The versatility of IoT devices increases the probability of continuous attacks on them. The low processing power and low memory of IoT devices have made it difficult for security analysts to keep records of various attacks performed on these devices during forensic analysis. The forensic analysis estimates how much damage has been done to the devices due to various attacks. In this paper, we have proposed an intelligent forensic analysis mechanism that automatically detects the attack performed on IoT devices using a machine-to-machine (M2M) framework. Further, the M2M framework has been developed using different forensic analysis tools and machine learning to detect the type of attacks. Additionally, the problem of an evidence acquisition (attack on IoT devices) has been resolved by introducing a third-party logging server. Forensic analysis is also performed on logs using forensic server (security onion) to determine the effect and nature of the attacks. The proposed framework incorporates different machine learning (ML) algorithms for the automatic detection of attacks. The performance of these models is measured in terms of accuracy, precision, recall, and F1 score. The results indicate that the decision tree algorithm shows the optimum performance as compared to the other algorithms. Moreover, comprehensive performance analysis and results presented validate the proposed model.

The growing use of nonlinear devices is introducing harmonics in power system networks that result in distortion of current and voltage signals causing damage to power distribution systems. Therefore, in power systems, the elimination of harmonics is of great significance. This paper presents an efficient techno-economical approach to suppress harmonics and improve the power factor in power distribution networks using Shunt Hybrid Active Power Filters (SHAPF) based on neural network algorithms like Artificial Neural Network (ANN), Adaptive Neuro-Fuzzy Inference System (ANFIS), and Recurrent Neural Network (RNN). The objective of the proposed algorithms for SHAPF is to enhance system performance by reducing Total Harmonic Distortion (THD). In our filter design approach, we tested and compared conventional pq0 theory and neural networks to detect the harmonics present in the power system. Moreover, for the regulation of DC supply to the inverter of the SHAPF, the conventional PI controller and neural networks-based controllers are used and compared. The applicability of the proposed filter is tested for three different nonlinear load cases. The simulation results show that the neural networks-based filter control techniques satisfy all international standards with minimum current THD, neutral wire current elimination, and small DC voltage fluctuations for voltage regulation current. Furthermore, the three neural network architectures are tested and compared based on accuracy and computational complexity, with RNN outperforming the rest.

Photovoltaic (PV) system has been extensively used over the last few years because it is a noise-free, clean, and environmentally friendly source of energy. Maximum Power Point (MPP) from the PV energy systems is a challenging task under modules mismatching and partial shading. Up till now, various MPP tracking algorithms have been used for solar PV energy systems. Classical algorithms are simple, fast, and useful in quick tracing the MPP, but restricted to uniform weather conditions. Moreover, these algorithms do not search the Global Maxima (GM) and get stuck on Local Maxima (LM). However, bio-inspired algorithms help find the GM but their main drawback is that they take more time to track the GM. This paper addresses the issue by using the combination of conventional Incremental Conductance (InC) with variable step size and bio-inspired Dragonfly Optimization (DFO) algorithms leading to a hybrid (InC-DFO) technique under multiple weather conditions, for instance, Uniform Irradiance (UI), Partial Shading (PS), and Complex Partial Shading (CPS). To check the robustness of the proposed algorithm, a comparative analysis is done with six already implemented techniques. The results indicate that the proposed technique is simple, efficient with a quicker power tracking capability. Furthermore, it reduces undesired oscillation around the MPP especially, under PS and CPS conditions. The proposed algorithm has the highest efficiencies of 99.93%, 99.88%, 99.92%, and 99.98% for UI, PS1, PS2, and CPS accordingly among all techniques. It has also reduced the settling time of 0.75 s even in the case of the CPS condition. The performance of the suggested method is also verified using real-time data from the Beijing database.

Examination cheating activities like whispering, head movements, hand movements, or hand contact are extensively involved, and the rectitude and worthiness of fair and unbiased examination are prohibited by such cheating activities. The aim of this research is to develop a model to supervise or control unethical activities in real-time examinations. Exam supervision is fallible due to limited human abilities and capacity to handle students in examination centers, and these errors can be reduced with the help of the Automatic Invigilation System. This work presents an automated system for exams invigilation using deep learning approaches i.e., Faster Regional Convolution Neural Network (RCNN). Faster RCNN is an object detection algorithm that is implemented to detect the suspicious activities of students during examinations based on their head movements, and for student identification, MTCNN (Multi-task Cascaded Convolutional Neural Networks) is used for face detection and recognition. The training accuracy of the proposed model is 99.5% and the testing accuracy is 98.5%. The model is fully efficient in detecting and monitoring more than 100 students in one frame during examinations. Different real-time scenarios are considered to evaluate the performance of the Automatic Invigilation System. The proposed invigilation model can be implemented in colleges, universities, and schools to detect and monitor student suspicious activities. Hopefully, through the implementation of the proposed invigilation system, we can prevent and solve the problem of cheating because it is unethical.

Photovoltaic (PV) solar energy is a very promising renewable energy technology, as solar PV systems are less efficient because of climate conditions, temperature, and irradiance change. So, to resolve this problem, two PV topologies are used, i.e., centralized and distributed PV systems. The centralized technique is quicker than the distributed technique in terms of convergence speed and a faster power tracking approach. In the event of uniform irradiance, the centralized system also has the benefit of supplying superior energy, but in PS scenarios, a huge amount of energy is lost. However, the distributed approach requires current and voltage measurements at each panel, resulting in a massive data set. Nevertheless, in the event of shading circumstances, the distributed technique is highly effective because a modular level power electronics (MLPE) converter is used. While in a centralized PV system, there is only a single DC-DC converter for the whole PV system. In this research work, a DFO-based DC-DC converter is designed for modular level, with an ability to perform a rapid shutdown of the module under fire hazard conditions, troubleshooting, and monitoring of a module in a very efficient way. The robustness of the proposed MPPT DFO algorithm is tested with different techniques such as Cuckoo Search (CS), Fruit Fly Optimization (FFO), Particle swarm optimization (PSO), Incremental conductance (InC), and Perturb and observe(P&O) techniques. The proposed technique shows better results in terms of MPPT efficiency, dynamic responsiveness, and harmonics. Furthermore, the result of MLPE and the centralized system is verified by using the Helioscope with different inverter companies like SMA, Tigo, Enphase, Solar edge, and Huawei. The results prove that MLPE is a better option in the case of shading region for attaining the maximum power point.

Patients with cognitive difficulties and impairments must be given innovative wheelchair systems to ease navigation and safety in today’s technologically evolving environment. This study presents a novel system developed to convert a manual wheelchair into an electric wheelchair. A portable kit has been designed so that it may install on any manual wheelchair with minor structural changes to convert it into an electric wheelchair. The multiple modes include the Joystick module, android app control, and voice control to provide multiple features to multiple disabled people. The proposed system includes a cloud-based data conversion model for health sensor data to display on an android application for easy access for the caretaker. A novel arrangement of sensors has been applied according to the accurate human body weight distribution in a sitting position that has greatly enhanced the accuracy of the applied model. Furthermore, seven different machine learning algorithms are applied to compare the accuracy, i.e., KNN, SVM, logistic regression, decision tree, random forest, XG Boost, and NN. The proposed system uses force-sensitive resistance (FSR) sensors with prescribed algorithms incorporated into wheelchair seats to detect users’ real-time sitting positions to avoid diseases, such as pressure ulcers and bed sores. Individuals who use wheelchairs are more likely to develop pressure ulcers if they remain in an inappropriate posture for an extended period because the blood supply to specific parts of their skin is cut off owing to increased pressure. Two FSR configurations are tested using seven algorithms of machine learning techniques to discover the optimal fit for a high-efficiency and high-accuracy posture detection system. Additionally, an obstacle detection facility enables one to drive safely in unknown and dynamic environments. An android application is also designed to provide users of wheelchairs with the ease of selecting the mode of operation of the wheelchair and displaying real-time posture and health status to the user or caretaker.