We experimentally demonstrate a novel nonlinearity-mitigation scheme based on a complex-valued neural network (CVNN) which is constructed by artificial neurons with complex-valued input and output. The in-phase (I) and quadrature (Q) components of optical signal are operated as complex values in the CVNN. A 40-Gbit/s optical 16QAM signal distorted by SPM was successfully compensated, improving error vector magnitude (EVM) by about 15%. The learning speed of the nonlinear equalizer was improved by using the CVNN, compared with conventional real-valued neural network (RVNN). Furthermore, the study show that CVNN has the potential to improve the computational complexity of RVNN.

We propose a novel self-homodyne scheme using a polarization-multiplexed and intensity-modulated pilot-carrier. The scheme has the capability to cancel nonlinear waveform distortion caused by self-phase modulation (SPM) in optical fibers, as well as the phase-noise of a laser diode. It has been known that a conventional constant-intensity pilot-carrier can cancel the waveform distortion caused by SPM to some extent. Theoretically, however, the proposed intensity-modulated pilot-carrier can completely cancel the effect of SPM. The performance in terms of the error vector magnitude was evaluated by numerical simulations of 50-km optical fiber transmission of 16-ary quadrature amplitude modulation (16QAM) signals.

We propose a novel SPM and phase-noise cancellation scheme using pilot symbols. The pilot symbols are intensity-modulated and time-division-multiplexed with a modulated optical signal. Each pilot symbol has the same optical intensity as the multiplexed modulated symbol and therefore has the same nonlinear phase shift caused by self-phase modulation (SPM). Additionally, each pilot symbol has the same phase noise as the modulated symbol because it is generated by the same laser diode (LD). At the receiver side, self-homodyne detection is performed using the pilot symbols as absolute optical phase references. The performance of the scheme was investigated by numerical simulations of 50-km optical fiber transmission of 16-ary quadrature amplitude modulation (16QAM) signals.

We investigate a novel frequency-quadruple optical two-tone signal generation method using a DP-QPSK modulator, which includes four Mach-Zehnder modulators (MZMs). One MZM is used to generate a basic two-tone signal. The other MZMs are used to compensate unwanted high-order harmonics and carrier components of the generated signal. In the experiment, a two-tone signal was successfully generated, and the power levels of fourth-order harmonics and the carrier were reduced to levels 36.8 dB and 40.3 dB smaller than the required signal, respectively. We also evaluated suppression ratio of the unwanted components against fluctuations of amplitude and bias voltages for the MZMs.

We propose a novel detection scheme for an optical twin single-sideband (twin-SSB) signal. The detection can be performed by one optical homodyne receiver with an electric butterfly operation which demultiplexes the lower side-band (LSB) and upper side-band (USB) signals without using optical filters. The dispersion-tolerant characteristics of the scheme were investigated by numerical simulation of transmission through 50 km and 100 km standard single-mode fiber (SSMF), in comparison with dual sideband (DSB) modulation. The optical modulation of the twin-SSB signal was performed using a dual polarization quadrature phase shift keying (DP-QPSK) modulator.

This paper describes a technique for predicting the received signal strength from a transmitter mounted on an automated guided vehicle (AGV). The predictor is part of a factory-based wireless communications anomaly detection system. We have proposed training a probabilistic neural network (PNN) to learn sliding-window sequences of the received fading signal. After training, the PNN conducts pattern matching between trained and observed sequences and returns an index to the closest matching set. Future deep-fading is predicted by adding an index offset to a fading look-up-table memory. We further propose to improve the performance by combining the predicted indices from multiple receivers. Performance results from ray-traced channels show that the mean squared error is reduced by almost two orders of magnitude by employing four receivers at 10 dB SNR.

Using virtualized CDN such as Amazon CloudFront in which cache servers are provided on virtual machines, content providers can flexibly select the regions of using cache servers according to the geographical demand pattern of their users, so they can expect to reduce the cost of using CDN. However, to maximize the profit of content providers, they need to carefully select the locations of cache servers which affect the quality perceived by users as well as the total cost. We propose a method of optimally selecting the geographical regions to use cache servers maximizing the profit of content providers.

A digital signage image-sensor-based visible light communication (IS-VLC) system can transmit data information by embedding it on visual information. The use of a motion video as visual information makes it difficult to completely erase the visual information in this system. The extraction of data information is an important issue because the incomplete extraction will degrade the bit error rate performance. In this letter, we expand the IS-VLC system to overlay a motion video with data information by introducing a cell-position correction method and a noise elimination method. Experimental results clarify the effectiveness of the extended system.

We numerically assess the performance and computational complexity of split-step Fourier method based digital backpropagation (DBP) with phase linear approximation (PLA-DBP) for reduced complexity on digital implementation. From Manakov equation, we can confirm that the phase of the received signal compensated with DBP changes almost linearly between optical amplifiers, thus the input signal phase even with dual-polarization can be estimated using linear approximation using only two points. We show the performance evaluation of 28 Gbaud single carrier dual polarization-16 quadrature amplitude modulation (DP-16QAM) format, and 20.0% reduction in complexity is obtained with transmission distance of 4000 km and input power of 3 dBm.

In global navigation satellite systems (GNSS), we need signals from more than four satellites to find out the location at the receiver. However, the satellite visibility is occasionally blocked especially in city areas. For augmenting the availability, placing the pseudo satellite, or pseudolite, has been proposed. A pseudolite broadcasts the signal in the same format as the satellites broadcast and acts as an additional satellite to the receiver. However, the satellite signals are often blocked by a pseudolite signal at the receiver in the close proximity to the pseudolite, and the phenomenon is referred to as the near-far problem. In this letter, a pseudolite with burst pulse transmission that the power varies pulse-to-pulse is proposed to mitigate the near-far problem.

In indoor wireless environments such as warehouses and factories, since radio propagation occurs in complicated ways and there are large moving objects like Automated Guided Vehicles (AGVs), from the perspective of network management, maintaining wireless communication quality is challenging. Because the failures in wireless communications may lead to undesirable delay or even stop in the manufacturing processes, it is desirable not only to detect degradations in link quality, but also to estimate the cause of the frame losses, to recover from the link failures as quick as possible. In this paper, we propose a sensing system that allows comprehensive analysis between physical and Medium Access Control (MAC) layer information, and allows estimation of the cause of the frame losses due to signal attenuation or interference. We implemented a prototype sensor node using an FPGA board and conducted an experiment to investigate the accuracy of the estimation. From the experiment, we confirmed that the prototype sensor node can determine whether the cause is due to attenuation or interference with high accuracy.

In this study, we utilize an effective method called Characteristic Mode Analysis (CMA) to analyze the physical mechanisms of U-shaped Folded Dipole Antenna (UFDA), which have been proposed for WiMAX device. The proposed antenna consists of the UFDA element and small Ground Plane. As a result, it is confirmed that two types of modes in the resonance of UFDA appear, the one is the mode of the Ground Plane and the other is the mode of the UFDA element. Subsequently, from these results, an easier method for broadband design of UFDA which can cover completely two frequency bands of WiMAX was investigated.

This paper presents penetration loss measurements of housing outer wall materials at 79 GHz to investigate the co-existence between indoor-use and outdoor-use radar sensors. The penetration loss was measured for building outer wall materials such as window and concrete wall. The penetration loss is also compared with 24 GHz-band because the band is available as indoor-use (e.g., non-contacted health-care monitoring) and outdoor-use (e.g., automotive radar) sensor systems. The loss of indoor-to-outdoor systems is experimentally investigated. As a result, the propagation path through outer wall material, at a distance of 5 m from the house, is found to be approximately 50 dB attenuated, which is 15 dB larger as compared with 24 GHz.

In this paper, two types of patch antenna that can be used for animal experiments are proposed.

The proposed antenna can generate a localized electric field at the head of the rat. The operating frequency of the system is 26.5 GHz. This frequency is used the fifth generation of mobile communication (5G). The localization of the electric field is evaluated by the ratio between target area filed and whole body field.

Rotation of rotor blades is typically micro-motion and would induce micro-Doppler modulation. However, accurate detection will become extremely difficult when the echo signal is corrupted by strong noise. This paper proposes a method to realize extraction of micro-Doppler feature from the rotating rotor blade echo signals with low signal-to-noise ratio (SNR). The proposed method is characterized by applying low-rank decomposition for saliency detection of spectrogram. MATLAB simulation results verify that the proposed method is effective in saliency detection of the spectrogram from rotor blade echo signals and is robust in estimating the maximum micro-Doppler shift under low SNR.