Neutrosophic and Information Fusion

Brain Computer Interface (BCI), especially systems for recognizing brain signals using EEG (Electroencephalogra-phy), is one of the important research topics that arouse the interest of many researchers currently. Convolutional Neural Nets (CNN) is one of the most important deep learning classifiers used in this recognition process, but the parameters of this classifier have not yet been precisely defined so that it gives the highest recognition rate and the lowest possible training and recognition time. This research proposes a system for recognizing EEG signals using the CNN network, while studying the effect of changing the parameters of this network on the recognition rate, training time, and recognition time of brain signals. As a result, the proposed recognition system achieved a 76.38% recognition rate and reduced classifier training time by 3 seconds by using Common Spatial Pattern (CSP) in preprocessing the IV2b dataset. A recognition rate of 76.533% was reached by adding a layer to the proposed classifier.

This research presents a novel simulation model of adaptive control to implement a control process using the MIT-rule adjustment mechanism for power oscillation damping in the single-machine infinite-bus (SMIB) system, and to measure its possible effects on the response of the damper by changing its parameters according to an external disturbance using Simulink. The results show that the use of the model-reference adaptive control (MRAC) technique maintains the response of the damper when the transfer function changes because of an external disturbance.

We studied the stability of the steady-state solutions for the Fisher equation in two cases. In the first case, with constant amplitude, we show that the steady-state solution u1 = 1 is always stable under any condition, whereas the other two solutions, u1 = 0 and u1(x) = Acos(nπX), are conditionally stable. In the second case, we studied the steady-state solutions for variable amplitude using two methods. The first is an analytical direct method, and the second is a numerical method using the Galerkin technique. Both methods show the same results: the steady-state solution u1 = 1 is always stable under any conditions, whereas the other two solutions, u1 = 0 and u1(x) = Acos(nπX), are conditionally stable.

The Q-learning algorithm in reinforcement learning is one of the algorithms that allows the robot to learn the surrounding environment without the need for prior training samples, using the principle of reward and punishment through interaction with the environment. Increasing the number of hidden layers of the deep neural network and adjusting some of its hyperparameters can increase the robot’s reward and thus obtain the best path to achieve the goal.

A k-arc in a plane PG(2,q) is a set of k points such that every line in the plane intersects it in at most two points and there is a line intersecting it in exactly two points. A k-arc is complete if there is no (k + 1)-arc containing it. This study is concerned with k-arcs, k = 4,5,...,10, and with the classification of projectively distinct k-arcs and distinct arcs under collineation. Using a computer program, we prove that complete k-arcs occur only for k = 6 and k = 10.