You are not logged in.
Hi,
I have few questions related to channel estimation in simulink.
-Is it done on per subcarrier basis or flat fading has been assumed ? I see that "sub car Index" is output from block "FFT and CHan ESt". Hence, channel estimation seems to be done at symbol and sub carrier levels. The signal 'sub car Index' traces back to subblock "Addr Gen". Can anyone please explain the relation between this subcarrier index and Addr Generation ? I am a bit confused.
-Channel estimation is done using long sequence by multiplying the received waveform by long sequence (esp. for WLANs). However, i couldnot find this multiplication in the blcok 'channel estimation'. Also, I am not able to understand subblocks H_A-A H_A-B H_B-A H_B-B estimate buffers.
Any help would be greatly appreciated.
Thanks,
Vijayant Bhatnagar
Offline
The channel coefficient for each subcarrier is estimated independently (i.e. we assume flat fading per subcarrier, independent fading between subcarriers).
The Addr Generation block generates the memory addresses fed to the buffers which store the received I/Q samples. For every 80 samples received, only 64 are fed to the FFT (thereby removing the cyclic prefix). This block generates the addresses at the right time to remove the prefix and feed the FFT 64 valid samples.
We use dedicated channel training symbols, transmitted after each packet's preamble. The number of training symbols per antenna is configurable. The training symbols are defined in the frequency domain, and contain only values of +1, -1 and 0. Thus, no multiplier is required to estimate the channel- just a mux to select the raw I/Q sample or its negative. This mux is inside the 'Mult by training' subsystem in the Channel Estimation block.
The H_A-A, etc. buffers contain the estimates for the four distinct channels; H_A-A stores the 64 channel coefficients for the Tx_antennaA - Rx_antennaA path.
Offline