1. Introduction

Multiple-input multiple-output (MIMO) systems have attracted much attention as a means to improve the capacity of wireless communications by increasing the number of antennas. However, implementing multiple antennas can be a problem, particularly in mobile terminals due to their space limitation. In this study, we focus on array antennas at the receiver to enhance the capacity.

To resolve the problem, several methods have been proposed to achieve multiple separate received signal components by using a single radio frequency (RF) frontend with electronically steerable passive array radiator (ESPAR) antennas [1]. The modulated scattering array antenna was proposed for diversity and MIMO receivers [2–6]. The antenna consists of an antenna element for receiving signals and several modulated scattering elements (MSEs) like ESPAR antennas. The impedance of an MSE can be modulated or changed by with an applied sinusoidal voltage of frequency f<sup>s</sup> to the variable reactance element connected to the MSEs. Then, the antenna patterns can vary also in a sinusoidal manner and can make the received signal frequency-shift by fs. The frequency-shifted components can be used for

diversity reception. The virtually rotating antenna was proposed also to diversify the received signal components in the frequency domain [7, 8]. The principle of setting diversity branches in this domain is similar to that of the modulated scattering array antenna. In the rotating antenna, a combination of reactance values, which generates a desired antenna pattern, is applied sequentially to the multiple reactance elements of the antenna to rotationally change the directivity of the antenna. Also investigated was an ESPAR antenna based on the diversity receiver whose antenna patterns are time variable in a sinusoidal manner and are suitable for the MIMO-orthogonal frequency-division multiplexing (OFDM) receiver [9]. Other researches have been studied on diversity and MIMO receivers with ESPAR antennas having periodically variable antenna patterns by both theoretical and experimental investigations [10–14]. They also investigated the reactance time sequence, which generates sinusoidal antenna patterns with suppressed higherorder harmonics [15, 16].

In the studies shown above, it can be seen that, instead of using a fixed antenna pattern that may satisfy some criteria, they constantly changed the antenna pattern to generate multiple received signal components in the frequency domain. In this study, we propose a concept of antenna pattern multiplexing (APM) for setting multiple virtual antennas at the same location without additional physical antenna elements. Since the proposed APM also periodically varies antenna patterns to build multiple diversity branches or virtual antennas, it may be possible to consider the APM as a generalised method of the previously mentioned related studies. In APM, instead of sinusoidal waveforms or a sum of sinusoidal waveforms with different frequencies, we apply the sum of a set of orthogonal code sequences as the waveform to change antenna patterns. Therefore, the received signal can be separated into code domains to exploit path diversity instead of using only the narrow frequency domain, which is the case for the previous studies.

antenna pattern hanging units, cables between antenna elements and receivers, and

The receiver models employing the conventional array antenna, array antenna with APM, and ESPAR antenna

The antenna elements are connected to AP changing units, which are weight multiplication for array antennas and variable reactance elements (VREs) for ESPAR antennas. In the conventional array antenna, constant weights w1, w2, and w<sup>3</sup> are multiplied to form an antenna pattern based on a criteria such as maximising the signal-to-interference-plus-noise ratio (SINR) or minimising the interference. In array antenna with APM, the weights are functions of time w1ð Þt , w2ð Þt , and w3ð Þt , which form multiplexed AP and make the received signals travelling through antenna elements separable. Thus, we can add the signals and carry them to the receiver by a single cable. That is, the array antenna with APM can reduce the number of cables between antenna elements and the receiver and decrease their calibration cost. Since the parasitic elements do not connect to the receiver in ESPAR antenna, the cable cost can be also minimised. In ESPAR antenna, the antenna pattern or the directivity can be changed by the reactance values contributed by the parasitic elements. The reactance values of VREs can be changed by the

connected to receivers, while ESPAR antenna has an element connected to the corresponding receiver and two parasitic elements which are connected to variable reactance components. In the figure, the shaded elements in ESPAR antenna show parasitic elements. In this part, the distance between neighbouring elements should be more than a half wavelength λ=2 for array antennas to reduce the correlation between the received signals obtained by the elements. On the other hand, since ESPAR antennas form antenna patterns by exploiting mutual coupling between antenna elements, the neighbouring elements need to be sufficiently close to each other. A study on ESPAR antenna with six parasitic elements describes that λ=4 is an appropriate distance [18]. In our previous work on two-element ESPAR antenna, appropriate distances between the elements are around λ=8 [16]. Hence, ESPAR antennas can reduce the space required for antenna elements less than a half of the space of array antennas in the case of three antenna elements. The increase in the number of antenna elements provides more gains in terms of reducing antenna sizes

In the part of antenna elements, the array antennas have three antenna elements

receivers which receive signals through the cables.

Antenna Pattern Multiplexing for Enhancing Path Diversity

DOI: http://dx.doi.org/10.5772/intechopen.89098

for ESPAR antennas.

77

Figure 1.

with APM of three antenna elements.

We introduce an antenna pattern matrix that consists of coefficients for each code sequence for each direction of received paths. With the matrix, we can derive the received signals of MIMO systems that use APM-based receivers in a form similar to the signals of the conventional MIMO systems. The ergodic capacity for the MIMO systems with APM technique is also derived<sup>1</sup> . Numerical results show that the capacity can be improved by increasing the number of arrival paths and the number of virtual antennas when the coefficients of APM are randomly distributed.
