Rapid climate change and global warming have accelerated the shift from traditional to alternative power sources. Solar Photovoltaic system (SPV) is most attractive among Renewable energy sources. In order to mitigate the intermittency and unreliability, Hybrid Renewable Power Generation (HRPG) is developed. Maximum Power Point Tracking (MPPT) is desirable for enhancing the performance of solar-based HRPG systems. As far as convenience and popularity are concerned, many solar-based hybrid power generation systems are employed. An MPPT algorithm is developed in solar hybrid power generation for enhancing the tracking efficiency. In this research, the suitability of the Incremental Conductance Algorithm (ICA) is inspected. Also, the performance parameters of overall solar hybrid power generation with the ICA MPPT algorithm are determined using MATLAB/Simulink environment.
The evolution of e-commerce accompanied with enormous amount of sentiment reviews drive practical difficulty for consumers to decide upon a purchase choice. Sentiment analysis assists in analysing the reviews for a consumer to decide a purchase choice and a manufacturer to follow up on their clients and manage their requirements. Aspect based sentiment analysis is a subtask of sentiment research that aims to identify and analyse sentiment towards specific aspects or features of a target entity within a given text. Aspect extraction from the text input reviews is another important function in sentiment analysis at aspect level. Identifying those aspects and selecting required aspects is very essential task in opinion analysis based on features. In the proposed analysis, we initiate a supervised machine learning technique-decision tree classifier for selecting the required aspects and sentiment research is conducted by utilizing various machine learning techniques. The proposed work gives an accuracy of 85 - 93%.
Worldwide, Type II Diabetes Mellitus (T2DM) is a major condition whose incidence is rising. It is a public health issue as well as the silent killer of humanity. Its diverse aetiology, which includes genetic predisposition and environmental influences, may offer a distinctive trait in the population that is useful for early diagnosis. As a result, researchers are seeking novel approaches to its early diagnosis and treatment. Now, manual methods are used to measure and compare the axial triradius (ATD) angles of both palms using a Goniometer for the early prediction of T2DM. The creation of datasets containing individuals' left and right palm prints is the primary objective of this project, which aims to develop a model that would assist in the early prediction of T2DM. The model would have the option to track down the distinction in the ATD points in the two palms and foresee assuming that the singular will be determined to have T2DM in the future. A total of 390 instances were collected out of which 130 are diabetic and 260 were non-diabetic patients. It consists of various image pre-processing techniques and delta-point detection methods. The model would be able to predict the results if the user simply uploaded images of both palms. The Image Thresholding and Segmentation Method was used to evaluate the best features and the crossing number method was used to automate the model, indicating that the model achieved an accuracy of 91.02% in predicting the ATD angle. Using biomarkers, this project can be implemented in the health sector for early detection of Type II Diabetes Mellitus.
The detection and analysis of cardiovascular diseases is a critical study field within the medical field. Cardiovascular diseases are one of the leading causes of mortality globally, and timely identification is vital to forego adverse effects and improve outcomes for patients with these conditions. The incorporation of machine learning technologies to forecast onset of cardiovascular diseases and analyzing their progression has shown promising results. This paper demonstrates the machine learning models that can provide near-accurate predictions and help identify the right data set based on accuracy. The choice of the best-fitting algorithm has been made by a comparative study and with the inclusion of K-fold cross-validation. We infer that the Random Forest algorithm is the best algorithm for detection and K-Nearest Neighbor Algorithm for the analysis of prominent heart diseases like Stable Angina, ST-elevated myocardial infarction, and non-ST elevated myocardial infarction which will serve as a direction to further diagnosis. Overall, the potential for machine learning to improve early detection and treatment of Cardiovascular Diseases, which could ultimately save lives and reduce healthcare costs, has been recognized. Index Terms—Artificial intelligence, Machine learning, Supervised learning, Unsupervised learning, Accuracy
Drones are a big advancement in precision agriculture and crop management based on GPS and big data as a method to boost crop productivity while addressing water and food shortages. According to studies, there has been an increased demand for systems which deal with accurate spraying of palms and on-board disease detection. This research presents the simultaneous localisation and mapping (SLAM) of palm plantations, and their geolocation using YOLOv4 detector. Autonomous Sequential Decision Process (SDP) is carried out to navigate using the Pixhawk’s inbuilt PX4 system that uses the palms detected through the camera as targets and navigates around them and successively to the next palm. Images captured by the drone are scanned for any signs of yellow leaf disease or rot by learning vector quantization (LVQ) and K-Means Clustering within the RGB channels. The key interest of this research is the trajectorial nozzle system which tracks the arecanut and actively adjusts the spray to match the arecanauts’ position.
The field of reconfigurable antennas has recently received major interest from the research community. The Origami based reconfigurable antennas provide enhanced capabilities to advanced applications. However, it is essential to compare performance of different types of Origami antennas to understand the specific benefits. In the current article, the performance of an Origami helical antenna and an Origami curlicue antenna is compared. Both the antennas are designed such that they are reconfigurable in nature, and for the antenna construction, aluminium is used as the material. Extensive experiments are conducted for validating and comparing the performance of the designed antennas. The experiment results demonstrate that the antennas operate between frequency range of 0.85-1.3 GHz. Further, the helical antenna operates appropriately in unfolded state whereas, the curlicue antenna operates efficiently in both, folded state and unfolded state.
In recent years the prominence of food product traceability has significantly increased due to growing concerns regarding food safety and ensuring equitable market access for farmers. Presently, supply chain information systems collect, transmit and store records in centralized systems. These conventional digital systems create fragmented data silos vulnerable to a single point of failure. These limitations pose significant challenges in establishing trust factors such as transparency, traceability, integrity, and privacy among the various actors involved in the agri-supply chain. To overcome these challenges, a decentralized technology like blockchain has immense potential to enhance traceability by addressing security and privacy aspects. This study proposes a blockchain-based framework called AgriFoodCredChain for food product traceability using Decentralized Identifiers (DIDs) and Verifiable Credentials (VCs). Initially, the study outlines a secure mechanism for establishing DID communication channels, assuring robust and reliable data exchange. Furthermore, methods for privacy-preserved issuance and verification of credentials for food traceability are developed and presented. The implementation of the framework is conducted utilizing the identity blockchain named Hyperledger Indy and Aries library, considering particular workflow in the food supply chain. Evaluating the framework w.r.t credential issue and verifying time proves that the solution implemented is scalable and satisfies the requirements for real-time adoption.
Lung nodules are small, round, or oval-shaped growths that can develop in the lungs. They are typically detected as abnormal spots or lesions on medical imaging scans, such as chest X-rays or computed tomography (CT) scans. Malignant lung nodules have the potential to be dangerous as they can be an early sign of lung cancer, one of the greatest causes of cancer-related deaths worldwide. Detecting and localizing these nodules accurately is crucial for early prognosis and effective treatment. Timely identification of malignant nodules can significantly improve patient outcomes by enabling early intervention and potentially increasing the chances of successful treatment. A Graph Convolutional Network (GCN) model was used to create an effective lung nodule identification and localization method in order to solve this problem. The LUNA-16 dataset, which has about 888 datasets, was used to train and test the GCN model. The GCN model was trained on a subset of 90 datasets and obtained an overall accuracy of 84% by utilising deep learning and graph-based representations. When categorising benign and malignant nodules, the model displayed clear performance variations. It demonstrated its capability to successfully distinguish non-cancerous nodules from potentially hazardous ones by achieving an accuracy of 60% for benign lung nodules. The GCN model also demonstrated excellent precision in diagnosing cancerous lung nodules, with an accuracy of 94% for malignant nodules. These findings highlight the potential of the GCN model in assisting medical professionals with the detection and localization of lung nodules. The promising results obtained in this research pave the way for further advancements in lung nodule analysis, contributing to the early treatment of lung cancer. By accurately identifying and localizing lung nodules, medical practitioners can make more informed decisions regarding patient care.
This manuscript presents a novel approach for accurately recognizing Indian sign language alphabets and numbers (1-9), which play a vital role in communication for individuals with hearing impairments in India. We conducted extensive experiments on our self-created dataset named "ISL 2022: A novel dataset creation on Indian sign language". This dataset contains images with non-uniform and complex backgrounds, reflecting the real-world scenarios where sign language gestures are performed. After identifying the region of interest i.e. the hand region from the images, we extract discriminative features using global image descriptors and Hu moments. Global image descriptors capture the global spatial layout of the gesture, providing valuable information about its overall structure and appearance. On the other hand, Hu moments capture shape characteristics essential for differentiating between different gestures. We employ the Random Forest and Naive Bayes algorithms to classify the gestures. The experimental results demonstrate the effectiveness and recognition rate of 98%. By combining the features of both the feature extractors, our approach showcases promising advancements in recognizing ISL gestures, especially considering the challenges posed by diverse backgrounds. The proposed methodology has the potential to contribute significantly to the development of practical and efficient sign language recognition systems.
Paraphrase detection, a crucial task in natural language processing, involves determining if two given sentences convey the same meaning. This research paper explores the advancements made by Microsoft Research in paraphrase detection using a range of machine learning models, ensemble techniques, and fine-tuned hyper parameters. The paper then presents the significant contributions of Microsoft Research, which involve employing a diverse set of machine learning models and leveraging ensemble techniques to enhance the accuracy and robustness of paraphrase detection. Additionally, the use of grid search cross-validation for hyper parameter optimization is explored to fine-tune the models' performance. The BERT (Bidirectional Encoder Representations from Transformers) model was also trained for paraphrase detection using the ktrain library. Experimental setups, including datasets, evaluation metrics, and preprocessing steps, are described in detail. Results and analysis showcase the effectiveness of the comparative analysis, demonstrating significant improvements over existing methods. Overall, this research paper provides valuable insights into the advancements achieved by Microsoft Research in paraphrase detection through the utilization of diverse machine learning models, ensemble techniques, and fine-tuned hyper parameters, thus driving the progress in natural language processing tasks.
This research article focuses on the detection of the mature stage of Aspergillus flavus fungus using an automated approach. Aspergillus flavus is a pathogenic fungus known for causing diseases in crops and producing carcinogenic mycotoxins. Detecting the mature stage is crucial for optimizing the production of antibiotics such as Aspergillic acid, which is abundantly produced during this phase. Manual color detection methods are labour-intensive and prone to errors. Existing techniques face challenges due to lighting conditions and background issues. To overcome these limitations, we propose a two-part automated process that involves background elimination using comparative thresholding techniques and mapping the mature phase on the Lab* scale. This approach eliminates lighting conditions and enhances accuracy in identifying the mature phase. By automating the detection process, our method offers a significant advancement in the production of antibiotics and reduces the reliance on manual intervention. The recommended methodology has the competence to upgrade efficiency in the pharmaceutical industry and educational research on fungal growth.
This paper presents the integration of using ML models with fault simulation to reduce the test volume in VLSI testing. It introduces a novel approach to VLSI testing, focusing on reducing the test volume while maintaining high test coverage. It has an upper edge when compared with the traditional testing as the test cost reduces drastically. Experimental results show an average of 30% test pattern reduction is possible
Cyclone is a meteorological phenomenon characterized by large-scale rotating and converging air mass around a center of low atmospheric pressure. They cause severe damage to the ecosystem and the global economy. Finding the direction that a cyclone is moving and how that affects the places where it causes damage to people and property is one of the elements that is crucial. The remote sensing community has long been interested in image classification to reduce injuries. To solve current issues, a different approach to predicting tropical cyclone intensity from satellite images has been used in many previous models on cyclone prediction to improve accuracy. EPF is prone to errors and results in serious catastrophes. The proposed model would Estimate, Predict and Forecast the TC using the collected satellite images. As a result, the suggested machine learning model contains techniques (EPF) for forecasting and cyclone strength identification.
Multilevel Inverters are preferred in low and medium power applications. Multilevel inverters have advantage of ease of control, and therefore they are widely used in Industries and Electric Vehicles. The output voltage of these inverters contains too many harmonics which causes heating and ultimately breakdown of the entire system and that leads to huge financial losses. The work carried out here describes a method to reduce the magnitude of lower order harmonics using a technique called Selective Harmonic Elimination (SHE). Genetic Algorithm (GA) is applied to calculate the firing angles for the switches in the single phase 7-level inverter. The behavior of the inverter is tested on a single phase Induction motor.
This paper studies the performance of outage probability in a dual-stage (DS) cooperative non-orthogonal multiple access (NOMA) system for power line communication (PLC) in the presence of Bernoulli-Gaussian impulsive noise. The network configuration includes a single source node and two destination nodes among which one the node act as a decode-and-forward (DF) relay node. This is similar to the existing research on DS NOMA setup, which primarily focused on additive white Gaussian noise. In this setup, both the source and relay transmit data to the destination node using NOMA across two consecutive time frames. We provide mathematical expressions for outage probabilities at both nodes in the context of Bernoulli-Gaussian noise. Furthermore, our analysis is validated using numerical methods and Monte Carlo simulations. Additionally, our results indicate that a DS-NOMA configuration exhibits lower outage probability when compared to a single-stage NOMA setup.
Telecommunication and computing technologies have become instrumental in solving various real-world problems. One pressing issue that developing countries still face is the safety of railway level crossings. These crossings pose significant risks, including accidents resulting in fatalities, injuries, and vehicle damage, all of which have detrimental effects on a country's economy. To address this challenge, the adoption of IoT technologies in the railway sector holds great potential. Previous research has proposed automated systems employing different IoT approaches, but upon thorough literature analysis, it is evident that these studies have their limitations. To overcome these shortcomings, this paper presents a novel system that combines IoT and machine learning concepts to create an efficient and effective solution for reducing fatal accidents at level cro Additionally, the system sends notifications to users, indicating the nearest railway station and suggesting alternative routes to reduce travel time based on latitude and longitude data.The suggested solution uses IoT to establish an intelligent and automated traffic control system at railway crossings. The solution features a centralized control unit that takes data from IoT devices and optimizes traffic management using complex algorithms and Google database approaches. movements, traffic density, and pedestrian safety. This ensures efficient traffic flow, reduces waiting times, and enhances overall safety at railway crossings.
Due to limited reserve capacity, shoot up of generation prices and environmental concerns, interest in renewable energy is growing as a substitute energy source to traditional fossil fuels. To excerpt extreme power output from a photovoltaic scheme an effective "Maximum Power Point Tracking" System is essential. The global research community is investigating all options for efficient energy conversion from freely available renewable energy sources. A comparative analysis for four MPPT algorithm has been carried out and a new hybrid method is proposed which involves fuzzy logic regulator and perturb and observe technique. All the algorithms have been investigated in MATLAB SIMULINK environment and a detail study have been carried out. The proposed technique precisely senses the MPP and can quickly trail it in existence of irradiance variations.
With the fast spreading of the Monkeypox disease worldwide, it would become an alarming situation for countries around the globe. The case might become worsen, leading to another global pandemic. Therefore, early detection and analysis of Monkeypox from various points of view become necessary. This paper provides a detailed study of the recent spreading of the Monkeypox disease. The study involves multiple perspectives like information about Monkeypox disease, its origin, general symptoms that a person affected with it can experience, its global impact, and recent developments in the state-of-the-art.
Individuals with disabilities face unique challenges in both their daily lives and education. As society becomes increasingly complex, visually impaired and physically disabled students may find it challenging to keep up. Therefore, it is essential to understand the eeds of these students and explore how technology can help them. Assistive technologies, including both hardware and software, can help individuals overcome barriers obtain better results in life and academic achievement. One example of an assistive technology is a voice-activated plotter that can recognize speech using a Bluetooth app and transcribe spoken content onto paper or other writing materials. Such a tool can be especially helpful for individuals with visual or physical impairments who may have difficulty writing or typing. However, creating an accurate and accessible voice-activated plotter can presentsome design challenges that need to be addressed
Humanoid behavior of robots has resulted in significant development in many fields, including medical science and automation. Understanding Human motions and movements plays a crucial role in designing such humanoid robots. Technological research development can apply human actions in significant fields like humanoid robots, medical analysis and assistance systems, and sports analytics. This paper presents a sensory system that interacts and reads the movement and orientation of individual arms and is used to control robotic arms. The proposed system is a 7 Degrees of Freedom (DoF) sensory system with a 7 DoF robotic arm miming the human transhumeral arm. A wearable sensor plays a vital role as the data collected can be used in numerous ways, from training robots and other modules, to analyzing the performance of sports personnel. The sensory band collects data to provide each joint's rotation angle. The sensory system was built using IMU sensors, flex sensors, and a potentiometer with Arduino Uno as the primary controller. The acquired sensory data was communicated to a robotic arm using an RF transceiver setup. This system can be inexpensive and easily used in various fields, including sports analytics in rural areas. The experimental results are promising, providing a linear connection between the sensory system and the robotic arm with average error for the joints below 2%. Incorporating higher-range sensors can improve the performance of the proposed model.
