arXiv subjects
Publications and source records attributed to Nikolaos Makris.
We demonstrate a centrally controlled dynamic switched-QKD network, with integrated PUF-based dynamic authentication for each QKD link. The performance of the dynamic switched-QKD network with real-time PUF-based authentication is analyzed.
We demonstrate a compact, high-isolation C-band optical ground terminal for satellite QKD, achieving more than 120 dB daylight background suppression and 135 dB crosstalk noise isolation. Successful QKD over an outdoor 100m FSO validate its feasibility for integration of satellite-QKD with urban fiber segments.
We analyze the performance of Time-bin Phase and Polarization based QKD systems on mixed 14Km underground and 16Km of aerial fiber using plug-and-play commercial QKD systems.
We present an approach to learning regular spatial transformations between image pairs in the context of medical image registration. Contrary to optimization-based registration techniques and many modern learning-based methods, we do not directly penalize transformation irregularities but instead promote transformation regularity via an inverse consistency penalty. We use a neural network to predict a map between a source and a target image as well as the map when swapping the source and target images. Different from existing approaches, we compose these two resulting maps and regularize deviations of the $\bf{Jacobian}$ of this composition from the identity matrix. This regularizer -- $\texttt{GradICON}$ -- results in much better convergence when training registration models compared to promoting inverse consistency of the composition of maps directly while retaining the desirable implicit regularization effects of the latter. We achieve state-of-the-art registration performance on a variety of real-world medical image datasets using a single set of hyperparameters and a single non-dataset-specific training protocol.
A MOSFET threshold voltage extraction method covering the entire range of drain-to-source voltage, from linear to saturation modes, is presented. Transconductance-to-current ratio is obtained from MOSFET transfer characteristics measured at low to high drain voltage. Based on the charge-based modeling approach, a near-constant value of threshold voltage is obtained over the whole range of drain voltage for ideal, long-channel MOSFETs. The method reveals a distinct increase of threshold voltage versus drain voltage for halo-implanted MOSFETs in the low drain voltage range. The method benefits from moderate inversion operation, where high-field effects, such as vertical field mobility reduction and series resistances, are minimal. The present method is applicable over the full range of drain voltage, is fully analytical, easy to be implemented, and provides more consistent results when compared to existing methods.
Variability of low frequency noise (LFN) in MOSFETs is bias-dependent. Moderate- to large-sized transistors commonly used in analog/RF applications show 1/f-like noise spectra, resulting from the superposition of random telegraph noise (RTN). Carrier number and mobility fluctuations are considered as the main causes of low frequency noise. While their effect on the bias-dependence of LFN has been well investigated, the way these noise mechanisms contribute to the bias-dependence of variability of LFN has been less well understood. LFN variability has been shown to be maximized in weak inversion (sub-threshold), while increased drain bias also increases LFN variability. However, no compact model has been proposed to explain this bias-dependence in detail. In combination with the charge-based formulation of LFN, the present paper proposes a new model for bias-dependence of LFN variability. Comparison with experimental data from moderately-sized NMOS and PMOS transistors at all bias conditions provides insight into how carrier number and mobility fluctuation mechanisms impact the bias-dependence of LFN variability.
A novel method for extracting threshold voltage and substrate effect parameters of MOSFETs with constant current bias at all levels of inversion is presented. This generalized constant-current (GCC) method exploits the charge-based model of MOSFETs to extract threshold voltage and other substrate-effect related parameters. The method is applicable over a wide range of current throughout weak and moderate inversion and to some extent in strong inversion. This method is particularly useful when applied for MOSFETs presenting edge conduction effect (subthreshold hump) in CMOS processes using Shallow Trench Isolation (STI).