A multiple‐input–multiple‐output (MIMO) microstrip patch antenna array of four elements is proposed to cover the 5.8 GHz WLAN band. The characteristic mode analysis is carried out to design a defected ground structure that improves the MIMO performance without impacting negatively the radiation characteristics of the four‐element array. Simulations and measurements demonstrate that the implemented prototype reaches a maximum mutual coupling of −32 dB and a gain‐per‐element of 5.3 dB with good MIMO attributes, surpassing many other proposals in the antenna literature: maximum envelope correlation coefficient of 0.0001, diversity gain very close to 10 dB, capacity loss near to 0.0023 bps/Hz at the resonant frequency, multiplexing efficiency of 0.935, and a total active reflection coefficient weakly dependent on the excitation phases. The presented design is proposed for operation on the IEEE 802.11a/n standard.

A detailed statistical analysis of the total active reflection coefficient (TARC) is carried out in this paper for three 4-port MIMO antennas featuring different levels of isolation across its ports. This analysis is very useful to determine the most likely performance of a MIMO antenna in a real communications scenario. The TARC parameter is commonly evaluated for only several combinations of the random phase with which a signal reaches every input port of a MIMO antenna. By contrast, we have evaluated a million combinations to obtain the probability density function of the TARC, using frequency as its parameter. In this way, an expected value of the TARC is obtained for each frequency, as well as a confidence interval ($\Delta C{I}_{\mathrm{TARC}}$) where the TARC values occur with 90% probability. Additionally, we have introduced the term “TARC shadow”, a visual representation of the TARC as a function of the frequency where the probability function is projected into this 2D graphic with different colors to identify the most likely values of the TARC. To demonstrate these concepts, a full TARC evaluation was performed for three 4-port MIMO antennas with increasing isolation of 12.9 dB, 25.4 dB, and 37 dB between elements, and different values of the $S_{nn}$ and $S_{nm}$ parameters, with n and $m=$ 1 to 4. From this study, the importance of the isolation among ports and its comparison with the return losses becomes evident in achieving a MIMO antenna array insensitive to random phase variations occurring in the communication channel.

In this paper, a novel technique for reducing the dimensions of patch antennas and their arrays is presented. The procedure uses complementary split ring resonators (CSRRs) for increasing the slow-wave effect on the radiator, and therefore augmenting its electrical length, which implicates the modified antenna gets a lower physical dimension than a conventional one for a given frequency. Led by the use of CSRRs, the patches may be closer to each other, without affecting the side lobes magnitude either the directivity of the array compared to a conventional one. As an example, a 4 × 1 patch array was designed and its total area was 40% smaller compared to a conventional one, considering the same design conditions with and without CSRRs. The gain, directivity, and coupling of the prototype were not affected.

In this paper, an easy-to-program mathematical expression to compute the Envelope Correlation Coefficient (ECC) between two antennas by using their radiated electric fields at the far-field region is derived. The main advantage of the resulting algorithm is its higher speed to obtain results compared to those used in commercial software. This novel approach presents a discrete equation that can be employed after obtaining the electric field in spherical coordinates by computer simulations. Thus, another alternative arises to report this metric for low or high radiation efficiency antennas, instead of the less accurate S-parameter-based equation, which does not ensure high accuracy when low radiation efficiency antennas are under study. A comparison between the presented algorithm and the results given by commercial software is shown and discussed, in which is clearly observed that the results are greatly convergent, but with the advantage of achieving a lesser time computing period by the proposed method.

In this Comments, the authors assure that the equation (9) presented by the authors is not appropriate to compute the Total Active Reflection Coefficient, and present the correct expression.