Extendable AI in Mechatronics Engineering Pushpalatha Naveenkumar, Surendar Rangaraju, T. Kokilavani, K. Kannan, M Praveen Kumar Handbook of AI Based Mechatronics Systems and Smart Solutions in Industrial Automation, 2025 Indeed, the use of artificial intelligence (AI) in mechanical engineering and mechatronics has recently started to pick up pace as there is increased advancement in technology. That is why as systems become more complex or in fact more automated, the need for extensible AI is rising. These systems are capable of learning and therefore engineers are capable of developing solutions which can change with the ever-changing needs and be more efficient. Beyond that, the fact that such systems may also exhibit a human-like behavior should be taken into consideration when designing the systems in the future generations of robots, cars, and healthcare ventures. Thus, it is made clear that AI is fundamental in the integration of mechanical engineering and mechatronics, and also due to extensiveness that AI has introduced, more new innovative inventions generated positively in the field can be explored.
The Resilience of Artificial Intelligence in the Mechatronics Sector: Challenges and Future Directions M. Praveen Kumar, S. Saradha, M. Meena, S. Vijayakumar, Pushpalatha Naveenkumar Handbook of AI Based Mechatronics Systems and Smart Solutions in Industrial Automation, 2025 In today’s changing field of mechatronics, integrating artificial intelligence (AI) has become a crucial driver of innovation, efficiency, and adaptability. This article explores how AI has not just survived but thrived amidst disruptions and market dynamics, in the mechatronics industry. By examining literature case studies and industry trends, our research aims to provide an understanding of AIs long-term impact. We begin by shedding light on how AI transforms mechatronics by optimizing processes, improving decision-making abilities, and enhancing system performance. Through in-depth analysis of AI applications in systems, we demonstrate the adaptability of AI algorithms to dynamic environments with exceptional self-learning and self-improvement capabilities. Furthermore, we delve into the challenges that AI faces in mechatronics such as concerns, data security issues, and privacy considerations. We explore solutions and frameworks developed to address these challenges while highlighting the versatility of AI in tackling ethical and regulatory problems. To conclude, we discuss the implications of AI resilience for the future of mechatronics – how it can drive sustainability, autonomy, and competitiveness within the sector. This chapter also presents the function of AI that handle the interruptions and avail the variations.
Deepfake Video Detection Using Inception_Resnet_V2: A Convolutional Neural Network Approach Mandala Praveen kumar, V Valli Kumari, K Swathi Lakshmi Durga Proceedings 2024 IEEE 16th International Conference on Communication Systems and Network Technologies Cicn 2024, 2024 Deepfake videos, generated by sophisticated AI algorithms, pose significant challenges to the authenticity and trustworthiness of multimedia content on the internet. In this study, we propose a deepfake video detection system utilizing the Inception_ResNet_v2 architecture, a deep convolutional neural network renowned for its effectiveness in image classification tasks. The system is trained and evaluated on the Deepfake Challenge Dataset sourced from Kaggle, comprising a diverse set of real and manipulated video clips. We preprocess the dataset, extracting frames from each video and augmenting them to enhance model generalization. The Inception_ResNet_v2 model is then fine-tuned using transfer learning, leveraging pre-trained weights from ImageNet to expedite convergence and improve performance. Through extensive experimentation and evaluation, we demonstrate the efficacy of our approach in accurately distinguishing between genuine and deep-fake videos. The proposed system achieves promising results in detecting deep-fake videos, showcasing its potential utility in combating the proliferation of synthetic media and safeguarding the integrity of digital content. Our findings underscore the importance of leveraging advanced machine learning techniques for addressing the evolving challenges posed by deep-fake technology.
BIG DATA CONGESTION ANALYTICS THROUGH QUALITY OF SERVICE FOR CROWDSOURCE USING DEEP LEARNING ALGORITHMS IN HEALTHCARE ENVIRONMENT Journal of Environmental Protection and Ecology, 2023
Resnet Based Blockchain Architecture for The Detection of Plant Leaf Disease in Agriculture Field B. Aruna Devi, M. Praveen Kumar, Lakshmana Phaneendra Maguluri, P. Tamilselvan 2023 International Conference on Disruptive Technologies Icdt 2023, 2023 The only way to get better crop yields is to find and treat crop diseases quickly. Deep learning models diagnoses the plant diseases by looking at the leaves. A residual neural network is developed for the detection of disease in maize leaf. The leaves are collected from the available dataset, where the detection architecture is decentralized using blockchain architecture. The residual neural network with decentralized blockchain enables an optimal classification of instances. The model is implemented with improved disease detection accuracy with reduced training time in a python simulator with keras library. The results of simulation show an improved rate of classification accuracy, precision, recall land f-measure in detecting the leaf disease than the existing convolutional neural network models.
A Survey on Wireless Network Traffic Analysis using Machine Learning Algorithms M. Praveen Kumar, P Ashwitha Noble, V S Esha Malavika, G Geethanjali, A S Farheen 3rd International Conference on Innovative Mechanisms for Industry Applications Icimia 2023 Proceedings, 2023 Analyzing and forecasting network traffic for different applications has gained popularity in recent years. Several tests have been carried out to find issues with the computer network apps that are in use today. Anticipating and evaluating network traffic can assist in implementing preemptive steps to guarantee safe, dependable, and excellent network connection. Numerous methods, including data mining approaches and neural network-based techniques, have been suggested and assessed for the purpose of predicting patterns of network traffic. For this goal, other linear and nonlinear models have also been developed. The main objectives of combining different network analysis and prediction techniques are efficiency and optimization. The goal of this survey study is to present a thorough review of the variety of methods used in traffic prediction and network analysis. The paper delves into the distinctive features and findings of previous research efforts and brings to light the patterns that have emerged from these studies. Furthermore, it provides an overview of the various areas in which network traffic analysis and prediction have shown their effectiveness.
Cloud Computing Based Diabetes Classification Using Binary Grey Wolf Optimization I. Jasmine Selvakumari Jeya, M. Praveen Kumar, B. Saravanan, V. Niranjani, V. Muneeswaran 2023 1st International Conference on Advances in Electrical Electronics and Computational Intelligence Icaeeci 2023, 2023 The surge in available technology has led to a significant rise in data within the medical field. This abundance of data has posed challenges for analyzing and extracting insights due to its sheer volume, making it difficult to identify meaningful patterns. To address this issue, methods that involve selecting specific data features have become crucial. These methods aim to streamline data processing and focus medical research by pinpointing important aspects. This research aims to assess the effectiveness of the BGWO algorithm in handling medical data. A comparison is made between BGWO and other feature selection techniques along with various classifiers. In this study, two additional methods, namely the genetic algorithm (GA) and binary particle swarm optimization (BPSO), previously employed to tackle medical classification problems, were applied to diabetic medical data. Employing feature selection techniques helps eliminate unnecessary and irrelevant data, reducing data complexity, model complexity, and computation time for classifiers. Typically, this results in improved classifier performance. The findings indicate that BGWO stands out as a successful feature selection algorithm for medical diagnosis, as evidenced by its strong performance on medical data.
Air Quality Image Classification using Ensemble Based Transfer Learning Techniques S. Saradha, M. Praveen Kumar, J. Asha, R. Saranya, S. P. Santhoshkumar Iet Conference Proceedings, 2023 In recent years, research has focused on the efficacy of neural networks, Learning by Transfer (TL), and Ensemble Learning (EL) techniques in image processing. Ensemble approaches, when employed on image classification tasks improve the overall predictive performance and robustness. This work examines the studies related to ensemble approach and applies the same for the classification of images based on different air pollution levels. The proposed research work uses three Convolution Neural Networks (CNN) models namely VGG, MobileNet, and EfficientNet and ensembles them using stacking technique for air quality image classification. The results exhibited in the suggested model generated an overall classification accuracy of 99%.
