An AI-driven intelligent design method for planar antennas is proposed by fusing multi-source structural information and electromagnetic prior knowledge. In the performance prediction stage, an Electromagnetics-Informed Multimodal Neural Network (EM-IMNN) is constructed by jointly modeling electromagnetic field distribution, structural parameters, and geometric topology, which improves the mapping accuracy between antenna structure and performance. In the structure generation stage, the basic electromagnetic principles of antennas, such as wavelength constrains and impedance matching requirements, are transformed into dozens of topology-guided patterns. These patterns are then fed into a Generative Adversarial Network (GAN) for pre-learning, thus effectively suppressing invalid structures and enhancing design efficiency. The EM-IMNN significantly alleviates overfitting and reduces the regression error by 90%. The GAN-based generator achieves a structure validity rate of nearly 100%. The proposed AI design agent is verified in dual-band, narrowband, and wideband scenarios. Results show that the design time is within 3 seconds, and the design target achievement rates reach 100%, 100%, and 96%, respectively, demonstrating high efficiency and practical value.

This letter presents an ultrawideband (UWB) circularly polarized (CP) monopole antenna with a defected ground structure, consisting of a monopole and a coupled patch with opposite openings, different sizes, the same rotational senses. characteristic mode analysis (CMA) is used to investigate CP generation mechanism, providing physical insight into different resonant modes and field distributions. Measured results exhibit a relative impedance bandwidth of 138.03% (3.30 GHz to 18.00 GHz) and a relative 3 dB axial ratio bandwidth of 123.60% (4.25 GHz to 18.00 GHz). The maximum of CP realized gain at the main beam direction is 8.32 dBic at 15 GHz. The overall dimension of the proposed antenna is 0.44 lambda(L) & times; 0.50 lambda(L) & times; 0.0165 lambda(L), where lambda(L) is the free-space wavelength at the lowest operating frequency. This antenna offers considerable application value in unmanned aerial vehicle detection, wireless sensor signal transmission, for its UWB CP radiation and high gain.

The power amplifier is one of the key devices in the RF front end, directly determining the performance of the transmission link. There are many indicators for power amplifiers, but in most scenarios, output power, gain, bandwidth, and efficiency are the four most important indicators. Therefore, in general, the figure of merit (FoM) is usually used to evaluate the comprehensive performance of a power amplifier. This article designs a 77 GHz wideband power amplifier with high FoM. The amplification circuit adopts a multi-frequency point superposition design approach to widen the bandwidth of the power amplifier, utilizes weakly coupled transformers to improve inter-stage matching, and employs a differential structure with neutralizing capacitors to enhance gain and output power. Based on the TSMC 65nm process, the circuit design and tape-out were carried out. Test results indicate that the power amplifier with the aforementioned circuit structure has a FoM value of 94.2. The peak gain is 30dB at 77 GHz frequency point, with a 3dB gain bandwidth ranging from 69.0 GHz to 81.0 GHz. Within the 3dB gain bandwidth, the saturated output power measured is greater than 13 dBm, and the output 1 dB compression point is greater than 10 dBm, The Power Added Efficiency is 16% at 77 GHz frequency point.

Traditional microstrip delay lines require sacrificing circuit area to achieve longer time delays, which is highly unfavorable for miniaturization of the electronic equipment, especially in the microwave low-frequency range. To increase the delay time of planar circuits, this letter proposes a combined structure of the open-circuit branch and parallel grounding strips. This combined structure enhances slow-wave transmission effects and reduces coupled eddy current effects. Thereby, the delay time is increased. Within the frequency range of 1.65 GHz to 2.15 GHz, for a single delay unit, the maximum group delay time can reach 1.3 ns, with an average transmission loss 0.5 dB. Moreover, this basic delay unit is also very easy to cascade to achieve longer delay times. The delay time of the four-unit cascaded delay line maximum reaches 7.4 ns, with an average transmission loss less than 2.6 dB.

Traditional RF front-end link typically arranges microwave devices with single functions sequentially in space dimension to process microwave signals, which becomes one of the bottlenecks hindering the softwarization of microwave systems and is particularly unfavorable for multi-scenario applications in radar. To solve the problem, here we propose a “time-for-space” microwave system architecture for multi-scene applications of radar. Dynamic definition of device functions through software enables the formation of sequential loops in the time domain for multiple functions, facilitating the processing of microwave signals. This architecture not only reduces the size of circuits but also, more importantly, allows for the software-defined configuration of the RF link according to application requirements, enabling the radar systems to have more flexible and diverse application scenarios. A redefinable 5-in-1 microwave device is designed and used for principle validation, theoretically capable of realizing over 55 = 3125 microwave links. By software-defining and pulse logic timing designing, we present three demonstration applications: microwave imaging, wireless communications, and radar detection, which show excellent measurement results: the detection positioning error is less than 5%, the imaging resolution is 0.5λ, and the communication bit error rate is less than 5%.

Based on the 55nm CMOS process, an LDO circuit with high PSRR in all frequency bands is designed to power the RF transceiver. An Active Low-Pass Filter (ALPF) is used to reduce the influence of the reference voltage on the PSRR of the LDO and improve the PSRR of the LDO in the mid-band band. Considering the problem of slow power-on charging of a large capacitor in the active low-pass filter, a Charging Circuit (CC) is designed to reduce the power-up time of the circuit. A compensation circuit is added to reduce the effect of the zero point on the circuit caused by the ESR resistance on the off-chip capacitor; The circuit is powered by 2.0V voltage, under the load of 60mA, the PSRR is less than −44dB in the band of from 1Hz to 1 GHz, the power-on time of the supply is $10\mu \mathrm{s}$, the power-on time of the LDO is $72\mu s$, and the static power consumption of the whole LDO is 270µA. Besides, the LDO has good transient response characteristics.

To realize "FPGA-like" microwave components and enhance the reconfiguration ability of microwave links, this article proposes a redefinable microwave component that breaks the limitations of conventional microwave components with single functionality. The redefinable microwave components adopt a "combining tangram" design approach to identify the "greatest common divisor" among several microwave functional structures. The patch + gap structure is used as the multiplex unit, and the RF switches are used to control the current flow direction, forming different electromagnetic distributions, thus achieving dynamic reconfiguration of microwave functions. The resonant performance is adjusted using variable capacitors, enabling the reconfiguration of the performance. The test results of the sample indicate that the microwave passive component can achieve software-defined antenna, filtering, and coupling functions, with adjustable center frequencies for all three functions. Compared with commonly used antenna, filter, or coupler, the performance metrics of the redefinable microwave components have not decreased. Furthermore, the component has also been applied for practical validation in radar systems. The demonstration results show that the redefinable microwave component significantly improves the reconfiguration performance of the microwave link, enabling the microwave system to meet the requirements of multiple application scenarios such as imaging, communication, and distance measurement.

Ingestible electronics have the capacity to transform our ability to effectively diagnose and potentially treat a broad set of conditions. Current applications could be significantly enhanced by addressing poor electrode-tissue contact, lack of navigation, short dwell time, and limited battery life. Here we report the development of an ingestible, battery-free, and tissue-adhering robotic interface (IngRI) for non-invasive and chronic electrostimulation of the gut, which addresses challenges associated with contact, navigation, retention, and powering (C-N-R-P) faced by existing ingestibles. We show that near-field inductive coupling operating near 13.56 MHz was sufficient to power and modulate the IngRI to deliver therapeutically relevant electrostimulation, which can be further enhanced by a bio-inspired, hydrogel-enabled adhesive interface. In swine models, we demonstrated the electrical interaction of IngRI with the gastric mucosa by recording conductive signaling from the subcutaneous space. We further observed changes in plasma ghrelin levels, the "hunger hormone," while IngRI was activated in vivo, demonstrating its clinical potential in regulating appetite and treating other endocrine conditions. The results of this study suggest that concepts inspired by soft and wireless skin-interfacing electronic devices can be applied to ingestible electronics with potential clinical applications for evaluating and treating gastrointestinal conditions. Existing ingestible electronics suffer from poor tissue contact, short dwell time, and limited battery life. Here, the authors report an ingestible, battery- free, and tissue-adhering robotic interface for non-invasive and chronic electrostimulation of the gut.

A soft and tunable microwave bandpass Chebyshev microstrip end-coupled filter is demonstrated on a polydimethylsiloxane substrate based on high electrically conductive composites made of silver nanoparticle-covered flakes, even with large strain. To achieve a precise design, key microwave parameters such as the impedance, S-parameters, and transmission matrix are defined as a function of strain. This strain vector polar coordinate is used to represent the relationship between the type and the performance of a filter. Then, the most important Chebyshev or quasi-Chebyshev type of filter can be accurately realized by looking up its polar coordinates to find the corresponding strain magnitude and direction of stretching. An end-coupled Chebyshev microstrip bandpass filter with a center frequency of 12.8 GHz and bandwidth of 22% was designed and fabricated for demonstration. Experimentally, at a lateral elongation of 8% and 10.7%, the center frequency decreases by 0.61 and 0.88 dB, respectively, matching the simulation results very well. One potential application for this tunable filter is a front-end module mounted on the biomimetic robot driven by its mechanic movements.

Aiming at realizing a reconfigurable microwave passive component programmable by software, whose counterpart in the digital domain is an field programmable gate array (FPGA), this work proposes a novel programmable passive microwave component. The component consists of basic functional units, transmission units and microwave switches, denoted as the basic building blocks. The states of the switches are controlled by software, resulting in different combination of the building blocks in different topologies where the distribution of the electromagnetic (EM) field is altered, thus different functions of the component. More specifically, the design is based on a planar micro-strip (MS) patch circuit. A resonance patch, a slot, a coupling MS line are used as the basic functional units, impedance transformation and MS transmission lines are used as the basic transmission units, which are connected by microwave p-i-n diode switches to form different topologies that can work as an antenna, a filter, a splitter, or a coupler with an FPGA acting as the control circuit. A prototype component is fabricated and measured with the following results. The antenna central frequency is 3.22 GHz, with realized gain 6.2 dBi. The passband of the filter is 3.35–3.53 GHz (a fractional bandwidth of 6.3%) with an insertion loss (IL) $\sim $ 2.7 dB. The splitter has a working band of 2.4–3.5 GHz, a 1.6 dB IL, a ± 0.4 dB amplitude unbalance and a ±3.4 ° phase unbalance. The coupler works at 2.5–5 GHz, whose IL $\sim $ 1.7 dB, coupling coefficient $\sim $ 30±1.4 dB. The four functions can be switched at will among each other by software control.

In this paper, a wireless module is presented using a high-efficiency and low-area current-mode DC-DC buck converter with an on-chip current sensor for feedback control. The DC-DC converter was fabricated using a standard 65 nm CMOS process and has a 1.1 mm 2 area. Its output voltage drops to 1.2 V from 2.5 V-3.6 V input, with an output ripple voltage of about 10 mV. To decrease the size of the external device while retaining high efficiency, a 2 MHz switching frequency and 2.2 μH inductance were chosen. Power efficiency exceeds 80% for output current ranging from 50 to 850 mA.

This paper introduces the structure and fabrication method of a bidirectional self-rolled-up inductor based on the self-rolled-up membrane (S-RuM) platform. Negative mutual inductance is produced by the coupling of two conductors with opposite current vectors, through by controlling the curling direction of the inductor's conductive strips and the length of the connecting lines, mutual inductance cancellation can be substantially reduced. Taking the 6-cell 10-turns inductor as an example, the inductance of the new bidirectional structure is increased by approximately 40%, achieving excellent electrical performance, and demonstrating potential for industrial production and practical applications.

High performance microstrip linear array antenna is demonstrated in this article through the optimization of the surrounding artificial magnetic conductor (AMC) units, which effectively improve the gain as well as the backward radiation. The in-plane design of the AMC units also achieves low-profile characteristic of the antenna structure, and therefore significantly simplify the fabrication processing. Systematic study of the AMC units, including the electromagnetic modeling, fabrication, measurement, and architecture reoptimization based on the performance discrepancy analysis, is implemented carefully. The best measured data of the demonstrative sample antenna show 14.5 dBi maximum gain which is enhanced by similar to 18% over the entire operating frequency band from 21.93 to 25 GHz compared to that of the controlled sample based on conventional design.

In this letter, a flexible three-dimensional (3-D) phase adjusting unit (PAU) lens is designed and loaded in front of a conventional Vivaldi antenna that is able to reconfigure with three typical statuses called planar, inward, and outward by physically bending. The distribution of energy and phase in the aperture of the antenna is therefore altered on demand for the enhancement of antenna gain and the elimination of radiation pattern distortion. The Vivaldi antenna samples loaded with reconfigurable 3-D PAU in the aperture are designed for demonstration, and their performance is compared in the interested frequency range of 6-18 GHz. The measured data show that the antenna gain of the sample with 3-D PAU lens is increased by similar to 1-2.5 dBi within the range of 11-18 GHz. In the higher frequency band of 13-18 GHz, its mainlobe is not split or distorted due to the effective suppression of sidelobe by similar to 1-5.5 dBi. Thus, the practical working bandwidth is significantly extended by 71%, from approximately 6-13 to 6-18 GHz, demonstrating the proposed method improves the antenna gain and stabilizes the radiation pattern of Vivaldi antenna concurrently over a wide frequency band.

As the critical component in the wireless communication systems, RF patch antennas fabricated on elastomer materials have attracted widespread attention and have been developing rapidly [1,2]. In general, stretchable RF antennas could be used in two primary cases : 1. In the first case, the electrical characteristics of a stretchable antenna are sensitive to the stretching of the soft substrate, so that a characteristic signal is generated in real-time for the monitoring of soft substrates morphological changes; 2. In the second case, the electrical performance of a stretchable antenna is insensitive to the deformation of the carrier, so that the performance of the wireless communication system is not significantly impacted under mechanical strain. Therefore, according to the different requirements of the above application scenarios, it is important to select the appropriate antenna type, and to alter the fabrication processes for the antenna in order to fit the elastic substrate and conductive material properties accordingly. The performance of the antenna will then become related not only to the frequency domain, but also to the strain of the substrate [3]. A general, simple method for the design, fabrication and application of stretchable RF patch antennas is highly desired. We proposed a practical and simple processing flow as shown through Figures 1(a) to 1(i). During the metal thermal evaporation, the PDMS substrate is intentionally pre-strained (1%~10%) as shown in Figure 1(f) in order to form metal wrinkles on the surface when relaxed. Figure 1(j) and 1(k) compare the surface morphology of the Au/Ni metal layers under different post-strain conditions, which shows “fishing net” contains smaller crack domains under 130% strain. The gaps between domains changes minimally and are negligibly small compared to the operating wavelength of the antenna in the GHz range. RF signals are able to pass across the gaps between metal domains easily by large electrical field coupling. As a result, the conductivity of the Au/Ni metal bilayer fabricated in the proposed way is affected little by stretching the substrate up to 140% at RF frequencies. Within a reasonable range of strain variation from 110% to 140%, the measured PDMS thickness and resulting relative dielectric constant as a function of strain is shown in Figure 1(l), and the expected trend of dropping characteristic impedance of microstrip transmission line when the PDMS substrate is stretched along the transverse direction is shown in Figure 1(m). Compared to traditional antenna design principle represented by equation (1) and based on FEM modeling and an artificial neural network (ANN) optimization algorithm, equation (2) summarizes the condition for an acceptable design of a stretchable antenna. Based on the above design route, a transmit-receive system using the two stretchable antennas attached conformally to a human wrist is setup as shown in Figure 2(a). It can be found from the amplitude of the received signal v.s. frequency plot (Figure 2(b) and 2(c)) that, for the stretchable monopole antenna, the received signal is relatively stable no matter how the wrist moves, and for the stretchable microstrip patch antenna, the frequency for which reception changes from 6 GHz to 5.6 GHz when the wrist moves. The two examples reported here suggest that the proposed design and fabrication methods for stretchable antennas provide remarkable levels of conformability to deformable substrates. Figure.1 Fabrication and design route for stretchable RF patch antennas. (a)~(i) fabrication processing flow with pre-strain strategy for improved stretchability of conductive layer at high frequency. (j) ~ (k) surface morphology of conductive layer under different applied strain. (l) ~ (m) Measured and calculated PDMS substrate thickness, dielectric constant and impedance of microstrip line on top of it in terms of applied strain up to 40%. Equation (1) and (2) represent the design principle of antenna with/without considering strain domain s (f: frequency, er: overall acceptable error,δ: frequency point error). Figure.2 The demonstration of the two stretchable antennas. (a) The simple transmit-receive system. (b) The comparisons between the relaxed status and stretched status of the monopole antenna. (c) The comparisons between the relaxed status and stretched status of the microstrip patch antenna.

A W-band low noise amplifier is designed based on the 0.13 μm SiGe BiCMOS process. The two-stage fully differential structure consisting of Cascode is used. In order to achieve good input and output matching performance, transformer matching is adopted between the two stages. Electromagnetic simulation is performed using 3D electromagnetic field simulation software. The simulation results show that the input and output matched well in the 90-100 GHz band. At 94 GHz point, the small signal gain is 15.6 dB meanwhile the input 1 dB compression point power is -14 dBm. The noise figure is 6 dB. The chip size is only 0.48mm × 0.19 mm. This low noise amplifier has good microwave characteristic which is suitable for communication, radar, imaging and other fields.

A 0.32-THz 7x7 SiGe HBT incoherent detector array with both horizontal and vertical polarization capability is implemented in 0.13-mu m SiGe BiCMOS technology. The average responsivity of the detector is 2.7 and 2.49 kV/W from 316 to 324 GHz for the horizontal and vertical polarizations, respectively, with a corresponding noise equivalent power of 23.8 and 25.8 pW/root Hz at an IF of 78.125 kHz. The contrast of metallic objects with features below the diffraction limit is enhanced by utilizing this chip's polarization diversities.

This paper presents a structural reliable high-gain Vivaldi array antenna over 8–12 GHz for airborne reconnaissance and surveillance operations. The overall size of the array is $190\mathbf{mm}\times 200\mathbf{mm}\times 170\mathbf{mm}$. The measured minimum gain is 20.5 dBi and the measured VSWR is less than 1.8 over the entire X-band. In the harsh operating environment, the specific shape aluminum plate can ensure the antenna work for long hours.

A broadband high-gain planar-printed end-fire antenna is presented for microwave imaging application. The proposed antenna consists of a conventional Vivaldi antenna (CVA) with slot edges (SEs) and a planar phase compensation lens (PCL). Taking into account the actual phase error along E-plane direction at antenna aperture, a more precise and detailed analysis for the printed antenna with PCL is carried out. The PCL with specific layout of rectangular patches is utilised to enhance antenna gain especially at high frequencies, while the SE technique is employed to further improve directivity at low frequencies. The final design combining PCL and SE develops an expected high-flat-gain Vivaldi antenna over the entire operating range. The CVA and the proposed antenna are fabricated and measured. The measured results agree with the simulated ones well. The proposed antenna provides a high gain of 10-11.7dBi in the range, which corresponds to a gain increase of 0.9-3.2dBi compared to the CVA.

Gain of the conventional Vivaldi antenna (CVA) decreases at higher frequencies. A phase correcting lens (PCL) made up of nonresonant rectangular patch unit cells (UCs) is proposed to correct this problem. The layout of the UCs is specially designed taking into account the actual phase error at antenna aperture. The detailed analysis, design and discussion for the CVA with the proposed PCL are performed. The CVA, CVA with one-layer PCL (CVA-OPCL), and three-layer PCL (CVA-TPCL) are fabricated and measured. The measured results agree well with the simulated ones. The measured results show that the proposed CVA-TPCL provides a gain variation of 5.75-14.9 dBi in the wide operating range from 3.5 to 16 GHz, which leads to 1.1-6.1 dBi gain enhancement than that of the CVA. In addition, the E-plane half power beamwidths become narrowed significantly at higher frequencies. An average improvement of 4.54 dB is achieved in terms of the front-to-back ratio. The sidelobe levels, on average, are reduced by 4.1 dB and 2.8 dB for E-plane and H-plane, respectively.