Sampath Kumar S
@sece.ac.in
Assistant Professor (Sl. Gr) and Department of Computer Science and Engineering
Sri Eshwar College of Engineering
RESEARCH INTERESTS
Blockchain, Drone, Deep Learning, Data Science
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
Kavitha S., Dileep M. R., Sampath Kumar S., Mohammad Shahid, P. Hemachandu, and S. Kaliappan
IGI Global
This study studies the implementation of machine learning (ML) algorithms to improve power distribution in an industrial context, concentrating on the essential issue of anticipating energy consumption. Various ML models, including Support Vector Machine (SVM), Artificial Neural Network (ANN), Decision Trees (DT), and Random Forests (RF), were extensively examined and compared for their usefulness in anticipating demand patterns within a sector encompassing machining, forging, CNC, and packaging stations. The models revealed various strengths, with SVM leading with an accuracy of 95.6%, closely followed by ANN at 94.33%, while DT and RF displayed accuracies of 87.6% and 85.6%, respectively. The research additionally gives a thorough comparison of actual vs expected demand levels over hourly intervals, illustrating the models' responsiveness to dynamic energy use patterns throughout the day.
Gautam Solaimalai, Kode Jaya Prakash, Sampath Kumar S, A Bhagyalakshmi, P Siddharthan, and K. R Senthil Kumar
IEEE
This exploration paper investigates the operation of deep underpinning literacy(DRL) for enabling independent drone navigation in cluttered surroundings. Navigating drones in cluttered spaces poses significant challenges due to the presence of obstacles and dynamic environmental conditions. Traditional navigation approaches frequently struggle to acclimatize to these complications. In this study, we propose a new frame using DRL ways to enable drones to autonomously navigate through cluttered surroundings while avoiding obstacles. The frame employs a deep neural network to learn a policy that guides the drone’s conduct grounded on environmental compliances. Through expansive simulations and real-world trials, we demonstrate the efficacity of the proposed approach in achieving robust and adaptive drone navigation in cluttered surroundings. The findings of this exploration have significant counteraccusations for colorful operations, including hunt and deliverance operations, surveillance, and package delivery, where independent drone navigation in cluttered spaces is pivotal.
Md Hussain Ansari, Balusamy Nachiappan, Sampath Kumar S, Ardly Melba Reena B, Shankar Nagarajan, and Jonnadula Narasimharao
IEEE
The purpose of this exploration composition is to probe the use of inheritable algorithms(GAs) for intelligent resource operation in cloud computing settings. The optimization of resource allocation and operation is becoming an increasingly delicate task as cloud computing continues to expand in both complexity and size. The operation of inheritable algorithms, which are deduced from natural selection and the generalities of genetics, gives a promising strategy for addressing this difficulty. inheritable algorithms are suitable to efficiently search and use the result space to gain near-optimal resource allocation ways. This is fulfilled by iteratively evolving a population of seeker results. In this exploration, we probe how inheritable algorithms(GAs) can be employed to perform tasks in cloud surroundings, similar to the placement of virtual machines, the scheduling of workloads, and the provisioning of coffers. This paper investigates the efficacity and scalability of GA-grounded resource operation strategies in cloud computing systems, with the thing of enhancing performance, resource application, and energy effectiveness. This is fulfilled by conducting a conflation of literature and case studies. One of the ultimate pretensions of this exploration is to donate to the development of resource operation results that are both intelligent and adaptive, and that can meet the ever-changing conditions of cloud computing surroundings.
Gautam Solaimalai, J. Maria Shanthi, Sampath Kumar S, Priyanka Khabiya, K. Geetha, and Gokul Talele
IEEE
The proliferation of virtual sidekicks in colorful disciplines has prodded a swell of interest in advancing Natural Language Processing(NLP) ways to enhance their effectiveness. This paper provides a comprehensive review of the current approaches and challenges encountered in integrating NLP into virtual sidekicks. It begins by outlining the foundational generalities of NLP and its vital part in enabling mortal- suchlike relations with virtual sidekicks. latterly, it delves into a discussion of the different methodologies employed in NLP for understanding and generating mortal language, ranging from rule-grounded systems to deep literacy models. likewise, the paper highlights the crucial challenges such as nebulosity resolution, environment understanding, and handling different verbal variations that stymie the flawless functioning of virtual sidekicks. By synthesizing perceptivity from recent exploration trials, this paper offers precious perspectives on the state-of-the-art NLP ways and directions for unborn exploration to overcome the challenges and propel the development of further intelligent and intuitive virtual sidekicks.
Kowshika M, Ooviya M, Pavithradevi B, Rashika K V, and Sampath Kumar S
IEEE
Search and Rescue (SAR) operations are critical endeavors in disaster management and public safety, often involving high-risk situations where human lives are at stake. Traditionally, these operations have heavily relied on human resources and ground-based technologies, which can be time consuming, resource-intensive, and constrained by various limitations. In recent years, the integration of drone technology into SAR operations has revolutionized the field, offering a powerful tool to enhance the efficiency, effectiveness, and safety of rescue missions. This study provides an in-depth exploration of the role of drones in human search and rescue operations. It examines the various facets of this technology, including hardware, software, and operational considerations, and highlights the manifold benefits it brings to SAR missions. The most important obligation during a natural disaster is to locate and rescue any trapped individuals as soon as conceivable. Unmanned aerial vehicles (UAVs) have seen a surge in usage recently due to their excellent durability, adaptability, cheap cost, and simplicity of implementation. This article includes a fresh thermal imaging dataset that was obtained by drones. The study discovers the best observers to prune and fine-tune the survivor detection network based on the sensitivity of the convolutional layer due to the restricted computational power and memory of the microprocessor.
Surendra Reddy Vinta, Ashok Kumar Koshariya, Sampath Kumar S, Aditya, and Annantharao Gottimukkala
European Alliance for Innovation n.o.
Despite rapid population growth, agriculture feeds everyone. To feed the people, agriculture must detect plant illnesses early. Predicting crop diseases early is unfortunate. The publication educates farmers about cutting-edge plant leaf disease-reduction strategies. Since tomato is a readily accessible vegetable, machine learning and image processing with an accurate algorithm are used to identify tomato leaf illnesses. This study examines disordered tomato leaf samples. Based on early signs, farmers may quickly identify tomato leaf problem samples. Histogram Equalization improves tomato leaf samples after re sizing them to 256 × 256 pixels. K-means clustering divides data space into Voronoi cells. Contour tracing extracts leaf sample boundaries. Discrete Wavelet Transform, Principal Component Analysis, and Grey Level Co-occurrence Matrix retrieve leaf sample information.
V Sathya Priya, Harshit Sharma, Neeraj Kumar Singh, Nilesh Dubey, Sampath Kumar S, and Prabhishek Singh
IEEE
In cloud computing is a sort of registering which can be considered as another period of processing. Cloud can be considered as a quickly arising new worldview for conveying figuring as a utility. High-level records technological know-how (IT) security, or cybersecurity, concerns like unauthorized records publicity and leaks, vulnerable get entry to controls, susceptibility to attacks, and availability disruptions affect regular IT and cloud structures alike. In any case, due to the accessibility of limited assets it is truly challenging for cloud suppliers to offer every one of the requested types of assistance. IoT is the between-systems administration of substantial gadgets, vehicles, developments, and various contraptions implanted with hardware, programming, sensors, actuators, and local area availability that permit these items to assemble and change data. Advances to the modern net will be sped up through extended local area dexterity, worked in manufactured cerebrum (AI), and the capacity to convey, robotize, organize, and firmly shut different use cases at hyper-scale.
S. Sampath Kumar, Kumar V. Ajay, Nataraj S. Arun, B. Devasarathy, and B. Hariharan
Trans Tech Publications Ltd
There is a communication lag between deaf-mutes and normal people. To overcomethat, we are providing information access and services to deaf-mute people in Indian Sign Language (ISL) and developing a flexible project that can be enlarged to capture the entire lexicon of Indian Sign Language via physical gestures like hand expressions and non-manual signs like facial expressions by developing and building a training model using machine learning algorithms. Sign language recognition uses image-based manual and non-manual gestures. Here we used figure recognition to identify manual and non-manual gestures. Finding expression gestures and analyzing finger movements to determine what the deaf-dumb individual is saying. In Python, the MediaPipe recognizes the hand signs and facial gestures of a person. These modules were developed to assist people with non-identical motions. This paper presents figure identification of Indian Sign Language via hand and facial gestures, as well asits integration with a chatbot as transcript output.
S. Sampath Kumar, M. Prithiv, S. Ravi Kumar, K. Senthil Kumar, and K. Vishnu
Trans Tech Publications Ltd
Forswearing of Service (DoS) and Denial of Service (DDoS) assaults are not kidding strings to the Internet. The recurrence of DoS and DDoS assaults is expanding step by step. Computerized instruments are additionally accessible that empower non-specialized individuals to carry out such assaults without any problem. Consequently, it isn't simply essential to forestall such assaults, yet additionally need to traceback the aggressors. Following back the wellspring of the assaults, which is known as an IP traceback issue is a difficult issue due to the stateless idea of the Internet and parodied Internet Protocol (IP) bundles. Various Internet Protocol traceback strategies present, yet they are have constraints like quantity of bundles needed or capacity computational workloads brought about at switches. A thought of an improved bundle stamping and traceback calculation for IP traceback to recognize an aggressor that works with the traceback of the parodied parcel to its starting point is proposed. The procedure proposed decreases computational time.
D. Vinod, Karimulla Syed, Sampath Kumar S, B. Sankara Babu, Shatrudhan Pandey, and Raja Velur Loganathan
IEEE
Throughout the cutting operation, among the most crucial steps is that has a direct impact on machining accuracy and product quality is tool wear. If you can accurately forecast how quickly tools will wear out, you can make the necessary adjustments early on, minimizing downtime and maximizing product quality. The evaluation of cutting tool wears during the production stage is crucial. The main objective of this study is to track tool degradation over time to guarantee regular tool changes and prevent workpiece waste and potential machine damage from excessive tool wear or unexpected tool breakage. For this reason, it is necessary to develop a system that is both intelligent and capable of providing accurate solutions to these issues. The criteria of intelligent production are great, though, and traditional methods simply cannot match them. Because of this, a deep learning-based innovative approach is presented to advance the precision of tool wear prediction. Parallel convolutional neural networks were constructed and used to achieve multi-scale feature fusion. To improve the model's efficiency, the channel attention mechanism was integrated with the residual connection to take into account the relative importance of each feature map. To prove that the created approach is superior, several experiments were carried out to anticipate tool wear, and the findings are more reliable and accurate than those obtained using existing methods. To guarantee the excellence of the engine cylinder and the steadiness of the machining procedure, a tool wear monitoring system was created. Using the PHM 2010 milling cutter wear dataset, tests were performed to validate the model effect. Based on the experimental data, this model has an average RMSE of 2.67 and an average MAE of 2.5. The trial outcomes proved the effectiveness of the suggested approach in evaluating the tool's wear.
Nikhil Srinivasan, Krishna M, Naveen V, Kishore S M, Sampath Kumar, and Subha R
IEEE
In recent years, machine learning algorithms have gained popularity in the field of economic forecasting. This study aims to predict the Indian Gross Domestic Product (GDP) using advanced machine learning algorithms. To achieve this, we collected data from various sources, including time series analysis and inflation rate. We analyzed the data using linear regression and polynomial regression techniques to determine which method produced the most accurate results. Our results showed that the polynomial regression model outperformed the linear regression model in terms of accuracy. The polynomial regression model was better able to capture the non-linear relationships between the independent variables and the dependent variable (GDP). Specifically, our findings showed that the polynomial regression model was able to predict the Indian GDP with an accuracy of 91%, compared to 87% for the linear regression model. This study highlights the importance of using advanced machine learning algorithms in economic forecasting. We found that the use of high-quality data sets and advanced techniques such as polynomial regression can significantly improve the accuracy of economic forecasts. Our findings have several implications for policymakers and businesses. Accurate predictions of economic indicators such as GDP can help businesses make informed decisions about investment and growth strategies, while policymakers can use these predictions to develop effective economic policies. Overall, our study provides valuable insights into the use of machine learning algorithms in predicting Indian GDP. Our findings demonstrate the effectiveness of polynomial regression in capturing non-linear relationships and improving the accuracy of economic forecasts. This study can be used as a reference for future research in this area and emphasizes the need for high-quality data sets and advanced machine-learning techniques in economic forecasting.
V. naveen Kumar, Gurpreet Singh, S. Rudresha, and S Sampath Kumar
IEEE
Due to the rising usage of sensors and networked equipment in manufacturing, artificial intelligence solutions help extract significant value from extensive data infrastructure. These methods may aid manufacturing sustainability and decision-making. Failure to check tool condition causes a lot of scrap in machining. This work develops an intelligent (IoT) tool condition monitoring system to detect sustainability-related production tradeoffs and optimum machining settings by monitoring machine tool status. A Pareto optimum front visualizes the ideal operating conditions found via evolutionary artificial intelligent (AI) algorithm-based multi-objective optimization.
S. Sampathkumar and R. Rajeswari
Informa UK Limited
The cultivation of crops, conservation of plants, restoration of landscape, and management of soil are the phases incorporated in agriculture and horticulture. During the cultivation and conservation stages, the plants and the crops are affected by various diseases such as Bacterial scourge, Bacterial Leaf Blight, Brown spot, Seeding blight, Leaf streak, Powdery Mildew, Fire Blight, Black Rot and Apple Scab. These diseases in plants will lead to losses such as manufacturing and financial loss in farming industry worldwide. To maintain the sustainability in horticulture, the detection of crop disease and maintaining the condition of the plants are important. The Computer Aided Detection (CAD) in the agriculture and horticulture is the emerging trend, based on the digital imaging that provides the detailed analysis about the disease by applying the image mining process. In this work, the Cross Central Filter (CCF) technique is proposed to perform the noise removal process in the image and the identification of objects in the image is applied by using the Cognitive Fuzzy C-Means (CFCM) algorithm to differentiate the suspicious region from the normal region. The evaluation is conducted against the diseases affected in the rice crop and apple trees. The performance evaluation proves that the proposed design achieves the best performance results compared to the other filters and the segmentation techniques.