Transactions on Machine Intelligence2821-16934220210601Detection of Pavement Damage Using Smartphones546220920910.47176/TMI.2021.54ENA.NouriM.Sc. in Civil Engineering, Department of Civil Engineering, Sharif University of Technology, Tehran, IranV.ToufighPh.D. in Civil Engineering and Associate Professor, Department of Civil Engineering, Sharif University of Technology, Tehran, IranJournal Article20210304Early detection and repair of pavement damage can significantly reduce the associated maintenance and repair costs. In recent years, efforts have been made to use digital hardware and software to streamline the inspection and diagnostic process. However, in a real-world scenario, the resource limitations, high cost, and time-consuming nature of these digital units have diminished their efficiency. In the past decade, smartphones have gained remarkable hardware capabilities. Mobile phones, aided by GPS, record location information and capture high-quality images with powerful lenses. This research aims to utilize machine learning algorithms to employ smartphones for pavement inspection. Deep learning, a method capable of pattern recognition and solving complex problems, has been chosen as the machine learning technique. The learning process of this algorithm is conducted using samples collected from pavement surfaces via smartphones. The study further aims to enhance the speed and processing power of the learning method through new parameters. Additionally, a defined framework is provided to assess the quality of damage, facilitating effective action by route managers.https://www.tmachineintelligence.ir/article_209209_38b3019166a8cc1b23d57e227823f731.pdfTransactions on Machine Intelligence2821-16934220210601Predicting Survival of Leukemia Patients Using a Support Vector Machine Based on the Bowerbird Algorithm636920921010.47176/TMI.2021.63ENA.MarousiDepartment of Computer Engineering, Khavaran Non-Profit Higher Education Institution, Mashhad, IranF.MarousiDepartment of Computer Engineering, Gonabad Higher Education Complex, Gonabad, Iran; Computer Instructor, Khorasan Razavi Department of EducationJournal Article20210306Unfortunately, research has shown that cancer incidence has been increasing in recent years. Leukemia is a type of blood cancer caused by an increase in the number of white blood cells. Generally, any type of blood cancer is extremely dangerous, and in most cases, there is no cure. Acute myeloid leukemia (AML) is a common and fatal type of this cancer. Predicting survival after diagnosis is one of the key indicators for evaluating treatment methods, which is the focus of the present study. In this research, we used a combination of Support Vector Machine (SVM) with the Bowerbird Algorithm to analyze survival status and predict mortality in patients. The data used pertains to patient information from Seyyed al-Shohada Hospital in Isfahan, with 197 samples and 9 features. MATLAB software was used to run the programs. Evaluation was based on diagnostic indices including sensitivity, specificity, and accuracy. The proposed SVM based on the Bowerbird Algorithm achieved a performance of 69.57% accuracy, 75.52% sensitivity, and 64.48% specificity, outperforming the combination of SVM with other optimization algorithms such as Cuckoo Search, Harmony Search, and Firefly Algorithm. Therefore, the proposed method is a promising tool for predicting survival in leukemia patients with improved diagnostic accuracy.https://www.tmachineintelligence.ir/article_209210_d11dc58a63bb29c3eeec83b639e86c45.pdfTransactions on Machine Intelligence2821-16934220210601Speech Signal Encryption in Time and Frequency Domains Using Chaotic Maps707520921110.47176/TMI.2021.70ENJ.ShiraziAssistant Professor, Department of Electrical Engineering, Gonabad Branch, Islamic Azad University, Razavi Khorasan, Gonabad, IranJournal Article20210224In this paper, with the aim of enhancing the security of speech signal transmission, a method based on encrypting speech signal frames in both time and frequency domains using a chaotic map is presented. To increase the complexity of the encryption, two stages of encryption in the time domain and one stage in the frequency domain were performed on the speech signal. In the first stage, the bytes of the speech signal were altered using random numbers generated from a chaotic map, and in the second stage, the resulting samples were scrambled based on random numbers. Discrete Cosine Transform (DCT) was then applied to the resulting samples, followed by scrambling based on random numbers. To further enhance encryption security, the sequence of random numbers used was periodically altered according to a specific order to increase the complexity of the encryption detection. Various criteria were used to evaluate the employed method, and the results indicate high levels of these criteria for the method. One advantage of the used method is its complexity due to combining time and frequency domains and the multi-stage nature of the encryption.https://www.tmachineintelligence.ir/article_209211_948ba442b7ebe1152b2dcee8637201a4.pdfTransactions on Machine Intelligence2821-16934220210601Energy Management in Distribution Network with the Presence of Electric Vehicles and Energy Storage Systems Using the Crow Search Optimization Algorithm768920921210.47176/TMI.2021.76ENP.RadmehrDepartment of Electrical Engineering, Faculty of Electrical and Computer Engineering, Yadegar-e-Imam Khomeini (RAH) Shahr-e-Rey Branch, Islamic Azad University, Tehran, IranA.ShahbaziDepartment of Electrical Engineering, Faculty of Electrical and Computer Engineering, Yadegar-e-Imam Khomeini (RAH) Shahr-e-Rey Branch, Islamic Azad University, Tehran, IranA.AhmarinezhadAssistant Professor, Department of Electrical Engineering, Faculty of Technical and Engineering, Central Tehran Branch, Islamic Azad University, Tehran, IranJournal Article20210329Considering the growth in electrical energy consumption and consequently the increasing investment in the distribution sector, as well as the significant portion of losses in the entire distribution network system, operators have been compelled to propose optimal solutions to mitigate these issues. Energy management aimed at reducing consumption and demand is an effective method for load management under peak load conditions and reducing energy consumption, a method that has been employed by several electricity companies for years. Studies conducted on this method are either based on network modeling, which requires precise information about the real-time status of network loads and the percentage contribution of various loads to the total load, or based on data received from measurement devices installed for consumers. In this paper, plug-in hybrid electric vehicles (PHEVs) combined with the integration of renewable energy systems into the power grid offer a promising method to address environmental problems. To this end, a multi-objective algorithm is proposed to optimally locate several renewable energy systems (RES), including parking lots for PHEVs in a distribution system. The proposed algorithm determines the number, locations, and sizes of RES and parking lots. The objective of the proposed algorithm is to minimize the total energy cost of the system. The problem is formulated as an optimization problem solved using the Crow Search Algorithm (CSA), considering power system and PHEV operational constraints. The proposed algorithm is tested on a standard 33-bus distribution system, and the obtained results demonstrate the effectiveness of the proposed method. https://www.tmachineintelligence.ir/article_209212_dac2e15187c48d25d470cfaf6d9acb19.pdfTransactions on Machine Intelligence2821-16934220210601Position Control of Permanent Magnet DC Motor Using IMC-Based PID Controller909920921310.47176/TMI.2021.90ENN.TalebiDepartment of Electrical and Computer Engineering, Yadegar-e-Imam Khomeini (RAH) Shahre Rey Branch, Islamic Azad University, Tehran, IranJournal Article20210425The position control of permanent magnet DC motors (PMDCMs) has gained significant attention in recent years, particularly with the advent of advanced control strategies such as Internal Model Control (IMC) combined with Proportional-Integral-Derivative (PID) controllers. This literature review synthesizes various research findings that explore the effectiveness of IMC-based PID controllers for position control applications, while also identifying knowledge gaps and suggesting future research directions. Today, permanent magnet DC motors hold a significant position in theoretical and industrial applications due to their simplicity in implementation and control. PID controllers, renowned as one of the most popular controllers in the industry, continue to maintain their esteemed status. This paper aims to leverage the robust capabilities of the Internal Model Control (IMC) method to develop a PID controller to regulate the position of permanent magnet DC motors. In the proposed controller, the proportional (kP), derivative (kD), and integral (kI) gains are determined by IMC. Simulation results demonstrate that the proposed design outperforms traditionally tuned PID controllers and is capable of effectively controlling the motor position under various conditions.https://www.tmachineintelligence.ir/article_209213_67311b561e26dacab1ebeaf2ee2b39b5.pdfTransactions on Machine Intelligence2821-16934220210601Image Encryption Using the TLBO Algorithm and Image Hash10010820921410.47176/TMI.2021.100ENM.AbedzadehDepartment of Computer Engineering, Faculty of Engineering, Shahid Bahonar University, Kerman, IranM. J.RostamiAssistant Professor, Department of Computer Engineering, Shahid Bahonar University, Kerman, IranJournal Article20210222An effective encryption algorithm must not only provide fast encryption but also be resistant to various attacks. Chaotic mappings are widely used in image encryption. Furthermore, since the hash of an image produces a distinct output for each image and changes drastically with even a single bit alteration, it can be an ideal candidate for the initial values in chaotic mappings. This paper explores the encryption of images inspired by the Teaching-Learning-Based Optimization (TLBO) algorithm, which is a metaheuristic algorithm. Initially, the image hash is computed using the SHA-512 hash function, and numbers between 0 and 1 are generated based on the obtained hash. The image is then divided into 16 equal parts. In the next step, the pixels of the image are permuted using a specific variant of the standard map proposed in this paper. In each iteration, the part with the best entropy is selected as the teacher. Among the generated values, the one that improves the entropy of that part is chosen, and other parts follow this value. Finally, the pixels of each part are altered using the logistic map. This algorithm offers both adequate execution time and high security. https://www.tmachineintelligence.ir/article_209214_22a2c0917cd80eb7e0aa746ad02b31f7.pdf