The telecommunication industry has experienced enormous growth over the last few decades. It all started when the first generation (1G) of mobile telecommunication standards, which were analog systems, were introduced in the early years of the 1980s. 2G telecommunication systems soon replaced 1G. Compared to 1G, 2G was digitally encrypted, more spectrally efficient and also enabled users to now exchange plain text-based messages with its Short Message Service (SMS). In 1998, the first 3G networks were introduced which for the first time allowed mobile broadband access of several megabits per second (Mbps) to smartphones and other mobile devices. To satisfy the increased demand for more mobile broadband internet access came 4G systems, which was first commercialized in 2008. 4G has made it possible for mobile users to browse the internet at a much faster speed, stream high definition videos, do video conferencing and play their favourite video games online. Research is currently underway in both industry and academia to develop the fifth generation of mobile communication networks, 5G. It is expected that, compared to 4G, 5G will provide higher data speeds, utilize the available frequency spectrum more efficiently, improve coverage and reduce the end-to-end transmission time. In addition, 5G networks are expected to open the door towards new application-cases such as the Internet-of-Things (IoT).
In all these standards, one key technology is how signals are generated and transmitted from form one device to the other over the wireless medium. This is known as modulation, which is basically a process of varying one or more properties of a periodic signal, or carrier signal, with an information signal. For 1G networks, analog modulation schemes such as frequency modulation (FM) or amplitude modulation ware used. Digital modulation schemes e.g. amplitude-shift-keying (ASK) and frequency-shift-keying (FSK) replaced their anolog counterparts in 2G and 3G networks. In 4G systems, multicarrier modulation technology was implemented, where digital data is mapped onto multiple carrier frequencies. The information generated by a user terminal is carried on a time domain signal which is referred to as a waveform. The waveform also determines the shape of the wireless signal in the frequency domain according to its shape in the time domain.
Orthogonal frequency-division multiplexing (OFDM) is the multi-carrier modulation method that was adopted for 4G networks. Information is transmitted over multiple overlapping subcarriers. A guard interval/band is introduced to maintain orthogonality between subcarriers especially for signals transmitted through fading channels. In addition, there is the use of cyclic-prefix between successive OFDM symbols to help reduce interference between symbols. The main advantage of OFDM is its ability to travel through severe channel conditions and still be able to be demodulated at the receive terminal. Also, it can effectively support the use of multiple antennas at the transmitter and receiver, a technique known as multiple-input-multiple-output (MIMO). Due to these and other advantages, OFDM has also been adopted for Wifi. However, to harness all these benefits, strict synchronization and subcarrier orthogonality must be maintained in OFDM. Other limitations include its low spectral efficiency and high-peak-to-average-power ratio (PAPR).
To address the drawbacks of OFDM and set the tone for the 5G air interface, new waveforms have been proposed. On one hand, there are proposals to maintain OFDM as the fundamental waveform for 5G by finding appropriate solutions to some of its shortcomings, making the new waveform compatible with existing technologies. On the other hand, different multicarrier schemes have been intensely studied as alternatives to OFDM. For the first release of 5G 3GPP (3rd Generation Partnership Project), which the standardization body for mobile communication networks, has opted for a modified version OFDM, instead of a different waveform, mainly because it will be compatible with existing 4G networks.
The internet of things (IoT) has immerged as an exciting application case in 5G networks. Basically, IoT involves the extension of internet connectivity beyond typical devices, such as laptop computers and smartphones, to everyday devices such as home appliances and vehicles. It will provide internet connection to these devices and make it possible for them to interact and exchange information. Most of these IoT applications will require asynchronous and uncoordinated user transmissions. Hence, the strict synchronization and orthogonality of OFDM may not serve these applications effectively. Also, there is a significant loss in spectrum efficiency in OFDM due to the use of large frequency guard bands to limit interference at the receiver.
Therefore, in order to support asynchronous and uncoordinated transmissions for IoT devices, the 5G waveform must relax the strict synchronization and orthogonality conditions of OFDM-based waveforms. To this end, non-orthogonal waveforms such as filter bank multicarrier (FBMC) and generalized frequency division multiplexing (GFDM) have been intensely investigated in the literature. These non-orthogonal waveforms address most of the limitations of OFDM. However, because of the lack of orthogonality in these waveforms, there is the problem of interference between the many subcarriers that make up the time domain symbols. This interference is known as inter-carrier interference (ICI). Hence, the key design problem for non-orthogonal waveform is how to overcome the inherent ICI.
There have been a number of proposed transmitter and receiver designs in the literature to reduce ICI in both FBMC and GFDM systems. In our research, we have initially set out to investigate a variant of FBMC systems known as FBMC-QAM (filter bank multicarrier-quadrature amplitude modulation). To help reduce the level of ICI at the receiver, we have proposed a receiver design that cancels some of the interference in the system by iteratively detecting the received signal a number of times. After each iteration, the detected interference signal is subtracted from the received signal, thereby improving the credibility of the final detection.
We are also considering other alternative approaches to eliminating the intrinsic ICI in FBMC-QAM. I will be able to share more about these solutions in my next article.