Ying Wang

Neural Network Inverse Modeling and Applications to Microwave Filter Design

In this paper, systematic neural network modeling techniques are presented for microwave modeling and design using the concept of inverse modeling where the inputs to the inverse model are electrical parameters and outputs are geometrical parameters. Training the neural network inverse model directly may become difficult due to the nonuniqueness of the input-output relationship in the inverse model. We propose a new method to solve such a problem by detecting multivalued solutions in training data. The data containing multivalued solutions are divided into groups according to derivative information using a neural network forward model such that individual groups do not have the problem of multivalued solutions. Multiple inverse models are built based on divided data groups, and are then combined to form a complete model. A comprehensive modeling methodology is proposed, which includes direct inverse modeling, segmentation, derivative division, and model combining techniques. The methodology is applied to waveguide filter modeling and more accurate results are achieved compared to the direct neural network inverse modeling method. Full electromagnetic simulation and measurement results of Ku-band circular waveguide dual-mode pseudoelliptic bandpass filters are presented to demonstrate the efficiency of the proposed neural network inverse modeling methodology.

Predistortion technique for cross-coupled filters and its application to satellite communication systems

This paper presents a novel adaptive predistortion technique for general cross-coupled microwave/RF filters with improved insertion loss and group-delay equalization. The method enables many potential applications of an almost abandoned technique, and permits a lower Q implementation technology to emulate the performance of a higher Q filter. 10-4-4 filters were built and tested at C- and Ku-band to verify the validity of the new method. The impact to satellite communication channels was also analyzed. Another novel concept of over-predistortion was proposed and evaluated and should lead to significant improvement for applications such as satellite transponder input multiplexers, where insertion loss can be traded off for in-band flatness, mass, volume, and even overall system performance.

True Inline Cross-Coupled Coaxial Cavity Filters

In this paper, a novel true inline configuration for cross-coupled coaxial cavity filters is presented, which is characterized by a simple and compact structure, improved performance, and good tunability. Instead of using folded structures, dedicated coupling probes, or extra cavities, as required by conventional techniques, cross coupling is realized by changing the orientation of selected resonators. Sequential coupling between adjacent resonators and cross coupling between nonadjacent resonators are effectively controlled by introducing small metal plates at different locations. A six-pole bandpass filter with two transmission zeros was built and tested. The measurement and simulation results agree very well, demonstrating feasibility of the inline filter configuration. It enables compact design with improved resonator Q compared to conventional combline filters. Furthermore, cross coupling can be readily adjusted using tuning elements.

Synthesis and Beyond

This article outlined several optimization techniques applied in different steps of the synthesis of filters and multiplexers. The importance of the subject has been appreciated by more and more researchers and engineers in this community. As it is impossible to cover all technologies available, the objective here is to provoke a broader thinking beyond basic filter synthesis. Established computer-aided design methodology is a true blend of both synthesis and optimization. In most cases, to meet a set of well thought-out specifications, the science of filter synthesis and trade-off applies. In other cases, the optimization of design is an art which creates opportunities and new possibilities.

Enhanced Microwave Multiplexing Network

A new approach for designing a microwave multiplexing network is presented in order to improve channel performance. Through design and optimization of the connecting transmission line structure, the new approach ensures that an extra pole is formed in the passband of each channel. The channel filter order is therefore increased by one without additional resonators. Significant improvement in the performance of the multiplexer is achieved without the penalty of increased size or weight of the hardware, which is extremely important for applications in communication satellite. Simulation and measurement results are used to demonstrate the feasibility of such enhanced microwave multiplexing networks.

A 59–66 GHz Highly Stable Millimeter Wave Amplifier in 130 nm CMOS Technology

The design and fabrication of four-stage cascaded mm-wave low noise amplifiers (LNAs) in a 130 nm CMOS technology are presented. The simultaneous high stability factor and low noise figure are obtained using proper inductors in both gate and source of the transistor. Measured gain of 14.7 dB with a 7 GHz bandwidth has been achieved. The larger inductors are realized with microstrip lines to improve the performance of the LNA and minimize the circuit size.

SIW Hybrid Feeding Network-Integrated 2-D DRA Array: Simulations and Experiments

Simulation and experimental results of substrate integrated waveguide (SIW)-integrated 2-D dielectric resonator antenna (DRA) array operating at millimeter-wave frequency band are presented. The antenna array consists of 64 elements arranged in 8 ×8. Hybrid feeding network combining series-fed and corporate-fed is used to provide a uniform power distribution over the entire aperture of the array. The antenna array is fabricated and tested. The measured reflection coefficient shows an impedance bandwidth ( S11 ≤ - 10 dB) of 2 GHz and a measured gain of 21.60 dB over the operating frequency band 35-37 GHz. A very good correlation between the measured results and simulated results is observed.

Hybrid Models for Effective Design and Optimization of Large-Scale Multiplexing Networks

In this paper, hybrid models for the design and optimization of large-scale multiplexers are presented. First, analysis based on the low-pass prototype filter circuit model proves that the first cavity of a channel filter next to the manifold is critical in the design of multiplexers. A hybrid model is proposed, combining a circuit model and the electromagnetic (EM) analysis of the coupling junction close to the manifold. The fact that only a single junction requires EM simulation facilitates the introduction of neural networks to multiplexer design for the first time. Another format of the hybrid model is then built by replacing the EM simulation with a neural model for the output coupling to the first resonator. The efficiency of the method is demonstrated using a six-channel multiplexer and a 13-channel multiplexer. Responses predicted by the hybrid models agree well with full EM simulations and measurement results. The proposed hybrid models are proven to be able to provide a high degree of accuracy in both in-band and out-of-band frequency ranges and fast, enabling iterative optimization involving all design parameters for large-scale multiplexers.

An efficient method to design optimum millimeter wave low noise amplifier

Recent advancement in CMOS technology has created new opportunities to design low cost receivers for conventional and emerging millimeter wave applications. Short range communication, wireless personal area network (WPAN), car radar and mmw imaging are some examples of millimeter wave transceiver application which have been in the focus of many researchers [1–5].

WAVE ANALYSIS FOR INDUCTIVELY MATCHED MILLIMETER WAVE AMPLIFIER DESIGN

A new design approach based on wave analysis has been implemented in order to derive voltage gain, center frequency and bandwidth of millimeter wave amplifler using parameters of transmission lines (TL). The derived formula allow one to design high frequency amplifler with predetermined bandwidth and centner frequency. It has been shown that in the case of lossy TL or at high frequency, circuit theory cannot predict the amplifler gain behavior while presented wave theory can accurately predict the frequency response of the amplifler in both low and high frequency ranges.