Wireless data rates, part 2: High-order and multi-carrier modulation

3.2 Higher data rates within a limited bandwidth: Higher-order modulation
As discussed in the previous section, providing data rates larger than the available bandwidth is fundamentally in-efficient in the sense that it requires un-proportionally high signal-to-noise and signal-to-interference ratios at the receiver. Still, bandwidth is often a scarce and expensive resource and, at least in some mobile-communication scenarios, high signal-to-noise and signal-to-interference ratios can be made available, e.g. in small-cell environments with a low traffic load or for mobile terminals close to the cell site. Future mobile-communication systems, including the evolution of 3G mobile communication, should be designed to be able to take advantage of such scenarios, that is should be able for offer very high data rates within a limited bandwidth when the radio conditions so allow.

A straightforward means to provide higher data rates within a given transmission bandwidth is the use of higher-order modulation, implying that the modulation alphabet is extended to include additional signaling alternatives and thus allowing for more bits of information to be communicated per modulation symbol.

In case of QPSK modulation, i.e. the modulation scheme used for the downlink in the first releases of the 3G mobile-communication standards (WCDMA and CDMA2000), the modulation alphabet consists of four different signaling alternatives. These four signaling alternatives can be illustrated as four different points in a two-dimensional plane (see Figure 3-2a). With four different signaling alternatives, QPSK allows for up to 2 bits of information to be communicated during each modulation-symbol interval. By extending to 16QAM modulation (Figure 3-2b), 16 different signaling alternatives are available. The use of 16QAM thus allows for up to 4 bits of information to be communicated per symbol interval. Further extension to 64QAM (Figure 3-2c), with 64 different signaling alternatives, allows for up to 6 bits of information to be communicated per symbol interval. At the same time, the bandwidth of the transmitted signal is, at least in principle, independent of the size of the modulation alphabet and mainly depends on the modulation rate, i.e. the number of modulation symbols per second. The maximum bandwidth utilization, expressed in bits/s/Hz, of 16QAM and 64QAM are thus, at least in principle, two and three times that of QPSK, respectively.

Figure 3-2. Signal constellations for (a) QPSK, (b) 16QAM and (c) 64QAM.

It should be pointed out that there are many other possible modulation schemes, in addition to those illustrated in Figure 3-2. One example is 8PSK consisting of eight signaling alternatives and thus providing up to 3 bits of information per modulation symbol. Readers are referred to [50] for a more thorough discussion on different modulation schemes.

The use of higher-order modulation provides the possibility for higher bandwidth utilization, that is, the possibility to provide higher data rates within a given bandwidth. However, the higher bandwidth utilization comes at the cost of reduced robustness to noise and interference. Alternatively expressed, higher-order modulation schemes, such as 16QAM or 64QAM, require a higher E_b/N₀ at the receiver for a given bit-error probability, compared to QPSK. This is in line with the discussion in the previous section where it was concluded that high bandwidth utilization, i.e. a high information rate within a limited bandwidth, in general requires a higher receiver E_b/N₀.

3.2.1 Higher-order modulation in combination with channel coding
Higher-order modulation schemes such as 16QAM and 64QAM require, in themselves, a higher receiver E_b/N₀ for a given error rate, compared to QPSK.

However, in combination with channel coding the use of higher-order modulation will sometimes be more efficient, that is require a lower receiver E_b/N₀ for a given error rate, compared to the use of lower-order modulation such as QPSK. This may, for example, occur when the target bandwidth utilization implies that, with lower-order modulation, no or very little channel coding can be applied. In such a case, the additional channel coding that can be applied by using a higher-order modulation scheme such as 16QAM may lead to an overall gain in power efficiency compared to the use of QPSK.

As an example, if a bandwidth utilization of close to two information bits per modulation symbol is required, QPSK modulation would allow for very limited channel coding (channel-coding rate close to one). On the other hand, the use of 16QAM modulation would allow for a channel-coding rate in the order of one half. Similarly, if a bandwidth efficiency close to 4 information bits per modulation symbol is required, the use of 64QAM may be more efficient than 16QAM modulation, taking into account the possibility for lower-rate channel coding and corresponding additional coding gain in case of 64QAM. It should be noted that this does not speak against the general discussion in Section 3.1 where it was concluded that transmission with high-bandwidth utilization is inherently power in-efficient. The use of rate 1/2 channel coding for 16QAM obviously reduces the information data rate, and thus also the bandwidth utilization, to the same level as uncoded QPSK.

From the discussion above it can be concluded that, for a given signal-to-noise/ interference ratio, a certain combination of modulation scheme and channel-coding rate is optimal in the sense that it can deliver the highest-bandwidth utilization (the highest data rate within a given bandwidth) for that signal- to-noise/interference ratio.