Article Preview
TopIntroduction
The future wireless mobile network's wheel turns rapidly, where the services of mobile Fifth Generation (5G) networks need to meet numerous, variant, and challenging requirements on data rates, time and frequency localization, latency, reliability, and Out-Of-Band (OOB) emissions (Xiang, Zheng, & Shen, 2017). Applications, that shall be supported by the 5G networks, can be grouped into three major categories: Internet of Things (IoT), Tactile Internet (TI), and gigabit wireless connectivity (Wunder et al., 2014). Every usage scenario needs different requirements and performance targets. IoT applications are characterized by a massive number of low-power consumption devices with small data-burst, which leads to scalability, long lifetime, and low-cost requirements (Wunder et al., 2014). TI is for real-time applications, comprised of extremely low latency and high-reliability communications (G P Fettweis, 2014). Gigabit wireless connectivity requires high throughput, a peak data rate of multiple Gbps, and an improved Spectral Efficiency (SE) (Xiang et al., 2017).
Cyclic Prefix-Orthogonal Frequency Division Multiplexing (CP-OFDM) has been widely utilized in Fourth Generation, thanks to its robustness in multipath fading channels, low-complexity implementation, and ease of generation based on Fast Fourier Transform (FFT) algorithms (Farhang-Boroujeny & Moradi, 2016).
CP-OFDM suffers from numerous disadvantages such as low SE due to the use of CP, high OOB and poor time-frequency localization, as a result of rectangular pulse-shaping for each subcarrier, and performance degradation caused by transmission over non-contiguous frequency bands (Farhang-Boroujeny & Moradi, 2016). CP-OFDM uses fixed parameters as CP length, transmission time interval, and subcarrier spacing that limit time and frequency multi-user scheduling (Zhang, Jia, Chen, Ma, & Qiu, 2015). Also, CP-OFDM imposes tight synchronization requirements which are unsuitable for asynchronous transmission. Following these arguments and because of the emergence of modern applications with broad requirements, the development of new waveforms becomes a necessity to overcome CP-OFDM limitations.
Generalized Frequency Division Multiplexing (GFDM), which was proposed by Fettweis (Gerhard P Fettweis, Krondorf, & Bittner, 2009), is a promising candidate for the 5G physical layer. This novel digital multi-carrier waveform, uses a circular filter bank approach on block-based structure where data symbols, called subsymbols, are grouped across several subcarriers and time slots.
GFDM transmits independent data blocks, which spread across a two-dimensional structure, with each block consisting of many subcarriers and subsymbols. This scheme offers high flexibility to allow for multi-user resource allocation in both time and frequency domains (Michailow, Krone, Lentmaier, & Fettwseis, 2012). In GFDM, each subcarrier is pulse-shaped using an adjustable prototype filter to reduce OOB emissions in a controlled manner, where the filter bank is a circularly time and frequency shifted version of the prototype filter (Michailow et al., 2014). Last but not least, GFDM scheme uses only a single CP for an entire symbols block, therefore it achieves an improved SE, through tail biting technique. Consequently, the existence of the CP enables simple and efficient FFT-based equalization (Gerhard P Fettweis et al., 2009). These features make this scheme suitable for cognitive radio and 5G applications.