Look-Up Table Based Implementation of Ultra-Low Complexity Narrowband OFDM Transmitters
Loulou, AlaaEddin; Yli-Kaakinen, Juha; Levanen, Toni; Lehtinen, Vesa; Schaich, Frank; Wild, Thorsten; Renfors, Markku; Valkama, Mikko (2019-10-21)
Loulou, AlaaEddin
Yli-Kaakinen, Juha
Levanen, Toni
Lehtinen, Vesa
Schaich, Frank
Wild, Thorsten
Renfors, Markku
Valkama, Mikko
IEEE
21.10.2019
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-202001081092
https://urn.fi/URN:NBN:fi:tuni-202001081092
Kuvaus
Peer reviewed
Tiivistelmä
In cyclic-prefix orthogonal frequency-division multiplexing (CP-OFDM) based radio access systems, the coexistence of different technologies without proper time-frequency synchronization is limited due to high out-of-band emissions. Therefore, filtered-OFDM (F-OFDM) type spectrum enhancement can play a key role to relax the synchronization requirements by delivering well-contained spectrum by using subband or resource block based filtering. This allows higher degree of flexible and dynamic spectrum use with minimized interference. However, this approach increases computational complexity compared with CP-OFDM. This paper presents a low-complexity solution for narrow-band F-OFDM transmitters based on the use of a look-up table (LUT) to store the F-OFDM waveform. This approach can be applied equally well for filtered version of the discrete Fourier transform-spread-OFDM (DFT-s-OFDM) waveform with small number of subcarriers. DFT-s-OFDM is commonly used in the uplink of OFDM-based systems to mitigate high peak-to-average power ratio (PAPR) of OFDM. Therefore, the scheme is particularly interesting for the transmitters of Internet-of-Things (IoT) or massive machine type communication (mMTC) devices. The LUT approach requires only memory units and relatively low number of additions. Moreover, we propose a low-complexity solution to deal with the CP-length variations within the transmission frame. The required memory wordlengths are evaluated through simulations. Comparisons with time-domain filtering and fast-convolution-based filtering solutions are included as well.
Kokoelmat
- TUNICRIS-julkaisut [17109]