Data classification based on fractional order gradient descent with momentum for RBF neural network.

The weight-updating methods have played an important role in improving the performance of neural networks. To ameliorate the oscillating phenomenon in training radial basis function (RBF) neural network, a fractional order gradient descent with momentum method for updating the weights of RBF neural network (FOGDM-RBF) is proposed for data classification. Its convergence is proved. In order to speed up the convergence process, an adaptive learning rate is used to adjust the training process. The Iris data set and MNIST data set are used to test the proposed algorithm. The results verify the theoretical results of the proposed algorithm such as its monotonicity and convergence. Some non-parametric statistical tests such as Friedman test and Quade test are taken for the comparison of the proposed algorithm with other algorithms. The influence of fractional order, learning rate and batch size is analysed and compared. Error analysis shows that the algorithm can effectively accelerate the convergence speed of gradient descent method and improve its performance with high accuracy and validity.

H ∞ Time-Delayed Fractional Order Adaptive Sliding Mode Control for Two-Wheel Self-Balancing Vehicles.

In this paper, a time-delayed fractional order adaptive sliding mode control algorithm is proposed for a two-wheel self-balancing vehicle system. The closed-loop system is proved based on the Lyapunov-Razumikhin function. The switching function is designed to make the system robust when facing uncertainties and external disturbances. It is designed to avoid monotonically increasing gains and can handle state-dependent uncertainties without a prior bound. The two-wheel self-balancing vehicle used in the experiment consists of a gyroscope MPU-6050 and accelerometer, a motor driving circuit composed of a motor driving chip TB6612FNG, and STM32F103x8B that is selected as the control core. The experimental results show that the time-delayed fractional order adaptive sliding mode control algorithm can make the vehicle achieve autonomous balance and quickly restore its stable state while appropriate disturbance is introduced.

Experimental Realization of an Extreme-Parameter Omnidirectional Cloak.

An ideal transformation-based omnidirectional cloak always relies on metamaterials with extreme parameters, which were previously thought to be too difficult to realize. For such a reason, in previous experimental proposals of invisibility cloaks, the extreme parameters requirements are usually abandoned, leading to inherent scattering. Here, we report on the first experimental demonstration of an omnidirectional cloak that satisfies the extreme parameters requirement, which can hide objects in a homogenous background. Instead of using resonant metamaterials that usually involve unavoidable absorptive loss, the extreme parameters are achieved using a nonresonant metamaterial comprising arrays of subwavelength metallic channels manufactured with 3D metal printing technology. A high level transmission of electromagnetic wave propagating through the present omnidirectional cloak, as well as significant reduction of scattering field, is demonstrated both numerically and experimentally. Our work may also inspire experimental realizations of the other full-parameter omnidirectional optical devices such as concentrator, rotators, and optical illusion apparatuses.

Manipulating surface plasmon polaritons with infinitely anisotropic metamaterials.

Guiding surface states through disorders recently has attracted attention of scientists from diverse backgrounds. In this work, we report a robust method to guide surface plasmon polaritons (SPPs) through arbitrary distorted metal surfaces (a kind of disorder), including slopes, bumps, and sharp corners. Almost total transmissions over a broad frequency range can be achieved by use of infinitely anisotropic metamaterials (IAMs). The SPPs are coupled into and out of the bulk modes in the IAMs, where the bulk modes are routed by altering the principle axis of the IAMs. Due to unique non-diffraction property of the IAMs, all processes are of high efficiency, which are explained from both microscopic and macroscopic perspectives. Several functional SPP devices, including adapter, cloak, and sharp bending waveguide, are presented in the simulations. Two proof-of-concept SPP devices are experimentally demonstrated, where the SPPs are mimicked by the designer SPPs at microwave frequency.