id="p-0047" num="0048">The proposed design concept and exemplary results are provided to crystalize the proposed approach to TMM antenna design.

(I): Design Concept and Exemplary Results for TMM Antennas

The design concept is to extend the time modulation of the antenna feeds for element weight control to the control of mutual coupling.

The basic idea uses the observation that if adjacent antenna elements are never allowed to be simultaneously in the “on” state (the “on” state is used here to define when the antenna port is connected to their respective receiver or transmitter), then the effect of mutual coupling can be significantly reduced. This concept is based on the property that antenna elements that are open circuited will not have an impact on an adjacent element. This is most straightforwardly appreciated by considering an N×N multiport impedance model of the antenna as:

$\begin{matrix} [\begin{matrix} V_{1} \\ ⋮ \\ V_{N} \end{matrix}] = [\begin{matrix} Z_{11} & \dots & Z_{N 1} \\ ⋮ & ⋱ & ⋮ \\ Z_{1 N} & \dots & Z_{NN} \end{matrix}] [\begin{matrix} I_{1} \\ ⋮ \\ I_{N} \end{matrix}] & Equation (1) \end{matrix}$

- where V_nand I_nare the voltage and current on antenna port n while Z_mnis the antenna impedance matrix. Note that all quantities are frequency dependent but we have dropped that from the notation for convenience.

If all ports in an antenna system in Equation (1) are left open, except for “port (1),” then the impedance seen at “the port (1)” will only be Z₁₁and no mutual coupling will be experienced because all other ports are open. This is the principle intended to be exploited in the design of TMM antennas in the present invention. That is, antennas with adjacent ports will never be simultaneously time modulated to be on together. Therefore, mutual coupling effects from adjacent antennas will be completely removed. While the effect of mutual coupling will be removed there are tradeoffs including power and harmonic issues and these are addressed in the first objective as explained earlier.

To verify the basic concept, exemplary simulations have also been performed.

FIG. 2 depicts an antenna design of a two-port planar inverted-F antenna (PIFA) utilized operating at 2.6 GHz according to one embodiment of the present invention. The PIFA have very low separation to demonstrate the effects of mutual coupling. As shown in FIG. 2, the geometry for a basic two-port PIFA design operating at 2.6 GHz, assuming perfect conductors for testing purposes, is provided. In the illustration, it can be seen that the two antennas are spaced only 15 mm apart (less than 0.2 wavelengths) and therefore mutual coupling effects will be significant. Computer Simulation Technology (CST) studio simulations have then been performed.

FIG. 3A and FIG. 3B show simulation results for the two-port antenna design in FIG. 2 when both ports are loaded and are therefore in the “on” state. Specifically, In FIG. 3A, it can be seen that efficiency is poor. From FIG. 3B, it can be seen that mutual coupling S₁₂is poor (the lines for S₁₁, S₂₂and S₁₂, S₂₁overlap due to the symmetry of the design). Both of these results are due to the high mutual coupling between ports. Further, in FIG. 3A and FIG. 3B, the efficiency and S-parameter results are provided for when both ports are fed and loaded normally with 50 ohms. As can be seen efficiency is poor at around 50% due to the coupling between antennas. In addition, from the S-parameters, it can be seen that S₁₂isolation is poor at around 6 dB. This result is expected because the mutual coupling between ports causes significant power to be returned to the adjacent source reducing efficiency.

FIG. 4A and FIG. 4B show simulation results for the two-port antenna design in FIG. 2 with an open circuit at port 2. From FIG. 4A, it can be seen that efficiency is excellent as predicted by equation (1). This is due to the open circuited port removing the effect of mutual coupling. From FIG. 4B, it can be seen S₁₁is now well matched at 2.6 GHz. S₁₂, S₂₁cannot be shown as port 2 is open.

In FIG. 4A and FIG. 4B, the results are presented for when port 2 is left open, showing significant differences compared to FIG. 3A and FIG. 3B. Efficiency at 2.6 GHz is now nearly 95%. Port 1 is now matched with S₁₁being less than-10 dB. These results have occurred because the adjacent port cannot return power induced by mutual coupling to its source. In FIG. 4A and FIG. 4B, it is also noted that S₁₂and S₂₂cannot be provided because port 2 is open. This is a subtle issue because the provided solution does not reduce mutual coupling. The mutual coupling remains the same as in FIG. 3A and FIG. 3B. However, the provided solution removes the effect of mutual coupling. Therefore, the measured efficiency is a main measure of how well the technique is performing. In this case, it has increased efficiency significantly as the effect of mutual coupling has been largely removed.

These exemplary results demonstrate that utilizing time modulation between adjacent ports can significantly reduce the effects of mutual coupling. Using this principle, further investigation is proposed to determine whether this approach can be useful in practice. Therefore, the following sections describe approaches to meet the objectives listed earlier.

(II): Understanding System Tradeoffs for TMM Antennas

Using the design concept in “(I): Design Concept and Exemplary Results for TMM Antennas,” it is possible to significantly reduce the effect of mutual coupling. However, there are some system tradeoffs that need to be considered first. The first is the effect of time modulation on receiver power and the second is the effect of the harmonics generated by the switching. These are addressed in the following.

The first issue is the direct reduction in power received by ports being time modulated. If adjacent ports need to be time modulated as open, then there will be less received power at that port. Assuming that sets of adjacent antennas can be modulated open and “on” alternatively the port “on” time will be reduced by approximately half. This will cause 3 dB loss of received power.

The second issue is the reduction in received power caused by the harmonics generated from the switches. These harmonics can be removed by filtering at the receiver side but will consequently reduce the received power again. To quantify this, it can be approximately assumed 50% of the received power is lost to harmonics. This will cause an additional 3 dB power loss.

One of the key implications of the power reduction discussed in the previous two paragraphs is its effect on SNR and therefore capacity. The asymptotic expression for MIMO capacity under the condition of high SNR and equal numbers of transmit and receive antennas is as follows:

$\begin{matrix} C \approx N \log_{2} (S N R) = 0.3 3 N SNR [dB] & Equation (2) \end{matrix}$

- where N is the number of transmit antennas. Assuming that TMM systems lose 6 dB of power compared to conventional systems, this would affect capacity. However, as seen from Equation (2), such a power loss does not alter the slope of the capacity line, which remains at N. The power loss only shifts the SNR. With an initial SNR of 30 dB, reduced to 24 dB (accounting for both switched and harmonic losses mentioned above), 25% more antennas are required to compensate for the 6 dB loss.

Although this may appear to require a substantial increase in the number of antennas, it is important to put it into perspective. For instance, in a massive MIMO array, using TMM and halving the inter-element spacing in each dimension would increase the number of antennas by a factor of 4, which far exceeds the impact of the 6 dB power reduction. To meet the requirement of 25% more antennas, reducing inter-element spacing by only 10% in each dimension would suffice. Thus, increased performance should still be achievable with the TMM approach, even when accounting for the power loss.

Although some preliminary results regarding the tradeoffs between TMM and conventional MIMO systems have been obtained, further analysis is necessary. First, the mutual coupling between every second element needs to be properly characterized and compared to adjacent elements to accurately assess the possible capacity gains. This involves using the full MIMO channel capacity model with mutual coupling effects. Additionally, a better understanding of time modulation effects is required. For instance, allowing some overlap between the open and on modulation states could reduce power loss, but it would also decrease the reduction in mutual coupling effects. This tradeoff needs to be thoroughly investigated, along with the inclusion of realistic switch harmonics in developing these tradeoffs. These issues will be addressed in the present invention.

(III): Development of TMM Antenna Design Guidelines

In principle TMM, antennas can be extremely close to each other with the effect of mutual coupling nearly completely removed. However, the patterns of the antennas of each of the elements will then also be nearly identical. This then leads to the question of whether channel correlation will be sufficiently small. For example, FIG. 5A and FIG. 5B show antenna patterns along x-z cut with both ports loaded. Specifically, FIG. 5A corresponds to port 1 (2.52 dBi) and FIG. 5B corresponds to port 2 (2.52 dBi). The patterns are different (with reflection symmetry due to symmetry in the antenna system) and will be different enough to exhibit low channel correlation. In FIG. 5A and FIG. 5B, the patterns for the antennas are shown when both antenna ports are loaded and fed normally. As can be seen the patterns are different. This difference will reduce channel correlation as it provides pattern diversity.

FIG. 6A and FIG. 6B show open circuit antenna patterns along x-z cut. Specifically, FIG. 6A correspond to port 1 pattern (1.77 dBi) with port 2 open circuited; and FIG. 6B corresponds to port 2 (1.77 dBi) pattern with port 1 open circuited. The patterns have fewer differences as compared to FIG. 5A and FIG. 5B and will exhibit more channel correlation.

The antenna patterns are provided for the case when the antennas are alternately time modulated, leaving the opposite port open. As observed, the patterns become more similar due to the removal of mutual coupling effects. However, the issue is that channel correlation becomes almost entirely dependent on the separation. Therefore, the channel correlation between two antennas separated by distance d will likely follow Jakes model which can be written, in terms of a Bessel function of the first kind, as:

$\begin{matrix} ρ (d) = J_{o} (kd) & Equation 3 \end{matrix}$

- where ρ(d) is the envelope correlation between the envelopes at port 1 and port 2, and k is wave number. It can be observed that separating the antennas will be necessary to reduce channel correlation, which directly conflicts with the TMM approach aimed at reducing antenna separation

To overcome these conflicts, it is therefore important that the patterns and/or polarization of adjacent antennas in TMM antenna design be different. If adjacent antennas are different, the correlation will not depend on the distance between antennas. It will depend on how different the patterns and/or polarization of the two antenna designs are. Therefore, an important design guideline for using the TMM concept is that adjacent antennas should have different patterns and/or polarization. In practice, this suggests that a massive MIMO array would have an interleaved structure. For mobile station antennas, it implies that adjacent antennas should be made as distinct as possible from one another.

In this task, the objective is therefore to identify pairs of antennas that can be easily interleaved while providing sufficient adjacent pattern diversity to reduce the effects of channel correlation. A fresh approach is needed to explore a wide range of antenna pair options for interleaving. Additionally, further investigation into the interleaved structure and pattern necessary to fully exploit this design concept is required.

(IV): Development of Reconfigurable TMM Antennas

Two aspects of TMM antenna design make it perfectly suited for use with reconfigurable antenna designs. The first is that for any two adjacent antennas only one is ever radiating. The second is that the patterns of the adjacent antennas need to be different. It is therefore possible to combine the two adjacent elements into a single reconfigurable antenna with two states. At any one instant only one of the antennas would be time modulated on and the patterns of the two reconfigurable states could be made very different. This approach would also have the significant added advantage of requiring less RF front ends. That is only one RF front end could be used to handle each reconfigurable antenna instead of two. The tradeoff would be that the frontends would need to have at least twice the bandwidth.

The TMM antenna design challenge then becomes the design of reconfigurable antennas and the development of novel techniques to allow them to be closely spaced. If the reconfigurable antennas can be as closely spaced as in conventional MIMO antenna designs then in a two-dimensional (2D) planar configuration it would be possible to achieve a four times increase in antenna density. Even if dual-polarization antenna arrays are considered, a double of antenna density can be achieved. This would easily overcome the reduction in capacity associated with power loss as discussed in relation to Equation (2).

The objective of this task is to identify reconfigurable antennas and configurations that are compatible with TMM. While the design of these antennas can leverage a significant body of prior research, an additional challenge arises: ensuring low mutual coupling between adjacent antennas. As a result, it will be necessary to design antennas in pairs, allowing mutual coupling between them to be understood and controlled. For example, in one configuration, low mutual coupling may be required, while in another, higher coupling could be acceptable, as the antennas will not operate in that state simultaneously. Additionally, the interleaving approach mentioned earlier could be applied, where adjacent reconfigurable antennas have different structures. Another critical consideration is how to implement the reconfiguration process. While PIN diodes have been used in previous designs, they can consume significant power when controlling hundreds of antennas. Therefore, the potential for using alternative reconfigurable elements, such as varactors, are also be explored in this task according to some embodiments.

(V): Investigation of Extensions to Transmit TMM Antenna Systems

In all the previous discussions, the target antenna designs have been assumed to be receiver systems. This is because the harmonics generated by the switching can be easily filtered out. Applying the proposed approach at the receiver also holds significant practical value, as the compact designs are particularly beneficial for mobile stations where space is limited. Consequently, this technique has the potential to enhance download speeds, addressing a common bottleneck in wireless communication.

Never-the-less extending this technique for transmission would also be very useful. To do so it is necessary to overcome the issue of the radiated harmonics. There has been significant effort in reducing harmonics as discussed in the review section. Special switching techniques have been developed with multiple phase and amplitude control.

The objective is to explore an alternative approach that has been proven effective for RF pulse shaping and orthogonal frequency division multiplexing (OFDM). This method involves systematic beam space decomposition along with electronically steerable parasitic array radiators (ESPAR) utilizing varactors. By smoothly varying the current at the varactors, virtually any pulse shape can be achieved. Applying this space-time approach to TMM antenna design for transmission systems could be a highly suitable solution. The pulse shaping technique based on varactors would allow for seamless reconfiguration from one beam shape to another, forming a reconfigurable element within the multiport antenna array, which would also reduce harmonics for transmission purposes.

In practical use, the proposed technical solution offers a general configuration for antennas. For example, a space-time multiport antenna system is provided with an antenna array, a plurality of feed ports, and a controller. The antenna array includes a plurality of antenna elements and the feed ports are connected to the antenna elements, respectively, so as to form a general base configuration. The controller can be applied to this general base configuration. The controller is connected to the antenna array via the feed ports and is configured to control the antenna elements for dynamically switching states of the antenna elements. Each pair of the adjacent antenna elements form a unit, and, in anyone of the units and at any given time, only one of the antenna elements is active and another one of the antenna elements is in a non-active state, which is advantageous to reduce mutual coupling of closely-spaced antennas significantly. The term “antenna array” mean it is formed by arranging the antenna elements into a N*M array, where N and M are positive integers greater than one.

FIG. 7A depicts a schematic drawing of a space-time multiport antenna system 100 according to some embodiments of the present invention; and FIG.