Does lteDLPerfe​ctChannelE​stimate command in LTE toolbox yields the true channel response or just an estimate?

Does lteDLPerfectChannelEstimate command in LTE toolbox yields the true channel response or just an estimate?

Answer 1

[Iain Stirling]

Hello Omer,

Yes, you've captured it well, lteDLPerfectChannelEstimate should serve as an upper bound. Doppler effects deemed "negligible" to the OFDM link performance will degrade the perfect channel estimate, but the degree of degradation should be similarly negligible.

Regards,

Iain.

[Omer Sholev]

Thank you very much

Answer 2

[Iain Stirling]

Hello Omer,

lteDLPerfectChannelEstimate operates by generating an impulse train (one impulse for each OFDM symbol), passing the impulse train through the channel and performing OFDM demodulation. Therefore it yields the true channel up to any limitations imposed by OFDM:

• if the channel varies (non-zero Doppler frequency) then orthogonality is lost and inter carrier interference (ICI) results, adding noise to the channel estimate
• if the channel delay spread exceeds the length of the cyclic prefix, then this also manifests as a varying channel, again introducing ICI

The degree of "imperfection" therefore depends on the exact operating conditions. Note that the waveform of interest being passed through the same channel will also suffer the same degree of ICI.

Regards,

Iain.

[Omer Sholev]

Hey Iain, thank you for the response. Is it possible to get the true (not estimated in any way) channel response that the command lteFadingChannel uses somehow? lteFadingChannel returns an info variable which contains inside it a variable called 'PathGains' but it doesn't seem like channel frequency response?

Answer 3

[Iain Stirling]

Hello Omer,

lteFadingChannel operates solely in the time domain, combining the path gains with time-domain filters that delay the inputs according to the path delays. Therefore the channel frequency response needs to derived by performing some kind of transformation from the time to frequency domain. This has theoretical limitations when the signal is varying over time.

Do you have a specific scenario where the channel estimate from lteDLPerfectChannelEstimate is not a good fit?

Regards,

Iain.

[Omer Sholev]

I'm trying to implement my own idea for channel estimation in LTE, and I just want to make sure that I compare my results with the true channel response. I do want to test my algorithm in a time varying scenarios so I do want model cases with Doppler affect.

But I thought that correct OFDM modulation requires the time of the transmit signal to be less than the coherence time so that the doppler affects are negligible. am I correct?

So basically the lteDLPerfectChannelEstimate would serve me as an upper bound?