This paper presents the design of odd-order near-linear phase IIR differentiators which is implemented as a parallel connection of two IIR allpass filters. Although this structure requires a slightly larger number of multipliers compared to delay line-based implementations, it provides a significantly lower phase error. The proposed design procedure utilizes direct control of the differentiator’s magnitude and phase responses, which differs from previously reported methods that rely on phase approximation of the individual allpass subfilters. As a result, the obtained magnitude and phase errors both exhibit an equiripple behavior. Furthermore, by introducing appropriate weighting coefficients in the optimization process, quasi-equiripple solutions are obtained that significantly reduce approximation errors at lower frequencies. The time-domain behavior of the proposed differentiators is also investigated using several input signals and compared with available state-of-the-art solutions, demonstrating superior performance.
This paper presents an efficient design procedure for generalized optimum modified Chebyshev filtering functions with symmetrically located transmission zeros, making it particularly suitable for enhanced even-degree filter designs. Two primary challenges are addressed in the approximation of this low-pass filtering function. The first is the analytical determination of the positions of an arbitrary number of symmetrically distributed distinct transmission zeros that ensure an equiripple stopband insertion loss. This is achieved by solving a system of nonlinear equations using Newton’s method, building upon a closed-form expression for calculating the frequencies of the local extrema in the characteristic function. The second challenge involves extracting the rational polynomial representation of the generalized optimum modified Chebyshev filtering function from its transcendental form. Additionally, it is shown that only a prototype network with a single pair of transmission zeros can be transformed into the microwave domain using the Richards–Kuroda procedure. The validity and applicability of the proposed method are demonstrated through a numerical example of a sixth-degree filter featuring one pair of transmission zeros.
A column-drain readout analog front end optimized for rolling-shutter DEPFET detector systems is presented. The architecture targets high-frame-rate applications demanding low-noise current signal conditioning and direct interfacing with high-speed, low-power analog-to-digital converters (ADCs). Each readout channel comprises a transimpedance preamplifier followed by a switched-capacitor fully differential amplifier, providing the signal conditioning required for differential-input ADC integration.The design employs low-power and compact circuit topologies to meet the stringent constraints of high channel density and fine pixel pitch, achieving an optimized trade-off among power consumption, noise performance, bandwidth, and silicon area. Implemented in a 65-nm CMOS process, the front end supports an input current range of 2–10 μA and occupies a channel area of 50 × 50 μm2. The fully differential output architecture enables seamless integration with high-speed ADCs capable of sampling the detector signal within ns range acquisition window.Post-layout simulations, including extended temperature corners validate the design robustness. The finalized front end achieves approximately 105× amplification of the input current pulse, a signal bandwidth of 26MHz per channel, a power dissipation of 1.5 mW, and an input-referred current noise of 56nA demonstrating its suitability for next-generation DEPFET-based detector readout systems.
Reliable low-altitude unmanned aerial vehicle (UAV) relay communications require each UAV to select a next-hop forwarding action while maintaining a high packet delivery ratio (PDR) under coupled link-reliability, queue, energy, delay, and packet-loss constraints. This paper formulates the decision as a constrained long-term communication-utility maximization problem and proposes LAAHN-MAPPO-ITA, a QoS-safe learning-assisted forwarding framework for reliable UAV relay communications. A shared MAPPO actor with initial-temporal attention estimates local next-hop preferences from compact neighbor context, while end-to-end delivery feedback and congestion–energy-aware credit allocation strengthen the training signal for relays on a packet path. The learned actor-attention model is exported from Python and reconstructed inside NS-3 for packet-level execution. A deterministic QoS-safe wrapper enforces link reliability, queue availability, residual-energy, and forwarding-progress feasibility before a packet is transmitted. The framework is evaluated in controlled 3-D NS-3 packet-level scenarios with 30, 35, and 40 UAVs against AODV, K-MORP, EARVRT, DEAR, DLGR-2DQ, and MARL-IAB-A2C. Additional ablation, robustness, and deployment-overhead studies quantify the component contributions and characterize the operating envelope of the framework. LAAHN-MAPPO-ITA achieves the best average QoS rank in all three main scenarios, ranks first in all six metrics under 30 and 35 UAVs, and remains best overall under 40 UAVs.
In this paper, we present an investigation into an ultra-wideband printed antenna designed for airborne applications. The antenna exhibits an ultra-wide fractional bandwidth of approximately 133% across the 1000–5000 MHz frequency range, maintaining a return loss better than 10 dB. It features a near-toroidal radiation pattern shape with a gain variation between 0 and 2.5 dBi. The paper details the antenna design, its operating principles, numerical simulations, and experimental measurements. Furthermore, the radiation patterns are analyzed for scenarios where the antenna is mounted on both dielectric and conductive planes. The underlying conductive surface has a negligible impact on the antenna's radiation characteristics, making it highly suitable for airborne applications and integration onto aircraft hulls. The compact dimensions of 93 × 52 × 0.813 mm3 (or 0.31λmax × 0.17λmax) underscore the key advantages of the proposed design.
In silica fiber supercontinuum generation, obtaining both broad spectral bandwidth and high coherence is still challenging, particularly in the 1.55 μm telecommunications window. Here, a simplified four-ring silica photonic crystal fiber infiltrated with nitrobenzene is proposed and numerically investigated. Ultra-flat chromatic dispersion is achieved through optimization of a small set of structural parameters, allowing the generation of supercontinua in both the all-normal dispersion (ANDi) and anomalous dispersion (AnD) regimes. Using 60 fs and 180 fs pump pulses, the proposed ANDi and AnD designs generate 1.51- and 2.23-octave supercontinua, respectively, at peak powers of 130 W and 200 W. Spectral coherence is evaluated through 200 stochastic simulations under different noise conditions. The ANDi fiber maintains high coherence, with an average first order coherence of g121= 0.956 over the 1.3–1.8 μm wavelength range and coherence exceeding 0.95 across approximately 2.3 μm of bandwidth, indicating its potential for high resolution optical coherence tomography. Meanwhile, the extended spectral coverage achieved in the AnD regime provides access to the mid-infrared region relevant to molecular absorption spectroscopy. This work provides useful design guidelines for broadband, low power supercontinuum sources based on liquid-infiltrated simplified silica PCFs operating at the 1.55 μm telecommunications wavelength.
Due to their ability to exhibit adaptive behavior depending on the power level of electromagnetic waves, energy-selective surface (ESS) layers have become one of the most important structures in the last decade for protecting sensitive electronic systems. Since ESS provides transmission to low-power electromagnetic waves while acting as a protective barrier against high-power electromagnetic threats, they are important elements, particularly in defense and communication applications. In this study, a double-layered ESS was designed and fabricated both to increase antenna gain at low power levels and to provide protection potential against high-power microwave (HPM) or electromagnetic pulse (EMP) effects for a patch antenna. The antenna design was implemented on an FR-4 dielectric substrate with a rectangular patch and line-fed configuration to operate in the 2.45 GHz ISM (Industrial, Scientific, and Medical) band. The ESS layer was positioned as superstrates at a certain distance on the antenna to form a periodic array structure. The design and analysis results show that the ESS superstrate layer can increase antenna gain around 2.27 dBi (~68.66%) in the transmission band, while effectively protecting (22 dB of shielding effectiveness) the antenna under HPM conditions. Thus, the proposed ESS structure can simultaneously behave as both a superstrate layer that improves electromagnetic performance and a protective shield layer against HPM threats.
Radar signature manipulation materials, such as radar-absorbing structures, have attracted broad interest for regulating the electromagnetic scattering characteristics of complex targets. However, most existing functional materials rely on dielectric or composite systems, which can limit environmental adaptability, durability, mechanical strength, and conformal integration. Low-cost solutions capable of achieving broadband scattering control and radar image signature regulation remain insufficiently explored. In this paper, a conformal all-metal coding metasurface is proposed to realize electromagnetic scattering suppression and cooperative control of radar image signatures using metallic structures. Two all-metal unit cells with an approximately 180° reflection phase difference are designed to enable broadband phase modulation from 10 to 18 GHz. By optimizing the coding sequence, the metasurface redistributes backward scattering energy and achieves broadband radar scattering reduction. In the X-band, an 8 × 8 coding array is employed to manipulate radar image signatures and generate preset illusion features. Owing to its substrate-free architecture, the proposed metasurface can be integrated with impact-resistant skins through advanced 3D printing for monolithic fabrication. This work expands the design space of electromagnetic scattering control materials and offers a promising route toward multifunctional protective structures in complex electromagnetic environments.
This paper focuses on the analysis of the performance of Single-Input Single-Output (SISO) wireless communication channels that are subjected to Lomax fading. Although the Lomax distribution has been introduced as a finite-dimensional system in the context of severe fading conditions, the fundamental issue of system reliability under the effects of non-Gaussian noise and hardware impairments is still largely understudied. This paper presents an effective approach as we provide innovative methods for the error analysis of the aforementioned systems in the presence of Additive White Generalized Gaussian Noise (AWGGN) and an imperfect phase errors. In order to deal with the analysis and resulting design complexities, we use tight approximations and mathematical tools, such as Laguerre polynomial and etc. We perform detailed exact simulations to serve as a benchmark for our analytical results and to validate the developed closed-form expressions over a number of given parameters, including fading severity, modulation schemes, and noise types. A key contribution of the research is the determination of the system's robustness, which implies the way in which performance decreases under different noise and phase error environments. This analysis will therefore identify the noise types (Laplacian, Gamma, Gaussian) and phase error conditions for the different wireless systems and enable the design and performance prediction for these systems.
This paper presents a new design of wideband circularly polarized (CP) antenna array based on the self-phase-shifting feeding scheme. To illustrate the operation mechanism, a prototype of a feeding network with inherent 90° phase shifting is initially presented and investigated. Subsequently, taking the slot loop antenna as an example, the feeding scheme for a CP antenna is implemented by virtue of two different types of microstrip-to-slotline transitions. In this context, a single-layer and low-profile wideband 2 × 2 CP antenna array is designed, fabricated and measured. Experimental results exhibit that the proposed CP array has achieved good CP radiation performances with a 45.7% impedance bandwidth, a 38.8% 3-dB axial ratio (AR) bandwidth as well as a 35.2% 3-dB gain bandwidth simultaneously, noting that only 0.006λ0 of profile is required. The results validated the feasibility and advantages of the proposed self-phase-shifting feeding scheme in wideband CP antenna array application.
Dynamic spectrum sensing remains a critical bottleneck for sixth-generation (6G) cognitive radio networks, as conventional algorithms are susceptible to low signal-to-noise ratios (SNRs), while deep learning models demand extensive labeled data and incur substantial deployment overhead. This paper proposes a hardware-software collaborative spectrum-sensing framework integrating a lightweight one-dimensional convolutional neural network (1D-CNN), parameter-freezing transfer learning, and a universal software radio peripheral (USRP) platform. The model is pre-trained on RadioML 2016.10a and fine-tuned on 3968 measured USRP in-phase/quadrature frames. It is compared with a standard CNN, a long short-term memory network, and MobileNetV2-1D under matched- and cross-domain conditions. The proposed model achieves 95.14% validation accuracy under matched-domain conditions and 79.67% average peak balanced validation accuracy over five cross-domain runs. Although the standard CNN attains 82.31%, the proposed model requires 83.3% fewer parameters and achieves 61.0% lower inference latency. It also achieves a detection probability of 0.92 at 0 dB SNR, exceeding the IEEE 802.22 requirement. These results demonstrate a favorable trade-off among transfer accuracy, stability, and deployment efficiency.
A 12-bit 4-GS/s two-way time-interleaved (TI) pipelined analog-to-digital converter (ADC) with a sample-and-hold amplifier-less (SHA-less) front-end is described in this paper. To address performance degradation caused by time skew in the SHA-less architecture, a joint magnitude-direction aperture error calibration technique with an extended Nyquist zone is proposed. By leveraging the dual-channel interleaved characteristics to evaluate the error direction, this technique overcomes the inherent Nyquist-boundary limitations of input frequencies, thereby allowing a seamless merging of the first and second Nyquist zones. Combined with the convergence of error magnitude, this approach achieves background-blind calibration, eliminating the need for extra input frequency information. Designed in 28-nm CMOS technology, post-layout simulation results demonstrate that the proposed aperture error calibration effectively mitigates the redundancy range consumption caused by timing skew, reducing the first-stage residue spread from +237/−232 LSB to +179/−178 LSB. With a 4 GS/s sampling rate, the two-way TI-ADC achieves a signal-to-noise-and-distortion ratio (SNDR) of 60.4 dB and a spurious-free dynamic range (SFDR) of 80.6 dB at the Nyquist input frequency. The differential nonlinearity (DNL) and integral nonlinearity (INL) are bounded within +0.55/−0.54 LSB and + 0.83/−0.77 LSB, respectively. Consuming a power of 268.4 mW in total, the ADC Schreier figure-of-merit (FoMS) is 156.1 dB.
