Transactions on Machine Intelligence2821-16932120190301Designing and Fabrication of Switching High Voltage Power Supply for Nuclear Medicine Imaging Systems1915972710.47176/TMI.2019.1ENM.MoradiDepartment of Medicine Engineering, Shahid Beheshti University, Tehran, IranZ.ArvanehDepartment of Power Engineering, Islamic Azad University, Ilam, IranJournal Article20190129This paper describes design considerations of the High voltage (HV) module with low ripple and good stability in order to improve the efficiency of nuclear imaging systems consist of Single Photon Emission Computed Tomography (SPECT). In the design of the current HV module, push-pull topology is preferred, due to the fact; it has a better efficiency in comparison with other technologies. The module receives a positive continuous voltage about 5-8 V as input, and provides a programmable positive or negative continuous voltage from 800-2500 V as output for a total load of 200-KΩ. The ripple of the output voltage is 10 mV that it is measured by laboratory equipment’s and safe high voltage (SHV) probe. The measured output power in the designed module is 25 W. Experiment results show that minimum efficiency at full load is around 65%. The proposed HV module shows a good stability, low ripple (noise), fast rise time. So, it can be used in SPECT or gamma camera imaging system to provide the constant voltage for PMTS.https://www.tmachineintelligence.ir/article_159727_4a6082dccb1fd65f87ed59de05a79292.pdfTransactions on Machine Intelligence2821-16932120190301Magnetometer Calibration Using Ellipsoid Fitting Method101915972810.47176/TMI.2019.10ENA.FarsimadanDepartment of Control Engineering, Damavand Branch, Islamic Azad University, Damavand, IranV.BehnamgolDepartment of Control Engineering, Damavand Branch, Islamic Azad University, Damavand, IranJournal Article20181125Given the wide range of applications for magnetic sensors, particularly magnetometers, precise calibration is essential to ensure reliable and accurate measurements. Magnetic sensors are highly sensitive to both internal imperfections and external disturbances, including environmental magnetic fields, which can significantly affect their output. To approximate real-world conditions, this study involved mounting a magnetometer on a vehicle and conducting measurements in a remote, low-interference region in Iran. The vehicle was driven within a limited radius to collect data under relatively undisturbed geomagnetic conditions. The raw data acquired by the sensor was recorded and subsequently processed using MATLAB software, employing the Ellipsoid Fitting Method for calibration. This technique is widely recognized for correcting both hard-iron and soft-iron distortions typically present in three-axis magnetometers. The study begins with a concise overview of common magnetometer calibration approaches, comparing their principles and practical implications. It then presents the experimental setup, data collection procedure, calibration process, and outcomes in detail. The results indicate a significant improvement in the accuracy of the magnetometer readings after calibration, demonstrating the effectiveness of the adopted method. This investigation highlights the practical considerations and challenges involved in magnetometer calibration under real-world conditions and contributes to the optimization of sensor performance for future field applications.https://www.tmachineintelligence.ir/article_159728_7a5950be6379345c43b3f47f48a8347a.pdfTransactions on Machine Intelligence2821-16932120190301TCP Low Rate DDoS Attack Detection203215972910.47176/TMI.2019.20ENM. R.RazianIran University of Science and Technology, Tehran, IranJournal Article20181125Undoubtedly one of the more significant attacks on computer networks is distributed denial of service (DDoS). DDoS assaults can be divided into two groups: high rate attacks and low rate attacks. In the high rate DDoS category, the attacker tries to use all of the bandwidth available on the channel by saturating it with packets. While maintaining a low average transmission rate, the attacker conducts a DDoS attack in the low rate DDoS category (also known as LDDoS). TCP LDDoS is a low-rate DDoS assault in which the attacker takes advantage of the way TCP handles congestion. In this article, we look into a system for stopping a TCP LDDoS attack and suggest a fresh approach. We offer several observations to help distinguish between appropriate behavior and an attack. Our system produces a priority queue of flows, where flows with a high priority are valid and flows with a low priority are suspect. Using the NS2 simulation environment, we assess the suggested system. Results demonstrate that our suggested approach can accurately distinguish between attack flows and genuine flows.https://www.tmachineintelligence.ir/article_159729_cab9a730117f9a6bbf8fb7e470b0af10.pdfTransactions on Machine Intelligence2821-16932120190301A SVM Based Approach for Real Time Detection and Classification of Vehicles at the Toll Gates Using Video Sequences333915973010.47176/TMI.2019.33ENA.MosayebiElectrical Engineering Department, Shahrood University, Shahrood, Semnan, IranH.KhosraviElectrical Engineering Department, Shahrood University, Shahrood, Semnan, IranJournal Article20181119This paper aims to present a real-time scheme for detection and classification of vehicles passing the toll gates in Iran. In our approach, a set of videos are captured using a stationary camera, placed on the roadside, at a little distance from the toll booth. The algorithm is designed in a way that there is no need for camera calibration. Based on our videos, 3 ROIs are defined, two of them are considered to determine if a vehicle is passing and the other one is the region containing the vehicle. This work starts with the training phase, in which, for each image in a manually gathered database, HOG vectors are extracted. Two SVMs are trained in this phase, one for distinguishing vehicles from non-vehicles, and one for classifying vehicles into light and heavy vehicles. After finishing the training, in the testing phase, firstly, foreground mask is obtained differencing two consecutive frames of the video. Then, those two aforementioned ROIs are checked in every frame and as soon as a vehicle is inside the interest region, that ROI is captured. Next, the captured frame is passed to the first SVM and it is classified as vehicle or non-vehicle. Those which are identified as vehicles are passed to the second SVM to be classified as light or heavy vehicle. Average true-positive and precision rates of the vehicle detection step are 92.5% and 97.5% respectively and the same rates, for the recognition step, are 98% and 0.99%.https://www.tmachineintelligence.ir/article_159730_687cf17ad64ae98deb46c36e8db65558.pdfTransactions on Machine Intelligence2821-16932120190301Speed Control Of The Interior Permanent Magnet Synchronous Motor Over A Wide Range Using Fuzzy Logic Controller404715973110.47176/TMI.2019.40ENM.BakhtiariUniversity of Isfahan, Isfahan, IranJournal Article20190102This paper proposes a Fuzzy Logic Controller (FLC) strategy for efficient speed control of an Interior Permanent Magnet Synchronous Motor (IPMSM) drive system operating across a wide speed range. The controller design incorporates Maximum Torque Per Ampere (MTPA) control for sub-rated speeds and field weakening control for speeds above the rated threshold. A key advantage of the proposed FLC is its ability to manage both torque and flux simultaneously across the entire speed spectrum, enhancing control precision and motor performance. The system's performance is evaluated for a 40 kW IPMSM using MATLAB/Simulink simulations. The results are benchmarked against a conventional Proportional-Integral (PI) controller under various dynamic conditions, including step changes in reference speed and load torque, single-phase disconnection scenarios, and parameter variations. The simulation outcomes highlight the superior performance of the proposed FLC approach in terms of fast dynamic response, robustness to disturbances and uncertainties, effective rejection of load variations, and elimination of overshoot and undershoot. Additionally, the controller achieves minimal settling time and negligible steady-state error, confirming its effectiveness and reliability for advanced motor drive applications.https://www.tmachineintelligence.ir/article_159731_85b13778db6b4065b75f0d79a727a587.pdfTransactions on Machine Intelligence2821-16932120190301Intelligent Intrusion Detection of Computer Networks Using Random Forest Algorithm485815973210.47176/TMI.2019.48ENS.MoshrefzadehDepartment of Computer Engineering, Faculty of Technology and Engineering, Dana Institute of Higher Education, Yasouj, Iran.O.Rahmani SeryasatDepartment of Electrical Engineering, Faculty of Technology and Engineering, Shams Higher Education Institute, Gargan, Iran.0000-0002-9289-6128B.RavaeiDepartment of Computer Engineering, Faculty of Technology and Engineering, Yasouj University, Yasouj, Iran.Journal Article20181022Intelligent Intrusion Detection Systems (IDS) are pivotal in safeguarding computer networks against unauthorized access and cyber threats. These systems are engineered to detect, identify, and classify potential attacks, while also recognizing security vulnerabilities, thereby enabling timely alerts for network administrators. This study delves into the application of the Random Forest algorithm as the core technique for intelligent intrusion detection. The efficacy of the proposed approach was evaluated using the NSL-KDD dataset, a widely recognized benchmark in intrusion detection research. This dataset comprises 125,973 samples with 41 distinct features representing various network traffic characteristics. The Random Forest algorithm, known for its ensemble-based nature, constructs multiple decision trees during training and outputs the class that is the mode of the classes (classification) of the individual trees. This method enhances predictive accuracy and controls overfitting. Experimental results indicate that the use of this algorithm significantly improves the accuracy of intrusion detection, achieving a remarkable detection rate of 99.89%. These findings underscore the potential of Random Forest in developing intelligent and reliable IDS, offering a robust solution for real-world network security applications. The study also discusses the algorithm's performance in terms of precision, recall, and F1-score, highlighting its effectiveness in various attack scenarios.https://www.tmachineintelligence.ir/article_159732_82ac9e8ca6407115baece9053bf9252b.pdf