A5080815432- Microwave Engineering and Waveguides
- Radio Frequency Integrated Circuit Design
- Antenna Design and Optimization
- Antenna Design and Analysis
- Photonic and Optical Devices
- Millimeter-Wave Propagation and Modeling
- Advanced Photonic Communication Systems
- Structural Response to Dynamic Loads
- Advanced Power Amplifier Design
- Superconducting and THz Device Technology
- Optical Network Technologies
- Metamaterials and Metasurfaces Applications
- Automotive and Human Injury Biomechanics
- Semiconductor Quantum Structures and Devices
- High-Velocity Impact and Material Behavior
- Traumatic Ocular and Foreign Body Injuries
- Cellular and Composite Structures
- Terahertz technology and applications
- Advanced Wireless Communication Techniques
- Advanced MIMO Systems Optimization
- Neural Networks and Reservoir Computing
- Semiconductor materials and interfaces
- Advanced Wireless Communication Technologies
Princeton University
2020-2025
Indian Institute of Technology Madras
2023
Bilkent University
2016-2017
We present a deep neural network-based framework for designing multiband microstrip antennas given desired impedance matching spectrum. The approach enables design methodology that generates the antenna structures rapidly (under second) through an effective learning-enabled search of large space and eliminates need extensive domain knowledge design. is built on our innovations in tandem networks consisting two cascaded networks. Our are parameterized exponentially discrete variables...
This work describes deep-learning-enabled inverse design of multi-port electromagnetic (EM) structures co-designed with circuits that can enable the synthesis novel high-frequency on-chip passives and designer scattering parameters in a rapid automated fashion. The EM for typically starts from pre-selected topology unit functional elements are subsequently optimized desired through time-consuming parameter sweeps, ad hoc optimization algorithms, or prior expert experience. In space all...
Millimeter-wave and Terahertz communications consist of complex analog mixed-signal transceivers where overall system performance is often limited by the weakest performing subsystem. While integrated circuits have significantly advanced, future 5G 6G transceiver design could be accelerated including artificial intelligence. In this combination, circuit operation would harness machine learning to identify, characterize, act upon variations anomalies in performance. Focusing on 6G, paper...
Deep learning and artificial intelligence, in general, is advancing scientific discovery technological inventions through its ability to extract inherently hidden features map it output a highly complex multi-dimensional space. Synthesis of electromagnetic (EM) structures with nearly arbitrary desired functional properties such an example high dimensional optimization In this article, we employ deep convolutional neural network (CNN) allow robust rapid prediction scattering planar on chip....
Microwave communications have witnessed an incipient proliferation of multi-antenna and opportunistic technologies in the wake ever-growing demand for spectrum resources, while facing increasingly difficult network management over widespread channel interference heterogeneous wireless broadcasting. Radio frequency (RF) blind source separation (BSS) is a powerful technique demixing mixtures unknown signals with minimal assumptions, but relies on dependent RF electronics prior knowledge target...
In this paper, we introduce inverse design of nearly arbitrary planar antenna structures with a deep convolutional neural network (CNN) modeling that allows rapid and accurate prediction performance (scattering parameters radiation patterns). Quite distinct from prior efforts ML-based antennas fixed template geometries finite degrees freedom, approach generalizing to opens up new space for properties beyond what can be achieved optimized library. By eliminating complex time consuming...
This paper includes results of the blast tests which were performed with aim comparing energy absorption and protection efficiency protective boots different sole configurations. Tests on a mechanical leg model vestured boots. Load three axis acceleration values measured during to determine boot samples. Herewith, it was understood that merely monolithic composite layer used in does not provide protection, whereas compressible metallic honeycomb material-based sandwich composites demonstrate...
Millimeter-wave and terahertz integrated circuits chips are expected to serve as the backbone for future wireless networks high resolution sensing. However, design of these can be quite complex, requiring years human expertise, careful tailoring hand crafted circuit topologies co-design with parameterized pre-selected templates electromagnetic structures. These structures (radiative non-radiative, single-port multi-ports) subsequently optimized through ad-hoc methods parameter sweeps. Such...
This work presents a compact, concurrent dual-band (36/64 GHz) linear power amplifier in 90 nm SiGe technology. To simultaneously overcome the gain and limitation of conventional common-emitter (CE) silicon device at mm-wave, stacked common-base (SCB) topology is studied utilized as PA cell. A highly efficient output network enables optimum load-line matching for two widely separated frequencies. The proposed achieves 20/18.8 dBm <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML"...
To enable a software-defined spectrally agile phased array operation across multiple bands from 30-100 GHz, we present 30-88 GHz broadband IQ vector modulator phase shifter in 90 nm SiGe. This was enabled by 1) A single-end to differential quadrature generator consisting of compact Marchand balun and 90°hybrid-based bandwidth extension network; 2) common-base (CB) stage as the input buffer VGA. Across 58 bandwidth, maximum error 5-bit is below half LSB resolution while achieving rms...
Joint communication and sensing utilizing wide bandwidth additional spectral bands within the 30–100 GHz range presents exciting opportunities for 6G networks. It enables improved spectrum utilization enhanced environmental awareness. However, achieving frequency agility in a universal array interface that seamlessly operates across licensed, unlicensed, shared poses significant challenges. This article addresses this challenge by presenting crucial component, specifically architecture of an...
With the evolving spectral allocations across bands ranging between 24–100 GHz, future mmWave front-ends will need to address multiple band allowing agility and spectrum sharing capabilities for high capacity robust links. Such characteristics are extremely challenging achieve in transmitter particularly at power amplifiers. This paper presents a broadband efficient PA with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML"...
High frequency wireless systems operating in the high mmWave and sub-THz frequencies can enable new applications communication, sensing imaging, if they operate with low latency resource-constrained environments. In particular, for one-to-many network nodes where receivers be highly energy-limited, processing traditional spectrally-efficient signals that utilize both amplitude phase modulation (e.g., QPSK, 16/64-QAM etc.) places a significant power burden on high-fidelity synthesis, Tx-to-Rx...
In this paper, we introduce novel machine learning based techniques to design multi-band microstrip antennas as per user specifications over a broad range of frequencies. The approach involves the and training neural network for approximating electromagnetic simulations antennas, so-called 'forward' problem. Here, antenna is parameterized in terms checker-board pattern metallic sub-patches. Additionally, second 'tandem' also designed, which takes specification desired return-loss spectrum...
Future mmWave and THz wireless chip-scale systems need to be incorporate complex functionalities including ability operate over multiple spectral bands in an agile fashion spread across 30-100+ GHz, support spectrum sharing concurrent multi-band transmission, allow joint sensing communication. Design methodologies enable such challenging features typically suffer from tradeoffs energy efficiency, bandwidth, linearity reconfigurability. These trade-offs arise domains, primarily the fact that...
<p>We present a deep neural network-based framework for designing multi-band microstrip antennas given desired impedance matching spectrum. The approach enables design methodology that generates the antenna structures rapidly (under second) through an effective learning- enabled search of large space and eliminates need extensive domain knowledge design. is built on our innovations in tandem networks consisting two cascaded networks. Our are parameterized exponentially discrete...
<p>We present a deep neural network-based framework for designing multi-band microstrip antennas given desired impedance matching spectrum. The approach enables design methodology that generates the antenna structures rapidly (under second) through an effective learning- enabled search of large space and eliminates need extensive domain knowledge design. is built on our innovations in tandem networks consisting two cascaded networks. Our are parameterized exponentially discrete...
<p>We present a deep neural network-based framework for designing multi-band microstrip antennas given desired impedance matching spectrum. The approach enables design methodology that generates the antenna structures rapidly (under second) through an effective learning- enabled search of large space and eliminates need extensive domain knowledge design. is built on our innovations in tandem networks consisting two cascaded networks. Our are parameterized exponentially discrete...
Joint sensing and communication utilizing the large bandwidths spectral bands distributed across $30-100 \mathrm{GHz}$ provide new opportunities for 6G networks more optimal spectrum resource utilization, environmental awareness. However, adaptive operation agility needed in such a universal array interface capable of operating all licensed/unlicensed/shared $(30-100 \mathrm{GHz})$ intelligent is extremely challenging. In this paper, we focus on an element array, particularly ultrawideband...