Multicarrier Modulation and Multiple Access

The requirements for the IMT-Advanced mandate that the utilized multiple access technologies are backward compatible with IMT-2000 (3G) systems. To support different services, both contention and contention-free multiple access should be supported. In addition, as a step towards interference control, FR should be supported. At the same time, in order to accommodate heterogeneity in regulations between different regions, both TDD and FDD duplexing schemes should be supported as well, including half and full duplex FDD.

In order to abide by these requirements while achieving the promised levels of performance, LTE-Advanced as well as WiMAX resort to multicarrier techniques. In particular, three techniques are used, namely OFDM, SC-FDMA, and OFDMA. While WiMAX uses OFDMA in both the uplink and the downlink, LTE-Advanced uses OFDMA for the downlink only while using SC-FDMA uplink.

An advantage of multi-carrier access techniques is their robust communication and stable interference management. In fact, multicarrier techniques facilitate fractional FR which will be discussed in subsequent sections. In addition, they also allows exploiting multiuser diversity at smaller granularities than was ever possible in CDMA-based networks. Another advantage of multicarrier techniques is enhancing system throughput by mitigating the frequency-selective randomness, that is, frequency selective fading. This enhancement is achieved by modulating orthogonal subcarriers, and allows these techniques to support different levels of user mobility and withstand different communication conditions as shall be elaborated further shortly.

OFDM

OFDM is probably one of the most striking advances in access technologies. It facilitates higher transmission rates with a reasonable equalization and detection complexities. This high transmission is achieved through modulating a set of narrowband orthogonal subcarriers. In particular, an OFDM block is built as shown in Figure 2.1. The sequence of L modulated symbols, x₀, x₁, ..., x_L−1, are converted into L parallel streams before taking the N-point Inverse Fast Fourier Transform (IFFT) of each. The possible mismatch between L and N is overcome by zero padding the remaining N − L inputs of the IFFT block. Next, the N outputs, s₀, s₁, ..., s_N−1 are converted back to a serial stream before adding the Cyclic Prefix (CP). Finally, the resulting OFDM block is converted to its analog form prior to sending it over the channel.

Using this architecture, an OFDM block can resist the Inter-Carrier-Interference (ICI) by employing orthogonal subcarriers, that is, as a result of using the IFFT. It is also capable of mitigating the channel time dispersion by inserting the CP. In fact, the insertion of the CP is a widely used technique to create a so called guard period between successive OFDM symbols. The CP is simply a repetition of the last part of the preceding OFDM symbol. The length of this repetition is made long enough to exceed the channel delay spread, hence mitigating the channel delay spread causing Inter-Symbol-Interference (ISI). In addition, the detection process turns to a circular convolution process which enhances the signal detection capabilities and simplifies the equalization process.

Figure 2.1 OFDM modulation using the IFFT.

OFDM Demodulation reverses the aforementioned processes. After converting the received signal back into the digital domain, the CP is removed. Next, the signal is converted into a parallel N data streams before performing an N-point FFT. Finally, the sequence is returned back into a serial one. These functionalities are shown in Figure 2.2.

Despite the many advantages of OFDM, actual implementations revealed some challenges. Probably the most famous one is the high Peak to Average Power Ratio (PAPR) problem. Simply put, high PAPR, which results from the coherent addition of the modulated subcarriers, reduces the efficiency of the power amplifier. The high PAPR also sophisticates the Analog to Digital (ADC) and Digital to Analog (DAC) processes [1]. While these two disadvantages can be overcome at the base station side, they form a serious challenge to the battery-powered Mobile Station (MS). Consequently, 3GPP replaced OFDM at the uplink in their IMT-Advanced proposal by SC-FDMA. However, before looking at this novel multiple access technique, let us look at the OFDM multiple access version, namely the OFDMA.

Figure 2.2 OFDM demodulation.

OFDMA

In OFDM, all subcarriers are assigned to a single user. Hence, for multiple users to communicate with the BS, the set of subcarriers are assigned to each in a Time Division Multiple Access (TDMA) fashion. Alternatively, an OFDM-based multiple access mechanism, namely the OFDMA, assigns sets of subcarriers to different users. In particular, the total available bandwidth is divided into M sets, each consisting of L subcarriers. Hence, a total of M users can simultaneously communicate with the BS. Subcarrier assignment can be either distributed or localized, as is shown in Figure 2.3.

While in localized assignment, chunks of contiguous subcarriers are allocated to each user, distributed assignment allocates equidistant subcarriers to different users.

Despite the relatively straightforwardness of OFDMA, it has very attractive advantages. Probably the most important of these is its inherent exploitation of frequency and multiuser diversities. Frequency diversity is exploited through randomly distributing the subcarriers of a single user over the entire band, reducing the probability that all the subcarriers of a single user experience deep fades. Such allocation is particularly the case when distributed subcarrier assignment is employed. On the other hand, multiuser diversity is exploited through assigning contiguous sets of subcarriers to users experiencing good channel conditions [2].

Another important advantage of OFDMA is its inherent adaptive bandwidth assignment. Since the transmission bandwidth consists of a large number of orthogonal subcarriers that can be separately turned on and off; wider transmission bandwidths, as high as 100 MHz, can be easily realized.

SC-FDMA

Amongst the many methods proposed and studied to reduce the PAPR of OFDM, SC-FDMA was practically adopted in both, LTE and LTE-Advanced. This OFDM-based multiple access method overcomes the PAPR problem through two additional processes, one at either side of the communication system. More specifically, an L-point Discrete Fourier Transform (DFT) stage is inserted just before the N-point IFFT at the transmitter side, while an L-point Inverse DFT (IDFT) is applied to the L outputs of the N-point FFT at the receiver side.

Figure 2.3 Distributed and localized subcarrier assignment strategies.

Since the only modification happens before assigning the different subcarriers, multiple access can be done in a similar way like OFDMA. Accordingly, SC-FDMA possesses the same advantages as OFDMA while experiencing lower PAPR. The adoption of SC-FDMA enhances the power utilization efficiency of the MS batteries, hence prolonging their lifetimes. In fact, LTE-Advanced MSs will use hybrid circuits, where SC-FDMA is used for long-range transmissions, that is, macrocell coverage, while OFDMA is used for short range transmissions, for example, femtocell coverage.