基於最小描述長度估計正交時頻空間調變系統的雜訊變異量

Noise Variance Estimation for Orthogonal Time Frequency Space Modulation Using Minimum Description Length

廖文鈞(Wun-Jyun Liao)

指導教授 ： 蘇柏青

國立臺灣大學/電機資訊學院/電信工程學研究所/碩士(2021年)

https://doi.org/10.6342/NTU202102035

摘要

正交分頻多工(OFDM)在時域—頻域(time-frquency domain)之中傳輸訊號，因此對於高都普勒頻率偏移十分敏感。而正交時頻空間調變(OTFS)則是在延遲域—都普勒域(delay-Doppler domain)之中傳輸訊號，可以把原本的衰減時變通道轉變為不受時間影響、且有固定複數通道增益的通道。因此正交時頻空間調變對於高都普勒頻率偏移的相容性更高，更適用於高速通訊的情境。Raviteja等人針對OTFS提出了以保護領航訊號(guard pilot)為基礎的通道估測方法，他們將雜訊變異量視為已知，並且以此來計算閾值。但事實上接收端並不會知道雜訊變異量，而是需要另行估計。因此，我們使用了最小描述長度(minimum description length)來估計雜訊變異量，以便通道估測的進行。模擬結果顯示透過此種方法，即使在不知道任何通道狀態資訊的情況下，我們也能夠準確地進行雜訊變異量的估計。

關鍵字

正交時頻空間調變 ； 估計雜訊變異量 ； 通道估測 ； 最小描述長度 ； 高速情境 ； 都普勒頻率偏移

並列摘要

Orthogonal frequency division multiplexing (OFDM) transmits signals in the time-frequency domain, thus it is sensitive to high Doppler shifts. To address this problem, orthogonal time frequency space (OTFS) modulation was proposed. OTFS modulation transmits signals in the delay-Doppler domain and converts the fading, time-varying channel into a time-independent channel with a complex channel gain. Hence, OTFS is more compatible with high Doppler shifts and suitable for high-speed communication scenarios. Raviteja \emph{et al.} proposed the guard pilot-based (GP-based) channel estimation for OTFS, which requires noise variance to calculate the threshold value. But in their work, noise variance was assumed to be known at the receiver, which is not practical. Thus, we use the minimum description length (MDL) algorithm to estimate noise variance. And the simulation results show that we can estimate noise variance accurately without any channel state information.

並列關鍵字

Orthogonal Time Frequency Space modulation (OTFS) ； noise variance estimation ； channel estimation ； minimum description length (MDL) ； high-speed communication scenarios ； Doppler shift

參考文獻

[1]P. Raviteja, Y. Hong, E. Viterbo and E. Biglieri, “Practical Pulse­Shaping Wave­forms for Reduced­Cyclic­Prefix OTFS,” inIEEE Transactions on Vehicular Tech­nology, vol. 68, no. 1, pp. 957­961, Jan. 2019.

[2]R. Hadani et al., “Orthogonal Time Frequency Space Modulation,”2017 IEEEWireless Communications and Networking Conference (WCNC), San Francisco,CA, 2017, pp. 1­6.

[3]P. Raviteja, K. T. Phan and Y. Hong, “Embedded Pilot­Aided Channel Estimationfor OTFS in Delay–Doppler Channels,” inIEEE Transactions on Vehicular Tech­nology, vol. 68, no. 5, pp. 4906­4917, May 2019.

[4]P. Raviteja, K. T. Phan, Y. Hong and E. Viterbo, “Embedded Delay­Doppler Chan­nel Estimation for Orthogonal Time Frequency Space Modulation,”2018 IEEE88th Vehicular Technology Conference (VTC­Fall), 2018, pp. 1­5.

[5]W. Shen, L. Dai, J. An, P. Fan and R. W. Heath, “Channel Estimation for OrthogonalTime Frequency Space (OTFS) Massive MIMO,” inIEEE Transactions on SignalProcessing, vol. 67, no. 16, pp. 4204­4217, Aug. 2019.