Introduction

In mobile communication channels the received signal is a combination of many components arriving from various directions due to multipath propagation resulting in a large fluctuation in the received signals. This phenomenon is called fading.
If two or more radio channels are sufficiently separated in space, frequency or time (and sometimes in polarization), their fadings are independent, i.e. it is rare for all the channels to fade simultaneously. Diversity techniques use this very fact to combat the effect of fading. Several received signals are used together to improve the SNR or some other performance parameter.

In general, the term “diversity system” refers to a system in which one has available two or more closely similar copies of some desired signal.

A rake receiver is a radio receiver designed to counter the effects of multipath fading by using several "sub-receivers" each delayed slightly in order to tune in to the individual multipath components. Each component is decoded independently, and later combined in order to make the most use of the different transmission characteristics of each channel. This could very well result in higher Signal-to-noise ratio (or Eb/No) in a multipath environment than in a "clean" environment.

Many of the current and emerging wireless communication technologies and systems make use of some form of diversity combining to combat the fading and shadowing effects induced by the harsh, complex, and dynamic channel. In designing such systems to achieve a specified quality of service, one is also faced with the practical reality of satisfying a predetermined degree of complexity. Thus, although the optimum (from a performance standpoint) diversity combining techniques such as maximal-ratio combining (MRC) (for coherent systems) and equal gain combining (EGC) (for non-coherent and differentially-coherent systems) are well-known, one often opts in favor of a less complex scheme at a sacrifice in performance.

Basic Model

To understand the basic idea behind diversity combining, consider a radar system which operates by storing the signal received during one scan and adds this to the signal received during the next scan. if f₁(t) and f₂(t) represent the two signals, then the composite signal is f₁(t) + f₂(t) = f(t). Now, since both the signals would have a desired message component and an undesired noise component, the composite signal may be written as:

f(t) = [s₁(t) + s₂(t)] + [n₁(t) + n₂(t)]

If the message components s₁ and s₂ are closely similar, their sum would be an approximate large copy of s₁ or s₂. On the other hand, the noise components n₁ and n₂ may be quite dissimilar, thereby partially cancelling each other. The signal f(t) may then be a better signal than f₁(t) or f₂(t) alone.
More generally, one may have N such copies f₁(t),f₂(t),f₃(t),...,f_N(t) each of the form f_j(t) = s_j(t) + n_j(t) and one may form the sum

f(t) = f₁(t) + f₂(t) + ... + f_N(t) = _j=1Σ^N f_j(t)

which may outperform, in some sense, the individual f_j(t).
As f_j may have fluctuating local statistics, it is more convinient to consider a weighted sum of the f_j components, i.e.

f(t) = a₁f₁(t) + a₂f₂(t) + ... + a_Nf_N(t) = _j=1Σ^N a_jf_j(t)

(1)

The combining coefficient a_j is propotional to the channel gain and may be allowed to vary in accordance with local charactistics of f_j(t).

Diversity Combining Techniques

Scanning Diversity

In this technique, at any given time, all but one of the a_j (refer equation (1)) will be zero, and the non-zero a_j will be 1. A selector device scans the channels in a fixed sequence until finding a signal above a preset threshold, uses that signal only until it drops below threshold, and then scans the other channels in the same fixed sequence until it again finds a signal above threshold. It is often applied to the case of two antennas supplying a single receiver through the switch, which is why it is sometimes called antenna selection diversity.

Selection Diversity

This technique is very similar to the one mentioned above except that in this case, the system picks out the best of the N noisy signals f₁,f₂,...,f_N and uses that one alone. The selection in such systems is mostly by electronic means.