. Try not to overwhelm the link though, unless you're using some sort of QOS

ping -i 1 10.11.10.2

Network Test (no hardware)

./concatenated_qpsk_modem_txrx_net.py --mtu 1115 --source-ip 192.168.1.231 --dest-ip 192.168.1.223 --source-port 52001 --dest-port 52002 --rx-user-data-rate 1000000 --tx-user-data-rate 1000000

./concatenated_qpsk_modem_txrx_net.py --mtu 1115 --source-ip 192.168.1.223 --dest-ip 192.168.1.231 --dest-port 52001 --source-port 52002 --rx-user-data-rate 1000000 --tx-user-data-rate 1000000