Hi Michael, my intention is to allow a communication between two USRP's hooked up to two different computers. Each USRP has one RFX2400 daughterboard and one antenna connected to the Tx/Rx port. I have a more of a CS than ECE background and my focus here is on the application level. I would like the radios to be able to send and receive messages asynchronously - which can mean asynchronous communication at the application layer, but synchronous below - perhaps not even observable above. In my scenario, radios send messages which are read from some buffer and at the same time receive messages and put them in a different buffer for further processing. How do I go about implementing such scenario? Do I have to implement a protocol which consists of some time-based switching of rx/tx? Can I achieve it with only one antenna per USRP, or do I need two of them - one for Tx//Rx and one for Rx2 port?
I imagine that in the case of only two radios, I could achieve this behavior by allowing one radio to Rx while the other is Tx-ing and after short time, switching the direction to the opposite, to allow the first radio to Tx. However, I don't see how this could be achieved if 2+ radios are involved - without the prior knowledge of the number of "neighbors". That is why I thought each radio should be allowed to Tx and Rx at the same time. Thank you for help! Jakub. On Mon, Apr 13, 2009 at 4:19 PM, Michael Dickens <[email protected]> wrote: > Hi Jakub - 2 possibilities, depending on how one interprets your question: > > A) If what you mean is "using a single host computer that controls the two > USRPs such that one transmits while the other receives and vice-versa, but > never a single USRP both Tx and Rx simultaneously", then yes, that can be > done; you could probably even use two host computers that sync the Tx/Rx in > a MAC-like way using the current GNU Radio software, so long as each USRP is > not trying to do simultaneous Tx and Rx. If using a single host computer: > on each USRP, you'd use set_auto_tr(true) and then control both USRPs from > the same Python script. My colleague Glenn has done this in a "TDMA > fashion" to compare properties of actual and theoretical relays; see his MS > thesis and related paper at UND < > http://www.nd.edu/~jnl/group/glenn-bradford/ >. > > B) If what you mean is "simultaneous Tx/Rx on each USRP, using a single RFX > and single antenna", then this problem is (1) possibly not do-able in the > current state of GNU Radio software and USRP hardware; and (2) even if it is > do-able, certainly not easily so in GNU Radio software without significant > hacking to make it work. The primary issue is in jointly estimating the > channel parameters (both the Tx -> Rx delay and filter coefficients), which > might be time-varying or dependent on qualities beyond the direct TX signal. > I'm happy to discuss this interpretation further off-list. > > Hope this helps! - MLD > _______________________________________________ Discuss-gnuradio mailing list [email protected] http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
