In the paper, the 5G air interface FBMC, GFDM, UFMC, and f-OFDM are simulated in MATLAB 2022b. The 5G channel model used for all the simulation is Vehicular Jake’s model with velocity 130 km/h with channel frequency of 2.5 GHz.
Table 1
List of Values Used in Simulation
Waveform Technique | Simulation Parameter | Value |
FBMC | Number of Sub carriers | 12 |
Overlapping factor K | 4 |
Modulation type | 64 QAM |
Filter | PHYDYAS, Hermite |
Length of the weighting filter window | 43 |
Sub carrier spacing | 15KHz |
f-OFDM | Number of Sub-bands | 10 |
Modulation type | 64 QAM |
Filter | Hanning window |
GFDM | Block length | 32 |
Number of Sub carriers | 128 |
RRC filter Roll off | 0.2 |
Modulation type | 16QAM |
UFMC | Number of Sub-bands | 10 |
Sub band Size | 20 |
Sub-band Offset | 156 |
Modulation type | 4, 16, 64, 256 QAM |
Size of IFFT | 512 |
Filter | Dolph-Chebyshev |
Length of the weighting filter window | 43 |
Sideband filter attenuation | 40 dB |
Firstly the spectrum for these waveforms is illustrated in terms of the PSD. It is concluded from Fig. 6 that FBMC is the most spectrally efficient waveform with minimum OOBE.
The comparison of a BER performance of all the 5G waveforms with respect to Signal to Noise Ratio (SNR) is shown in Fig. 7. The figure shows that f-OFDM is having higher BER as compared to other multi-carrier modulation techniques. In contrast to this UFMC achieves optimal BER performance.
(a) FBMC simulation
The key feature in FBMC technique is its frequency selectivity through the use of spectrally efficient Prototype Filters which helps in coping with inter-channel interference (ICI) [15]. The characterization of the prototype filter is K. The factor of overlap K is the number of multicarrier symbols which overlap [16, 17]. Varying the value of K gives different filter windows depicted in Fig. 8. The analysis of FBMC with the metric of BER with different filters is depicted in Fig. 9.
(b) GFDM simulation
GFDM symbol is simulated with CP length = 0, Using 4 bits per symbol (16 QAM), and varying the block size. PSD for GFDM signal with different block size is shown in Fig. 10. From the figure it is observed that with increase in block size the PSD performance decreases and spectral leakage increases. GFDM has the flexibility to choose the number of blocks.
(c) UFMC simulation
Simulation parameters used for UFMC model implementation are 200 Subcarriers with Dolph Chebyshev filter for sub band filtering [13]. The performance parameters PSD and BER are simulated. The Fig. 11(a) shows the normalized spectrum of the UFMC system with the number of sub carriers 20 in 10 sub bands. To achieve good frequency spectrum utilization, higher order QAM mapping technique is to be applied. Figure 11 (b) shows UFMC system performance for 4QAM, 16QAM, 64 QAM, and 256QAM and concludes that BER increases as QAM order increases.
Table 2
Comparison of the 5G waveforms [18–20]
S. No. | Metrics for comparison of waveforms | Performance report |
1 | Compatibility with 4G (LTE ) and MIMO | UFMC and F-OFDM waveforms are compatible with LTE but FBMC and GFDM does not support backward compatibility |
2 | Latency | FBMC has highest latency and UFMC is used for low latency applications |
3 | Out of band emission (Spectral leakage) | Best filter performance and lowest OOBE is in FBMC |
4 | Compatibility with high mobile applications | FBMC is the best suitable waveform for high mobile applications |
5 | Flexible to adapt in different scenarios | UFMC is the most flexible to with various numerology |
6 | Compatibility for Short burst traffic | UFMC and GFDM are most suitable for short burst traffic |
7 | Transceiver Design complexity | UFMC and GFDM has most complexity |
8 | Best spectrum utilization | F-OFDM uses scattered frequency band so saves bandwidth |