Deb Chatterjee

Small-size wide-bandwidth microstrip patch antennas

In many applications in wireless communications, it is desirable to use microstrip patch antennas which are small in size but have bandwidths in excess of 10%. There are three common methods of reducing the patch size: (1) the use of microwave substrates with large permittivities; (2) the use of shorting walls; and (3) the use of shorting pins. When applied to the regular patch, the resulting small-size antenna usually has only a few percent bandwidth, not enough to meet the requirement of many applications. In this paper, we apply the three size reduction techniques to two wideband patch antennas - the U-slot patch and the L-probe patch. Simulation results using ENSEMBLE 6.0 are presented first, followed by some measured results.

The impact of mutual coupling on MIMO radar emissions

The effects of mutual coupling between antenna elements are considered with regard to the impact upon co-located MIMO radar emissions. Because this sensing scheme intentionally couples the spatial and fast-time (waveform) domains, it is shown that MIMO radar is sensitive to any electromagnetic mutual coupling effects that are not adequately characterized in the transmit array manifold. This sensitivity leads to mismatch that will degrade the radars sensitivity on receive.

Characteristic Mode Analysis of a Class of Empirical Design Techniques for Probe-Fed, U-Slot Microstrip Patch Antennas

In this paper, characteristic mode analysis (CMA) of three empirical design techniques for the probe-fed, symmetrically located, U-slot microstrip patch antenna, on a single-layer grounded substrate, is presented with supporting experimental data. The first method, resonant frequency (ResF), utilizes the existence of the four distinct ResFs, while the second one, dimensional invariance (DI), relies on the property of DI, for the design of the U-slot microstrip patch. In both these methods, the optimization of the probe location is necessary for further enhancement of the 10-dB return loss bandwidth. The third method, dimensionally invariant ResF, that optimally combines the features of the previous two is developed here and shown to yield better bandwidth performance with minimal or no probe location optimization, and hence is superior to the other two for rapid prototyping. CMA is carried out for critical parameters, such as substrate electrical thickness, slot width, probe radius, and feed location variations, to assess their dominant influence on the characteristics of the U-slot microstrip patch antenna.

A hybrid formulation for the probe-to-patch attachment-mode current for rectangular microstrip antennas

In this paper, we present a hybrid representation for the attachment-mode current existing on the suirface of a coaxially fed rectangular microstrip patch antenna. The hybrid representation consists of a one- or two-term residue anid an eigenfunction series which gives the attachment-mode curirent at all points on the patch surface, including the probe-to-patch junction. It is numerically demonstrated that the residue and eigenfunction series blend smoothly in the close vicinity of the Stokes region which includes the probe-to-patch junction. The Stokles region is a very narrow region over which the attachment-mode current changes rapidly. The residue and eigenfunction series are used outside and within this Stokes region, respectively. This hybrid representation can be used in either spectral- or space-domain techniques for a full-wave solution to the microstrip antenna problem. The residue series is obtained by using an equivalent contour integral representation of the infinite eigenfunction series, which subsequently reduces to the Watson transform under the assumption of a slightly lossy substrate. At or near the Stokes regions the residue series is inadequate, as it cannot model the rapidly varying junction currents there. Examination of the residue series shows that the attachment mode is a superposition of exponentially attenuated, leaky, surface-traveling modes. This hybrid representation is expected to provide the optimum trade-off between speed and accuracy for computationis involving electrically large finite arrays.

Analytical evaluation of Sommerfeld integral tails for layered-media Green's functions

Greens functions for layered media find applications in analysis of radar cross sections, radiation problems for array sources embedded within complex materials, prediction of electric fields from lightning, remote sensing, shipborne EMI/EMC issues and other various applications. Central to the predictions, for such applications, is the need for accurate evaluation of the appropriate Greens functions that contain Sommerfeld integrals. In this paper a method is proposed for direct, real-axis integration of Sommerfeld integrals, obviating the calculation of pole locations and their residue contributions, with the integral tail evaluated analytically in closed form. The computational efficiency of the proposed algorithm is illustrated here with preliminary results by analyzing the Gzx component of the Greens function for a HED inside a PEC-backed, double-layer dielectric media. Finally, the present method appears to be better amongst other similar approaches because it can better model the inhomogeneity of the constitutive electrical parameters in the vertical direction of the PEC-backed layered structure.

On reducing the patch size of U-slot and L-probe wideband patch antennas

The coaxially-fed U-slot rectangular patch antenna and the L-probe fed rectangular patch antenna are two recently developed single-layer single-patch wideband microstrip patch antennas. In both cases, a second resonance is introduced near the main patch resonance, either by the U-slot or by the L-probe. The U-slot or the L-probe also introduce a capacitance which counteracts the inductance of the coaxial feed, allowing for the use of thick substrates (0.08-0.1 /spl Lambda//sub 0/) where /spl Lambda//sub 0/ is the free space wavelength. Using foam substrates (with /spl epsiv//sub r//spl ap/1) the impedance bandwidths of these antennas operating in the fundamental mode are in the 30-40% range, with stable pattern and gain characteristics. These bandwidths are more than sufficient for most wireless communication applications. The resonant length of the fundamental mode is about half of the free space wavelength. For many applications, it is desirable to reduce the size of the patch to conserve real estate space. For this reason, there have been extensive investigations on patch size reduction techniques. One method uses microwave substrates with values of /spl epsiv//sub r/>1. Another method uses a shorting wall at the location of zero electric field so that the resonant length is halved, resulting in the quarter-wave patch. Yet another method uses a shorting pin near the feed. This introduces capacitive coupling to the patch resonance, thereby increasing the effective /spl epsiv//sub r/, and reducing the frequency, which means that, for a given resonant frequency, the patch size becomes smaller. In this paper, results of some of these investigations are presented.

Phase-integral formulation of the single-layer microstrip Green's function

The three-dimensional (3-D) Greens function for a continuous, linearly stratified planar media, backed by a PEC ground plane, can be expressed in terms of a single contour integral involving one-dimensional (1-D) greens function. The phase-integral and WKB methods are applied here to obtain the 1-D greens function in the direction of continuous stratification, i.e., the z-axis. It is shown that for the spatially local wavenumber, k(z) with one simple zero, the 1-D greens function contains Airy functions and hence need careful evaluation across the Stokes lines. The phase-integral or WKB approach avoids the computational complexities associated with discretization of the continuous media into several layers with piecewise continuous electrical parameters, and hence a single 3-D Greens function can be derived to calculate all fields for arbitrary source and receiver locations.

Source-proximity limitations in high-frequency code solvers for modeling conformal cylindrical arrays

The NECBSC high-frequency code solver is an efficient computational tool for solving electrically large problems. The features of the UTD-based algorithm in this solver is gainfully utilized to calculate the embedded element pattern of a 5×5 cylindrical dipole element array located before a PEC circular cylinder of electrical radius ka = 96.5. The central element of the dipole array which is kept at 0.257λ off the curved surface of the cylinder, is excited with other dipoles terminated in a matched load of 50Ω. Since NECBSC formally does not calculate the array mutual coupling, this is included by calculating the induced currents on the elements derived from a full-wave integral equation based solver, FEKO. Comparison against the available experimental data shows remarkably good agreement between NECBSC, FEKO in the lit regions. The FEKO and NECBSC yield different results in the deep shadow regions dominated by creeping waves, indicating the necessity of further investigations.

A new approach for improved evaluation of sommerfeld integral tails for PEC-terminated single layered media

Sommerfeld integrals [1] appearing in the Greens function for layered media pose significant challenges in computation of the tail part of the integrand when the observer and source locations have large lateral separations with the observer close to the interface(s) [2]-[5]. Analytical [3],[4] and novel numerical [5] methods have been utilized to circumvent such problems, and is the subject of this paper. The purpose of this investigation is to obtain improved analytical forms for closed-form evaluation of the Sommerfeld integral tail that are more general compared to [3],[4].

Investigations into some new integral forms of bessel functions with applications to scattering from cylinders

The mutual coupling between two antennas on a convex surface is, for example, solved by using the uniform theory of diffraction (UTD) for electrically large bodies. However when new generalizations of asymptotic solutions are developed, they are compared against the exact solution to the same canonical problem. This validation process, in general, indicates the necessity for accurate computation of the exact solution. The subject of this paper is the accurate computation of the radiation and scattering from circular cylinders with electrically large radii (ka rarr infin).