8.4. USRP Implementation

• The ML non-coherent demodulators for both \(M\)-ASK and \(M\)-DPSK use demodulation statistics that are obtained by sampling the MF (matched to the TX pulse shape) output at the symbol rate.

• Assume that fine signal timing is obtained by a joint acquisition and symbol synchronization step at the beginning of the packet. Let the interpolation and decimation factors of \(U\) and \(D\) respectively are needed to convert from the USRP sampling rate to the symbol rate.

• One simple way to generate the sequence of MF output statistics is to perform MF with an interpolation factor \(U\) on the RX samples from the USRP and then decimate the MF output by a factor of \(D\) with the correct timing obtained from the synchronization step.

• A more efficient way is to perform by directly employing a \(\frac{U}{D}\)-rate polyphase overlap-save implementation. Note that the decimation step takes place at time instants \(nD\) in the interpolated domain. Therefore, additional delay may need to be added to the MF impulse response to make sure that the decimation timing aligns with the fine symbol timing obtained from the synchronization step.

8.5. References and further reading

1

J. G. Proakis and M. Salehi. Digital Communications. McGraw-Hill, 5th edition, 2008. ISBN 978-0-07-295716-7.