(WORK IN PROGRESS)

Launching an experiment on the USRP board requires the following steps:

1. Reserve a number of USRP nodes (not implemented yet)
2. Create a Task as a directory (say taskFolder ), this is explained in the Create a task section.
3. Activate the task creation in Minus:
   

   username@srvairlock:~$ create_task taskFolder

This will result in the creation of a file taskFolder.job

1. Launch the task with minus:
   

   username@srvairlock:~$ submit taskFolder.job

2. Analyze the results

Creating a task for an experiment amounts in:

1. providing the programs targeted to each USRPs included in the experiment (explained in the Providing USRP programs section)
2. providing a scenario.desc file necessary for minus to launch the task (explained in the Providing scenario file section)

In CortexLab, USRP are programmed using the GnuRadio toolchain. GNU Radio is a free & open-source software development toolkit that provides signal processing blocks to implement software radios. In order to avoid arbitraty changes comming the GnuRadio developpement community, CorteXlab maintain a copy of the GnuRadio repository which contains exactly the same version of GnuRadio package that the one installed on the CorteXlab nodes. This distribution can be downloaded Here

Programming a USRP with GnuRadio is not an easy task, the user should be familiar with that, and should refer to the GnuRadio website if it is not. A CortexLab program consists of a number of GnuRadio python script executed on the PCs connected to the USRPs (of course many USRPs can share the same GnuRadio code). For instance, on the my_task example present on the CortexLab account, the following script should be executed by the PC connected to the USRP1:

$ benchmark_tx.py --antenna="TX/RX" --tx-amplitude=0.2 -v -W 2M -f 2.49G

The command could be anything, it is up to the user to decide the name of the script, its parameters etc. The only constraints is that it should be able to run with the GnuRadio packages present on CorteXlab. It the user wants to add more packages or libraries, he should put the files in the Task directory. For instance, on the my_task example, the benchmark_tx.py script uses the transmit_path class which is defined in the tansmit_path.py package, which is present in the my_task directory.

Again, using the my_task example, this example uses 2 USRPs (one transmiting, the other receiving), the command to be launched on the PC connected to USRP2 is:

$ benchmark_rx.py --antenna="TX/RX" --rx-gain=25 -v -W 2M -f 2.49G

Once the programs to be launched on each USRPs are set up, remains to indicate the time duration of the experiment. Minus will stop the experiment after this time duration, and the result will be available for analysis.

The scenario file is the Minus script that indicates to Minus which program will be launch on which USRPs and how long the experiment will last. For each node (i.e. each usrp), it should indicate:

1. a command (or entry) which is the program executed on the PC connected to the USRP
2. a commande line option (or parameter) which is the parameter given to the command
3. an exit script (optional) which is executed at the end of the experiment

In addition with these information, the scenario file should contain the description of the experiment and a duration in minute. The exact syntax of the scenario description file is presented here. Below is a simple example used for the my_task example:

# Example scenario description file
desc base scenario for CorteXlab
#
# Experiment maximum duration  integer (minutes)
durat 3
#
# Node list
#
node2:
   entry benchmark_rx.py
   params --antenna="TX/RX" --rx-gain=25 -v -W 2M -f 2.49G
   exit none
node1:
   entry benchmark_tx.py
   params --antenna="TX/RX" --tx-amplitude=0.2 -v -W 2M -f 2.49G
   exit none