The antenna pattern is an essential part of the design of RF systems and affects the performance and capabilities for many applications in communications, radar, and sensing. There are many applications which require specified antenna patterns with specific directivity, beamwidth, and sidelobe level (SLL). Single-element antennas usually have simple and specific patterns which are difficult to be shaped to meet more complicated pattern requirements. For instance, the popular parabolic reflector antenna uses a reflector which can be shaped to produce a desired radiation pattern with high directivity. However, it has a large structure and can only produce single fixed-beam patterns. On the other hand, array antennas consist of multiple antenna elements which together can be used to synthesize antenna patterns with narrower beams and lower sidelobes as compared to single-element antennas. More specifically, many applications which require high directivity, narrow beam patterns with low sidelobes include: (1) radars, which often use a narrow beam to detect targets for achieving a better angular resolution, higher signal-to-noise ratio (SNR), and low sidelobes to avoid ambiguity coming from signal returns from other directions; (2) modern cellular phone base stations which employ specially shaped beam patterns to provide uniform signal strength with the coverage area while minimizing radiation into the sky; (3) newest satellite communications/broadcasting systems which adopt spotlight beams to cover specific zones while reducing interference into neighboring areas for enhanced security and SNR.
The first array antennas for producing shaped directive beam patterns were introduced during World War II for early radar systems using an array of dipole elements. The disadvantages of such a dipole array were that the dipole elements were large 3D objects requiring manual labor to produce and the design was difficult to use for higher frequency such as for X band or higher. The Yagi-Uda antenna, an end-fired dipole array, uses parasitic dipoles to produce high gain and have also been widely used for receiving broadcast and communication signals in the VHF and UHF frequency range. The later generation of radar systems used slotted waveguides, with the arrays of slots cut into the sides of waveguide walls. These slots are sequentially excited with proper phase and magnitude as the electromagnetic waves propagate along the waveguide, constituting a type of series-fed array. The main disadvantage of the slotted-waveguide antenna lies in their 3D structure, which is not easy to fabricate accurately, especially at frequencies above 18 GHz.
In modern applications, printed circuit board (PCB) antennas using microstrip patch elements are preferred since they can provide small, lightweight, low-profile planar designs which are easily manufactured using modern PCB fabrication. PCB antennas are also favored for their ease in interfacing the antennas with modern monolithic microwave integrated circuit (MMIC) devices used in RF systems. Phased array antennas are used in many modern applications which require actively scanning the beam to different angles. In earlier phased array systems, a single high-power source such as a Klystron or Traveling Wave Tube (TWT) produced an RF signal with MW or GW of total power distributed over all array elements. This approach requires careful management of the high-voltage and thermal issues. With the advances in small high-power solid-state amplifiers, modern phased arrays replace the single high-power source and distribution architecture with a small individual transmit/receive (T/R) module behind every array element. This makes PCB-based array antennas even more desirable for its ease in interface between array elements and T/R modules.
While phased arrays using T/R modules are mainly used for military and defense high power radar applications, many commercial communication systems applications need small, lightweight, low-profile, and low-cost high-gain antennas with specified radiation patterns with narrow fixed beams and low sidelobes. Some of these examples include (1) automobile radar systems which require broad elevation beamwidth and narrow azimuth beamwidth, with low sidelobe patterns for detecting people and objects, (2) base station antennas on cellular towers which require a special shaped pattern for providing uniform signal strength over the coverage area at different distances from the tower, and (3) the user-segment of the satellite communications systems which requires specific antenna gain levels and patterns that meet certain FCC gain-pattern envelope masks for avoiding interference to neighboring satellites. For these applications, the antenna array must be designed to produce a specific radiation pattern by controlling the amplitude and phase distribution across the array aperture which is related to the far-field pattern via the Fourier transform relationship. Without using T/R modules on each array element as in the more expensive military and defense phased arrays, a sophisticated feeding network would be required to properly distribute power and phase to all array elements according to the desired aperture distribution.
Most conventional feeding networks adopt corporate feeds which split the signal from the source using a network of power dividers to distribute power to the array elements. Corporate and parallel feeds are favored for straight-forward implementation of the approximate aperture amplitude distribution by using unequal power dividers and mismatch lines. However, such power dividers cannot realize the aperture amplitude distribution with a high accuracy, and thus cannot achieve accurate patterns and very low sidelobe level (SLL). Furthermore, the corporate feed network approach; (1) requires a large number of power dividers for a large number of elements, (2) suffers from increased ohmic and dielectric losses from long feed transmission lines, and (3) produces undesired spurious radiation due to mismatches accumulated from all power dividers and final mismatch lines. Alternatively, the series-fed architecture which uses a common feedline running through elements to distribute power sequentially minimizes feed line length and eliminates power dividers, thus minimizing power loss and spurious radiation. Therefore, the series-fed array design approach is very popular at very high frequency ranges such as for 68-82 GHz mm-wave radars for vehicles.
In the conventional series-fed array designs, the aperture amplitude distribution for a desired pattern is implemented by varying the width of the patch array elements, which varies the amount of power radiated from each element. The drawback to this conventional design method is that the bandwidth and pattern of individual elements also change with the patch width, which causes the final array pattern to deviate from the designed array pattern predicted from array factor multiplication. Another conventional series-fed design method is to vary the widths of the feedlines interconnecting patch elements. However, when using this method, the phase velocity along the series-fed array varies which then requires the patch elements to be adjusted to maintain resonance and for element spacings to be adjusted for all elements to be in phase. Furthermore, there is a limit to both the available range of coupling coefficients and accuracy in implementing the amplitude distribution due to limits in the line width and tolerance in manufacturing. Therefore, it remains a major challenge to design a series-fed array for specified patterns due to the lack of a systematic design methodology that can achieve streamlined and accurate implementation of specified amplitude distributions in a way that is simple and practical without requiring multiple design iterations.
This dissertation introduces a novel series-fed array design process for systematically implementing a desired aperture amplitude distribution for series-fed microstrip antenna arrays. In contrast to the conventional series-fed arrays which vary amplitude coupling by tapering the width of patch elements or feedlines, the different amount of coupling at each array element is achieved by varying the gap between the main feedline and a novel T-coupler which can achieve a wide range of coupling coefficients from -5 dB to less than -20 dB. The coupling coefficients from the normalized aperture amplitude distribution are then used to obtain the corresponding coupling gap for each array element. The proposed streamlined series-fed array taper design process allows effective control of the shape of the far-field pattern within only one design iteration. Since the taper amplitude distribution is controlled only by the coupler between the main transmission line and array element, all elements in the array can be identical and there is extreme flexibility in the choice of element types.
The dissertation includes several design examples to demonstrate the efficiency, effectiveness, and versatility of the proposed design methodology. So far, a design example was completed for an X-band dual-channel array with a customer-specified Taylor aperture taper in the azimuth plane for use in Synthetic-Aperture Radar (SAR) and Ground Moving Target Indication (GMTI) in an airborne radar system.
The novel design method presented here will significantly advance the state-of-the-art in designing series-fed array antennas and enable new capabilities of series-fed array antennas. For instance, the phase between elements on the feedline are fixed in conventional series-fed arrays and so they are rarely used for beam scanning. Using the proposed novel design method, it will be possible to place phase shifters between the couplers and elements to produce a beam-scanning series-fed array. Series-fed arrays also suffer from what is known as beam squint, which is the deviation of the main beam direction with change in frequency. This limits the operational bandwidth of series-fed arrays intended for patterns that are constant over frequency. A dispersive transmission line can be used to compensate the change in phase between elements over frequency and eliminate beam squint. This can also help to ensure a constant array pattern over frequency. Using the novel design method presented here and with the added capabilities of beam-scanning and wider operational bandwidth by eliminating beam squint, future series-fed array designs can be used for a wider range of applications. The novel design method presented here lays the foundation for new series-fed array designs that improve upon the current conventional designs and can add new capabilities for more applications in the future.