A Survey of Intrusion Detection Using Deep Learning in Internet of Things

The use of deep learning in various models is a powerful tool in detecting IoT attacks, identifying new types of intrusion to access a better secure network. Need to developing an intrusion detection system to detect and classify attacks in appropriate time and automated manner increases especially due to the use of IoT and the nature of its data that causes increasing in attacks. Malicious attacks are continuously changing, that cause new attacks. In this paper we present a survey about the detection of anomalies, thus intrusion detection by distinguishing between normal behavior and malicious behavior while analyzing network traffic to discover new attacks. This paper surveys previous researches by evaluating their performance through two categories of new datasets of real traffic are (CSE-CIC-IDS2018 dataset, Bot-IoT dataset). To evaluate the performance we show accuracy measurement for detect intrusion in different systems.

Graphene as a Piezoresistive Material in Strain Sensing Applications

High accuracy measurement of mechanical strain is critical and broadly practiced in several application areas including structural health monitoring, industrial process control, manufacturing, avionics and the automotive industry, to name a few. Strain sensors, otherwise known as strain gauges, are fueled by various nanomaterials, among which graphene has attracted great interest in recent years, due to its unique electro-mechanical characteristics. Graphene shows not only exceptional physical properties but also has remarkable mechanical properties, such as piezoresistivity, which makes it a perfect candidate for strain sensing applications. In the present review, we provide an in-depth overview of the latest studies focusing on graphene and its strain sensing mechanism along with various applications. We start by providing a description of the fundamental properties, synthesis techniques and characterization methods of graphene, and then build forward to the discussion of numerous types of graphene-based strain sensors with side-by-side tabular comparison in terms of figures-of-merit, including strain range and sensitivity, otherwise referred to as the gauge factor. We demonstrate the material synthesis, device fabrication and integration challenges for researchers to achieve both wide strain range and high sensitivity in graphene-based strain sensors. Last of all, several applications of graphene-based strain sensors for different purposes are described. All in all, the evolutionary process of graphene-based strain sensors in recent years, as well as the upcoming challenges and future directions for emerging studies are highlighted.

Blending Colored and Depth CNN Pipelines in an Ensemble Learning Classification Approach for Warehouse Application Using Synthetic and Real Data

Electric companies face flow control and inventory obstacles such as reliability, outlays, and time-consuming tasks. Convolutional Neural Networks (CNNs) combined with computational vision approaches can process image classification in warehouse management applications to tackle this problem. This study uses synthetic and real images applied to CNNs to deal with classification of inventory items. The results are compared to seek the neural networks that better suit this application. The methodology consists of fine-tuning several CNNs on Red–Green–Blue (RBG) and Red–Green–Blue-Depth (RGB-D) synthetic and real datasets, using the best architecture of each domain in a blended ensemble approach. The proposed blended ensemble approach was not yet explored in such an application, using RGB and RGB-D data, from synthetic and real domains. The use of a synthetic dataset improved accuracy, precision, recall and f1-score in comparison with models trained only on the real domain. Moreover, the use of a blend of DenseNet and Resnet pipelines for colored and depth images proved to outperform accuracy, precision and f1-score performance indicators over single CNNs, achieving an accuracy measurement of 95.23%. The classification task is a real logistics engineering problem handled by computer vision and artificial intelligence, making full use of RGB and RGB-D images of synthetic and real domains, applied in an approach of blended CNN pipelines.

Cloud-Based Parkinson’s Disease Diagnosis Using Machine Learning

Parkinson’s disease (PD) harms the human brain's nervous system and can affect the patient's life. However, the diagnosis of PD diagnosis in the first stages can lead to early treatment and save costs. In this paper, a cloud-based machine learning diagnosing intelligent system is proposed for the PD with respect to patient voice. The proposed system is composed of two stages. In the first stage, two machine learning approaches, Random-Forest (RF) and Long-Short-Term-Memory (LSTM) are applied to generate a model that can be used for early treatment of PD. In this stage, a feature selection method is used to choose the minimum subset of the best features, which can be utilized later to generate the classification model. In the second stage, the best diagnosis model is deployed in cloud computing. In this stage, an Android application is also designed to provide the interface to the diagnosis model. The performance evaluation of the diagnosis model is conducted based on the F-score accuracy measurement. The result shows that the LTSM model has superior accuracy with 95% of the F-score compared with the RF model. Therefore, the LSTM model is selected for implementing a cloud-based PD diagnosing application using Python and Java.

On-line Measurement Method and Technical Scheme of Transmission Line Crossing Point Based on Machine Vision

With the increasing demand for high voltage and UHV transmission, higher and higher requirements are put forward for transmission line monitoring and fault diagnosis. The traditional line measurement method has many shortcomings, which cannot ensure the measurement accuracy, measurement efficiency and high cost. Based on this, this paper first studies the principle of machine vision on-line measurement, then analyses the on-line measurement process of transmission line crossing point based on machine vision, and finally gives the error source and amelioration measures of transmission line crossing point on-line measurement.

Fabrication of Implantable Capsule-Type Channel Sounders for the High-Accuracy Measurement of the Signal Loss of Out-Body to In-Body Channels

This paper proposes a channel sounder to measure the channel properties of an implantable device that transmits data from inside to outside the human body. The proposed channel sounder measures the receiving power of a signal transmitted from outside the human body. The channel sounder is equipped with a Bluetooth module that enables the wireless transmission of the receiving power outside the human body. Wireless transmission enables the channel measurement by isolating the transmitter and receiver inside the channel sounder. Using the proposed channel sounder, the channel properties can be measured without any interference between the transmitter and the receiver.

Low-Cost Polyethylene Terephthalate Fluidic Sensor for Ultrahigh Accuracy Measurement of Liquid Concentration Variation

A low-cost polyethylene terephthalate fluidic sensor (PET-FS) is demonstrated for the concentration variation measurement on fluidic solutions. The PET-FS consisted of a triangular fluidic container attached with a birefringent PET thin layer. The PET-FS was injected with the test liquid solution that was placed in a common path polarization interferometer by utilizing a heterodyne scheme. The measured phase variation of probe light was used to obtain the information regarding the concentration change in the fluidic liquids. The sensor was experimentally tested using different concentrations of sodium chloride solution showing a sensitivity of 3.52 ×104 deg./refractive index unit (RIU) and a detection resolution of 6.25 × 10−6 RIU. The estimated sensitivity and detection resolutions were 5.62 × 104 (deg./RIU) and 6.94 × 10−6 RIU, respectively, for the hydrochloric acid. The relationship between the measured phase and the concentration is linear with an R-squared value reaching above 0.995.