The optical signal-to-noise ratio (OSNR) and fiber nonlinearity are critical factors in evaluating the performance of high-speed optical fiber communication systems. Recently, several deep learning based methods have been put forward to monitor OSNR of a fiber communication system. In this work, we propose a long short-term memory (LSTM) network based method to simultaneously estimate OSNR and nonlinear noise power caused by fiber nonlinearity. In the training step, LSTM network extracts the essential features in frequency domain of the input signal. Then, with the built model in the training step, the LSTM output the OSNR and nonlinear noise power of the signal under test. The simulation by VPI software is carried on a 5-channel long haul optical transmission system with the launched optical power of −3.0~ + 3.0dBm per channel. The results show that the test error of OSNR is less than 1.0dB with the reference OSNR from 15 to 30dB for QPSK, 16QAM and 64QAM signal. The test error of nonlinear noise power is less than 1.0dB for QPSK and 16QAM signal when the Laser linewidth is 6 KHz and 100 KHz respectively. The proposed method is a promising candidate for nonlinearity-insensitive OSNR and accurate nonlinear noise power estimation in multi-channel long haul optical fiber communication systems.

In this letter, we propose an indoor visible light positioning technique that combines deep neural network based on the Bayesian Regularization (BR-DNN) with sparse diagonal training data set. Unlike other neural networks, which require a large number of training data points to locate accurately, we realize the high precision positioning with only 20 training points in a $\rm 1.8m\times 1.8m\times 2.1m$ location area. Furthermore, we test a new optimization method of training data set which is the diagonal set. To verify our ideas, we experimentally demonstrate three different training data acquisition methods which contain the common choice of training points (even set), arbitrary selection (arbitrary set) and diagonal selection (diagonal set). Experimental results show that the average localization accuracy optimized by the BR-DNN is 3.40 cm with diagonal set, while the average localization accuracy is 4.35 cm for the arbitrary set and 4.58cm for the even set. In addition, the training time and positioning time are only 11.25ms and 8.66ms due to significant reduction of the sparse training data. All of above experimental results show that the algorithm and training data optimization we proposed provide a new solution for real-time and high accuracy positioning with neural network.

With the rapid development of smart technology, the need for location-based services (LBS) increases every day. Since classical positioning technology such as GPS cannot satisfy the needs of indoor positioning, new indoor positioning technologies, such as Bluetooth, Wi-Fi, and Visible light communication (VLC), have already cut a figure. VLC positioning has been proposed because it has higher accuracy, costs less, and is easier to accomplish in comparison to the other indoor positioning technologies. However, the practicality of VLC positioning is limited since it is easily affected by multipath effects and the layout of LEDs. Thus, we propose a fusion positioning system based on extended Kalman filters, which can fuse the VLC position and the inertial navigation data. The accuracy of the fusion positioning system is in centimeters, which is better compared to the VLC-based positioning or inertial navigation alone. Furthermore, the fusion positioning system has high accuracy, saves energy, costs little, and is easy to install, making it a promising candidate for future indoor positioning applications.

Modulation format identification is an important part of optical performance monitoring in elastic optical networks. We propose a modulation format identification scheme based on the information entropy analysis of received optical communication signal. In our method, the information entropy is calculated based on the amplitude distribution of received signal and employed to identify 4 commonly used modulation formats like QPSK, 16QAM, 32QAM, and 64QAM. The results demonstrate that the modulation format can be identified with the required OSNR not higher than the FEC threshold. The influence of the residual chromatic dispersion on the proposed scheme is also discussed. It is insensitive to phase noise caused by laser linewidth and frequency offset, and can tolerate a range of fiber nonlinearity, which is also an advantage of our method. Compared to other blind modulation format identification methods, the proposed method needs no additional hardware or training samples, is more simple and cost-effective.

In this letter, we propose an indoor visible light positioning technique using a Modified Momentum Back-Propagation (MMBP) algorithm based on received signal strength (RSS) with sparse training data set. Unlike other neural network algorithms that require a large number of training data points to locate accurately, we have realized high-precision positioning for 100 test points with only 20 training points in a 1.8 m × 1.8 m × 2.1 m localization area. In order to verify the adaptability of the MMBP algorithm, we experimentally demonstrate two different training data acquisition methods adopting either even or arbitrary training sets. In addition, we also demonstrate the positioning accuracy of the traditional RSS algorithm. Experimental results show that the average localization accuracy optimized by our proposed algorithm is only 1.88 cm for the arbitrary set and 1.99 cm for the even set, while the average positioning error of the traditional RSS algorithm reaches 14.34 cm. Comparison indicates that the positioning accuracy of our proposed algorithm is 7.6 times higher. Results also show that the performance of our system is higher than some previous reports based on RSS and RSS fingerprint databases using complex machine learning algorithms trained by a large amount of training points.

Simultaneous localization and mapping (SLAM) are fundamental elements for many emerging technologies, such as autonomous driving and augmented reality. For this paper, to get more information, we developed an improved monocular visual SLAM system by using omnidirectional cameras. Our method extends the ORB-SLAM framework with the enhanced unified camera model as a projection function, which can be applied to catadioptric systems and wide-angle fisheye cameras with 195 degrees field-of-view. The proposed system can use the full area of the images even with strong distortion. For omnidirectional cameras, a map initialization method is proposed. We analytically derive the Jacobian matrices of the reprojection errors with respect to the camera pose and 3D position of points. The proposed SLAM has been extensively tested in real-world datasets. The results show positioning error is less than 0.1% in a small indoor environment and is less than 1.5% in a large environment. The results demonstrate that our method is real-time, and increases its accuracy and robustness over the normal systems based on the pinhole model.