convolutional network

Natural renewable energy sources are abundant and economically attractive, with zero or very low carbon emissions. On the other hand, harsh environmental and weather conditions, such as soil and dust accumulation, affect the efficiency of renewable energy sources and systems. Accordingly, the need for automated inspection of photovoltaic panels is becoming more critical as the demand for new solar energy system manufacturing and installation increases worldwide. This study introduces a new dataset of images of dusty and clean plates. Furthermore, a new convolutional neural network architecture is proposed to detect the voltage generated by the panel. In the following, the parameters taken from the environment and the voltage estimated by the proposed neural network are analyzed using the random forest regression algorithm, and the panel's efficiency is calculated. The proposed process deals explicitly with detecting dust accumulation in photovoltaic panels. The results obtained in this work have experimentally shown that the proposed system produces high detection rates. The proposed new method leads to the implementing of a more effective and efficient automatic cleaning technique for photovoltaic panels.

In the world of digital communication, authentication is an important concern and the need for a safe and secure system increases the necessity of designing authentication systems. To perform authentication, biometric-based approaches are of great interest due to the property of being alive and resistant to forgery. In this study, an authentication system based on heart signal is designed. Due to the process of receiving heart signals, their data usually has a lot of noise. In order to prepare the data, in the proposed system, the heart signals are first cleaned and then transferred to the frequency domain for feature extraction. Also, they are converted into an image by applying the Wigner-Ville distribution, so that each image contains the signal information of each person’s heart and is unique. In the proposed authentication system, these images are used for training and evaluation in a deep convolutional neural network. The output of this system provides the possibility of people’s identification. The data of this study are taken from the NSRDB and MITDB databases, and significant results have been obtained compared to previous studies.

Smartphones are now well integrated with advanced capabilities and technologies such as the Internet. Today, due to the facilities and capabilities and the widespread use of smart mobile devices, mobile security has become a vital issue worldwide. Smartphones are not properly protected compared to computers and computer networks, and users do not consider security updates. Recently, mobile devices and networks have been targeted by one of the most dangerous cyber threats known as botnets. Mobile Bantent An enhanced example of Boutons has the ability to perform malicious operations such as denial of service attacks, data theft, eavesdropping, and more. Bunters use three communication protocols: HTTP, SMS and Bluetooth to communicate with each other; So when users are not connected to the Internet, botnets are able to communicate with each other. In this study, to identify mobile batonet from 14 Android baton families, including 1932 samples of Android mobile devices applications and 4304 samples of safe and secure Android mobile devices applications have been used. Application permissions were extracted for reverse engineering to automatically classify and detect types of botnets, then based on these permissions, each application was converted to an equivalent image using the proposed method. Labeled images were then used to train convolutional neural networks. The results of evaluation and comparison of this method with classical methods including backup vector machine and decision tree showed that the proposed method is able to achieve higher efficiency in detecting different types of botnets and separating it from healthy programs

In a business environment where there is fierce competition between companies, accurate demand forecasting is vital. If we collect customer demand data at discrete points in time, we obtain a demand time series. As a result, the demand forecasting problem can be formulated as a time series forecasting task. In the context of time series forecasting, deep learning methods have demonstrated good accuracy in predicting complex time series. However, the excellent performance of these methods is dependent on the amount of data available. For this purpose, in this study, we propose to use time series augmentation techniques to improve the performance of deep learning methods. In this study, three new methods have been used to test the effectiveness of the proposed approach, which are: 1) Long short-term memory, 2) Convolutional network 3) Multihead self-attention mechanism. This study also uses a multi-step forecasting approach that makes it possible to predict several future points in a forecasting operation. The proposed method is applied to the actual demand data of a furniture company. The experimental results show that the proposed approach improves the forecasting accuracy of the methods used in most different prediction scenarios