In this section, receiver structures that can be used for separation and processing of multipath signals are presented. The uplink case is considered in which smart antennas are implemented at the Base Station. In general antenna system can be separated into two parts as illustrated in Figure 1.

• Analog Part: It consists of antenna arrays, RF modules (amplifiers, filters, mixers, oscillator), inphase / quadrature (IQ) modulators, ADCs etc. Array consists of antenna elements distributed in any pattern, usually linear equally spaced (LES)(Figure a & c), and uniformly circular (Figure b) or uniformly spaced planar array (Figure d) as illustrated in Figure 2. Here each dot in array represents the antenna element and $\Delta x$, $\Delta y$ and $\Delta z$ represent the spacing between elements in $x, y$ and $z,$ direction respectively. $\Delta \phi$ represents the angular separation between elements in circular array.

• Digital Part: As illustrated in Figure 1, It consists of DSP to perform complex signal processing operations, weight calculation and beamforming.

i) Signal Processing: DSP performs various complex signal processing operations such as,

a) Estimation of the direction of an arriving signal: The most important signal processing operation in smart antenna is evaluation of the direction of an arriving signal. This is normally referred as Direction of Arrival (DoA). DoA is estimated using various techniques available in the literature such as,

• MUSIC (Multiple SIgnal Classification),
• estimation of signal parameters via rotational invariance techniques (ESPRIT) algorithms,
• Matrix Pencil method or one of their derivatives.

These algorithms involve finding a spatial spectrum of the antenna array, and calculating the DOA from the peaks of this spectrum. These calculations are computationally intensive. Matrix Pencil is very efficient in case of real time systems, and under the correlated sources. The detailed explanation of these techniques is beyond the scope of this book.

b) Separation of signals of interest, interfering signals and multipath signals: Among many received signals at receiver, the desired signal is distinguished from other interfering signals by adaptive process using temporal information (reference signal) or spatial information (direction of user).

ii) Weight Calculation: It is known that in an antenna array, the direction of main lobe and nulls in radiation pattern can be changed by changing the length of elements, spacing between elements, the amount of magnitude and phase of signal fed at each element. If once all these parameters are fixed, a fixed radiation pattern is obtained. However in smart antenna adaptive radiation pattern is required according to the direction of user. This can be achieved by changing the amount of magnitude and phase of the signal fed to each element.

Each branch of the smart antenna array has a weighting element associated with magnitude and a phase. It is used to decide the amount by which the amplitude and phase of the received signal is to be changed. Based on the received signal the weights (in terms of magnitude & phase) are calculated by adaptive process using temporal information (reference signal) or spatial information (direction of user). Thus by assigning different weights to different signals, the main beam of the array is directed towards the desired user, while the nulls are adjusted toward the interferer.

iii) Beam Forming: The purpose of a smart antenna is to form dynamically main lobe of the beam towards the user of interest, and null beams in the direction of interferers. This is accomplished by adjusting the weights of antenna array elements. This requires digital beam forming (DBF) software, which forms an integral part of a smart antenna. The smart antenna is also known as digital beam former antenna.

The output $Y(i)$ of the DBF at discrete time i is given by a linear combination of the data at the M antenna elements and is given by,

$Y(i) = W^{H} X(i) \cdots (1)$

Where,

$X(i)$: The received signal vector on the antenna at time instant i

$θ_0$: Single direction of arrival in the discrete domain

W : The weight vector whose elements $(w_1,w_2... w_M)$ are the weights applied to the received signal on each branch of the array.

Superscript H represents Hermitian matrix of W, i.e. conjugate transpose.

$X(i)$ can be expressed as, $ X(i)=A(\theta) S(i) \cdots(2)$.

Where, $A(\theta)$ is the array response vector and $S(i)$ is the incoming signal at the receiver at time instant i.

If $(\theta_1, \theta_2... \theta_M)$ are the direction of arrival of different signals.

For half wavelength element spacing, array response vector is represented as,

$\mathrm{A(\theta)}=\left[\quad \mathrm{e}^{\mathrm{j} \pi \sin \theta_{1}} \mathrm{e}^{\mathrm{j} \pi 2 \sin \theta_{2}} \ldots \ldots \mathrm{e}^{\mathrm{j} \pi(\mathrm{M}) \sin \theta_M}\right]^T \cdots(3)$

If we let the weights be,

$\mathrm{W}=\left[\quad \mathrm{e}^{-\mathrm{j} \pi \sin \theta_{1}} \mathrm{e}^{-\mathrm{j} \pi 2 \sin \theta_{2}} \ldots \ldots \mathrm{e}^{-\mathrm{j} \pi(\mathrm{M}) \sin \theta_M}\right]^T \cdots(4)$

Then we have,

$\mathrm{W}^{\mathrm{H}}=\mathrm{A}\left(\theta \right) \mathrm{M}\cdots(5)$

Substituting (2) and (5) into (1), we get

$Y(i)=W^{H} X(i)=M S(i) \cdots(6)$

Thus, in an ideal situation the signal can be totally recovered with a gain of M. In practice many constraints are taken in to considerations, including estimation of the delay, the DOA, and multipath fading, etc.. Digital beam former weight vector (W) can be adjusted to form a beam in the desired direction and deep nulls in the direction of interfering signals. This is illustrated in Figure 3. Thus cochannel interference can be reduced significantly.

• The Steepest Descent Gradient method,
• Maximum Likelihood method,
• Minimum Variance method, and
• Least Mean Squares (LMS) algorithms.

Interested readers can refer the literature to understand the weight calculation and beamforming algorithm in detail.

Types of Smart Antennas: Smart antenna systems are categorized as either switched-beam or adaptive-array systems.

a) Switched Beam Antennas: A switched-beam antenna is also known as multi-beam antenna or phased array antenna. It is an antenna array that forms numerous numbers of fixed beams in different directions based on fixed weights assigned to each received signal. i.e., beams direct in certain fix discrete directions as illustrated in Figure 4. Among many beams, one beam will turn on or will be steered towards the desired user. A switch is used to select the best beam to receive a particular signal, offering maximized received power. In other words, the desired signal beam moves along with the movement of the user.

Advantages:

The main advantages of the switched-beam approach are as follows:

• Simplicity: It is easier to implement as it requires only a beam forming network, an RF switch and control logic to select a particular beam.
• Economical: All processing is done in the RF domain, so that only a single signal has to be converted to the baseband and processed there. Since down conversion circuits are among the most expensive components in today’s wireless systems, this is a significant advantage.

Drawbacks:

The drawbacks of this scheme are as follows:

• Limited Flexibility: The main beam can only be pointed in certain fixed directions, and it is not necessary that user should be present in the direction of any one beam, thus the gain in the actual direction of the user might not be the maximum achievable value.
• Limited Interference Reduction: In this mechanism, nulls cannot be pointed in arbitrary directions, the position and number of nulls is fixed. Hence the null cannot be created in the direction of undesired signal (interferer) i.e. Nulling of interferers is very ineffective. For these reasons, switched-beam antennas seem to be more suitable for CDMA applications, where signal enhancement is critical, and not for SDMA applications, where interference suppression is essential.

b) Adaptive Array Antenna: In this type of antennas, directions of arrival from the users are first estimated, and then the weights of the beam formers are calculated in accordance with the specified directions. The signals that are received will be weighted and later combined to increase the desired signal to interference in addition to the signal and noise power ratio [S/N]. The mechanism is adaptive, so as desired user moves, there will be a change in the beam pattern. The main beam points towards the desired user and nulls towards interferer. Thus, the interference will be balanced as the desired signal will be in the direction of the main beam. The Figure 5 illustrates the beam pattern of an adaptive array antenna. The antenna can easily steer the main beam to any direction, while at the same time nullifying the interfering signal. The direction of the beam can be calculated using the Direction of arrival algorithm method.

The main features of adaptive antenna arrays are as follows:

• Requirement of DOA estimation is required.
• Calculation of weights of Beam-formers in accordance with the environment in specified directions is required.
• Using DOA algorithms, the direction of interferes are determined.
• Beam patterns are adjusted to null the interferers.
• Used for Downlink with FDD & TDD modes.
• Offers maximum received power and simultaneously reduces interference.
• This approach optimizes the signal to interference ratio and mitigates the effect of multipath Fading.
• It is applicable to NLOS environments.
• No Reference signal is required.