Hossam M. H. Shalaby

Presents corrections to the paper, Neural Network Ensemble for Precise Laser Spot Position Determination on a Quadrant Detector.

A new highly accurate optical biomedical sensor is proposed for cancer detection characterized by high sensitivity, small footprint, low cost, and low limit of detection. The sensor is based on double-ring resonators made of silicon on insulator. The type of the waveguide is critical in determining the sensor performance. To compromise the advantages and disadvantages of strip and slot waveguides, a mixed design of both has been introduced in literature at the expense of increased footprint compared to traditional sensors. Indeed, almost 27-fold footprint increase is required to improve the sensitivity by only one third of that of slot waveguide’s sensitivity. In this paper, we introduce a new design that keeps the same footprint of traditional sensors, while achieving high sensitivity. This sensitivity depends on the resonance wavelength shift due to different refractive indices of the biosample. It has the value 109.8 nm/RIU compared to 55.57 nm/RIU and 129.621 nm/RIU for strip and slot waveguides, respectively. The hybrid waveguide quality factor is 537.7 while the quality factors of the strip and the slot waveguides are 627.99 and 380.76, respectively. In addition, the new design achieves the minimum limit of detection (0.0255) when compared to that of traditional designs. Furthermore, a new method of detection is proposed with the same design, providing a higher sensitivity over both traditional waveguide types with a value of 15.3, compared to 13.2 and 11.5 for strip and slot waveguides, respectively. In this method, the sensitivity relies on various values of output transmitted light at the same wavelength due to altering the biocell refractive index. The biosensor output equation is developed. In addition, the relationship between the supermodes and the sensitivity is determined at variance conditions. It is found that there is an inverse relation between them.

A quadrant detector (QD) is a widely used technology for laser spot position sensing in establishing and maintaining laser communication links. There is a nonlinear relationship between the output signal offset and the precise location of the laser spot on the QD, which impairs position detection accuracy. To address the problem, this work employs a neural network ensemble solution. An ensemble of neural networks combines the predictive power of multiple artificial neural network (ANN) models, resulting in improved generalization and prediction stability compared to a single neural network. Test results indicate that our solution demonstrates a significant improvement in both the estimated position error and the generalization capability when compared to individual ANN solutions.

We propose and numerically analyze a broadband, ultracompact hybrid plasmonic mode converter/polarization rotator. The proposed device converts the fundamental transverse magnetic mode (TM₀) to the first-order transverse electric mode (TE₁). The mode converter is designed using a tapered hybrid plasmonic waveguide in which metal parts are directly surrounding both the input Si waveguide and the tapered output waveguide without a thin low-index layer. After optimizing the proposed structure, a TM₀-to-TE₁ mode converter with a modal conversion efficiency of 98.2% is achieved. A broad operating bandwidth of 100 nm is achieved with a compact footprint of only 0.8µm×12.5µm for the whole device. At a wavelength of 1550 nm, the insertion loss ranges from 0.91 dB to 1.19 dB using different metal materials, and the extinction ratio of the TE₁ and TM₀ modes is higher than 21 dB in the output Si waveguide.

An ultrafast, compact, zero-biased, and complementary metal–oxide semiconductor-compatible graphene photodetector (PD) based on a silicon-on-insulator hybrid plasmonic waveguide (HPWG) is proposed. Lumerical MODE solver is employed to investigate the modal characteristics of TM-polarized modes in the HPWG composing the PD. It is shown that the input light can be completely coupled into the photonic-like and plasmonic-like fundamental TM modes at the PD section. These two modes are exploited together in the photodetection process to enhance the PD performance. A rigorous analysis is performed in order to extract the optoelectronic characteristics of the single-layer graphene (SLG) used in the proposed structure. Lumerical 3D-FDTD solver is then employed to quantify the light interaction of the two aforementioned optical modes with the SLG. With a proper design at a wavelength of 1550 nm, the PD voltage responsivity reaches 2.8 V/W, and the photocurrent responsivity is obtained as 18.5 mA/W, while the corresponding absorption length is kept below 8 µ m and the noise equivalent power is limited to 3.7 p W / H z . Moreover, as the PD operates under zero bias, its photoresponse is predominated by the photothermoelectric mechanism, exhibiting a bandwidth that exceeds 180 GHz while avoiding the dark current.

A converter that manipulates the energy exchange between two arbitrary guided modes having different mode orders and polarization states is proposed. This objective is achieved through engraving a periodic grating on a strip waveguide. Theoretical analysis, based on the coupled-mode theory, is developed to match the proposed structure. In addition, coupling analysis is addressed to provide a comprehensive view regarding the various couplings that can be achieved, including acceptable margins for perturbation designing parameters and their optimum values for each coupling. Moreover, our converter is implemented via both 3D-FDTD simulation and computational solution to examine the validity of the proposed approach and verify the coupling analysis through performing two conversions, namely TM1-to-TE0 and TM1-to-TE3. The first conversion is obtained with an insertion loss of-1 dB and a crosstalk lower than-17.5 dB at a conversion length of 9.15 mu m. The second is executed with an insertion loss of - 1.5 dB and a crosstalk lower than -15 dB at a conversion length of 16.937 mu m. Furthermore, tolerance fabrication analysis is implemented to confirm the degree of stability that the proposed structure can achieve. Finally, the findings reveal that the proposed design achieved its function at a compact length and without being restricted by the hybridization approach.

A hybrid multi-state orbital angular momentum-multi pulse-position modulation ( $N$ OAM-MPPM) scheme over gamma-gamma free-space optical ( ${\Gamma \Gamma }$ -FSO) channel is studied in this paper. In our study, all atmospheric and pointing error impacts are taken into account. Expressions for the parameters of ${\Gamma \Gamma }$ -FSO-pointing error channel are derived. In addition, approximate-tight upper bounds on the bit-error rates (BERs) of $N$ OAM and $N$ OAM-MPPM techniques are developed over ${\Gamma \Gamma }$ -FSO-pointing error channels, considering the influences of beam divergence and pointing error (PE). The ${\Gamma \Gamma }$ -FSO-PE channel parameters and the BER expressions are evaluated numerically and verified by simulation. It turned out that the analytical results are nearly the same as those obtained from simulation under different turbulence scenarios and OAM modes. The results demonstrate that under variable turbulence conditions, the $N$ OAM-MPPM technique outperforms both ordinary $N$ OAM and MPPM systems. Furthermore, different deep learning (DL) techniques, namely random forest (RF), convolution neural network (CNN), and auto-encoder (AE), are employed to get the optimum classification accuracy using different datasets of $N$ OAM-MPPM over ${\Gamma \Gamma }$ -PE channel model. Finally, the results indicate that AE has the best performance metrics compared to other models using different datasets.

Orbital angular momentum-shift keying (OAM-SK), which is the rapid switching of OAM modes, is vital but seriously impeded by the deficiency of OAM demodulation techniques, particularly when videos are transmitted over the system. Thus, in this paper, 3D chaotic interleaved multi-coded video frames (VFs) are conveyed via an N-OAM-SK free-space optical (FSO) communication system to enhance the reliability and efficiency of video communication. To tackle the defects of the OAM-SK-FSO mechanism, two efficient deep learning (DL) techniques, namely convolution recurrent neural network (CRNN) and 3D convolution neural network (3DCNN) are used to decode OAM modes with a low bit error rate (BER). Moreover, a graphics processing unit (GPU) is used to accelerate the training process with slight power consumption. The utilized datasets for OAM states are generated by applying different scenarios using a trial-and-error method. The simulation results imply that LDPC-coded VFs achieve the largest peak signal-to-noise ratios (PSNRs) and the lowest BERs using the 16-OAM-SK model. Both 3DCNN and CRNN techniques have nearly the same performance, but this performance deteriorates in the case of larger dataset classes. Moreover, the GPU accelerates the performance by almost 67.6% and 36.9% for the CRNN and 3DCNN techniques, respectively. These two DL techniques are more effective in evaluating the classification accuracy than the other traditional techniques by almost 10 - 20%.

Data offloading is a promising low-cost and power-efficient solution for the expected high demands for high-speed connectivity in the near future. We investigate offloading efficiency in a cellular/light fidelity (LiFi) network. This offloading efficiency is a measure of the ratio of traffic carried by the LiFi network to the total traffic carried by both LiFi and cellular networks. We consider the two scenarios of opportunistic and delayed offloading. Effects of user density, user mobility, LiFi-signal blocking, and channel characteristics are investigated. We use Zemax to simulate LiFi channels in the proposed model. Based on our results, delayed offloading can achieve up to 60% offloading efficiency while opportunistic offloading achieves up to 18% offloading efficiency.