Proceedings of the 7th International Conference on Communications and Cyber Physical Engineering

Inhaltsverzeichnis

1. Enhancing Accent-to-Accent Translation Using Python

   Sunil Bhutada, V. Kakulapati, T. Venkat, A. Bharati, R. Anvitha

   Abstract

   Accent is the basic pattern of acoustic features and pronunciation. It can identify the person's social and linguistic background. It is an important source of inter as well as intra-speaker variability. The accent-dependent dictionary or model can be used to improve the accuracy of the speech recognition system. Many voice translation apps struggle with understanding and translating accents accurately. This leads to confusion and makes communication challenging, especially for non-native speakers. Additionally, there's a lack of tools that help users visualize how words are pronounced in different accents. To solve this, we need to design and implement an Accent-to-Accent Translator system capable of accurately translating spoken and written content between distinct English accents. This system should address the differences in pronunciation, intonation, and vocabulary to ensure clear and contextually appropriate communication, stimulating better understanding and collaboration. In this work, using statistical analysis to convert a voice into a foreign country accent. The process of back translation, especially in cases where there are no corresponding words in the target language can be effective. Analysis of results according to sound stimuli based on their context (technological, human, nature) can reveal useful information. This work can translate your voice and read aloud the translated results. Allowing you to travel, communication and social networking is no longer a language barrier. Voice Translator can also be used as you learn and understand a language tool, carrying your custom dictionary. Your voice will be translated into the local accent of a particular country. This application will give more support to students who wish to study in other countries.

2. Analysis of Most Played Telugu Songs on Spotify

   V. Kakulapati, Sunil Bhutada, Bolledhu Nandini Trishna, Rentala Jyothirmai, Pyaram Akhil, Varakala Pranay Goud

   Abstract

   The contemporary electronic age is redefining the dissemination of music with computerized methods that can collect and analyze large amounts of data from the web. The objective of this research is to use various machine learning and statistical techniques to examine the acoustic characteristics of songs. We have experimented with several models including random effects, since we are particularly interested in understanding the determinants of a song's popularity and acquiring more insights into these elements. Our study will be beneficial for those seeking to predict the achievement of new goods in the music industry since it delineates the essential aspects that affect a song's popularity. This study seeks to elucidate the correlation between a song's auditory characteristics and its popularity, as well as to discover the key aspects that lead to a song's popularity. The proposed system offers an essential contribution to our comprehension of the music industry and its future course in this fascinating area of research. The work proposes the use of order-limiting methods in K-means clustering to achieve stable outcomes and improve accessibility, which might decrease categorization mistakes in activities involving measure identification. This approach also enables the automated extraction of the musical framework of popular song content.

3. Exploration of Big Data Analytics Using Machine Learning (ML): A Comprehensive Survey

   E. JohnAlex, P. Venkatapathi, Chintala Cury, Yasmeen, S. Samatha, B. V. Krishnaveni

   Abstract

   The rapid evolution of data in various domains has necessitated the development of advanced analytics techniques to extract meaningful insights from vast datasets. In this study, we look at how Big Data analytics and machine learning work together, and how this synergy is changing data analysis. It provides an overview of key machine learning techniques, including supervised, unsupervised, semi-supervised, and reinforcement learning, and their applications in handling Big Data. The paper also discusses the essential tools and technologies, such as Hadoop, Spark, TensorFlow, PyTorch, and cloud platforms, that facilitate efficient data processing and analysis. By examining case studies and addressing future trends, the paper illustrates how machine learning can enhance decision-making, optimize processes, and drive innovation. The challenges associated with scalability, data quality, and computational complexity are also considered. This comprehensive review aims to provide a deeper understanding of how machine learning techniques and Big Data technologies can be leveraged to unlock valuable insights and address complex analytical problems.

4. Improving Power Quality in Distribution System Using ZSI-DVR

   K. Vijaya Bhaskar Reddy, Sunitha Tappari, B. Bhuvan Reddy, A. Hanumantha Rao, Srinivasa Rao Balasani, P. Archana Rao

   Abstract

   Power quality problems are the power system's primary concerns. Examples of these phenomena include voltage swells, sags, flickers, voltage notches, harmonic distortions, transients, and brief pauses. The majority of these problems mostly affect industrial clients. The best way to reduce these problems is to use custom power system components, such as UPQC, DVR, D-STATCOM, and others. DVR is a superior option to the others because of its better, more dependable, and more efficient operation. VSI (Voltage Source Inverter) is the main component of DVR and it receives power via a DC link capacitor or battery. The obtainable output voltage of a voltage-sensing instrument (VSI) is limited by the DC source voltage. On the input side of the VSI, a separate DC-DC boost converter is required for a PV-supplied system. The boost converter DC-DC requires additional switches. This results in high switching losses and overall system volume, as well as increased system costs. Additionally, when both of the same leg's switches are turned on, the inverter is damaged and loses dependability, resulting in shoot-through (ST). The primary goal of this work is to enhance DVR performance by substituting a ZSI for a VSI and using Unit Vector Template-Pulse Width Modulation (UVT-PWM) as the DVR controller. This study examines the ZSI-DVR's harmonic mitigation performance in addition to its voltage compensation performance under voltage swell and sag scenarios. The performance of ZSI-DVR and conventional VSI-DVR will also be compared in this paper. As an enhancement we are replacing the power MOSFET given by the reference author using power IGBT.

5. Performance Analysis of Super Capacitor for Energy Storage

   Thamatapu Eswara Rao, Jonnalagadda Ankamma Rao, Nomula Malla Redddy

   Abstract

   Battery technology has a number of disadvantages, despite being widely used and well-established. These disadvantages include mass, weight, high internal resistance, low power density, and poor transient response. On the other hand, thanks to developments in materials science and other technologies, energy storage devices like electrostatic double layer capacitors (EDLC), ultra capacitors, and super capacitors (SCs) look extremely promising. They are appropriate for a range of applications due to their lower internal resistance, lighter weight, smaller volume, higher transient responsiveness, and higher power density. Super capacitors benefit from their high power density and quick recovery from high currents. Super capacitors are a useful tool for lowering battery peak current and extending battery life in energy storage systems (ESS). The present paper describes the implementation of a field-programmable gate array (FPGA)-based pulse-width modulation controller for a buck-boost converter. This controller helps in controlling the super capacitor’s performance. With the built hardware prototype, multiple modes of operation of the super capacitors were evaluated at different duty cycles and under charging and discharging situations utilizing an FPGA-based pulse width modulation (PWM) in a bidirectional converter.

6. Empowering a Pollution-Free Tomorrow: AI-Driven Innovations for Universal Clean Air

   Pragathi Jogi, T. Charan Singh

   Abstract

   Our conventional approaches to air quality monitoring are showing their shortcomings in the face of the increasing air pollution challenge. Understanding and resolving the challenges of modern air pollution requires a more sophisticated and proactive strategy due to the complicated interaction of industrialization, urbanization, and natural processes. Traditional monitoring approaches that depend on stationary devices and erratic data gathering are not meeting the needs of providing the precise, real-time information needed for efficient management. Here is where artificial intelligence (AI) becomes an innovation, offering a paradigm-shifting opportunity for us to tackle the challenges that are associated with air quality dynamics. With its advanced machine learning algorithms and data processing powers, artificial intelligence (AI) sheds a light on new frontiers for dynamic, flexible, and perceptive air pollution monitoring. The many ways that artificial intelligence (AI) is changing the field of real-time air quality assessment are examined in this paper. It explores concrete examples and case studies that eloquently demonstrate the transformative potential of AI in addressing the multiple issues presented by air pollution, going beyond theoretical talks. AI can be a sophisticated tool that enhances our understanding of air quality. It goes beyond being a flashy gadget; instead, it acts as a trusted companion, working alongside us to improve and sustain a cleaner and healthier atmosphere. Whether it's providing insights into daily air conditions or predicting potential pollution events, AI transcends its role as a mere assistant – it becomes an integral force in reshaping our world into a healthier environment with cleaner air accessible to everyone. This review paper serves as a guided journey, shedding light on how AI accomplishes this, using uncomplicated examples to underscore its relevance for all of us.

7. Enhancing Extractive Text Summarization Through Ensemble Techniques

   B. Tirapathi Reddy, B. Sai Krishna Athul, J. Lasya, G. Kiran

   Abstract

   In today's constantly evolving and data-rich world, the rapid development of information has necessitated an evolutionary change toward the retrieval of accurate, concise, and comprehensible content. In this context, it is essential to demonstrate the ability to summarize user-generated content while preserving its core meaning, thus text summarizers have been introduced. This study integrates natural language processing (NLP) with ensemble extraction, a novel approach for enhancing the extraction methodologies. The core of this approach consists of intricate processes including HTML Tag Removal, Hyphenated Words Normalization, Whitespace Normalization, Unicode Normalization, Quotation Marks Normalization, and Bullet Points Normalization. People these days are so busy with numerous tasks that they hardly have the time to read everything that may be found online. The issue is addressed by the suggested approach, which offers greater understanding without demanding a lot of reading by summarizing significant study findings. Machine learning models and algorithms have been previously used for text summarization, but they often don't meet expectations and fail to create an impact on society. Our project aims to close this gap by offering a powerful platform for summarizing long texts.

8. Supporting of Crop Yield Prediction Using Machine Learning Algorithm Techniques

   S. Vasundhara, Madhavi lata Mangipudi, Supriya Vaddi, Hema Neelam

   Abstract

   A major factor in India’s financial expansion is agriculture. The country’s increasing population and continuously shifting climate influences crop productivity and food safety. Crop selection is influenced by a multitude of elements, such as government policy, soil type, rainfall, temperature, market price, and production rate. Numerous adjustments are needed in the agriculture industry in an effort to strengthen the Indian economy. The paper focuses on different types of machine learning algorithms to approximate the crop on different variety of crops in the Indian states. The outcome shows from all the algorithms what we applied; Linear Regression SVM, Decision Tree Ensemble and Neural networks the linear regression and SVM performed better from the remaining algorithms with on 0.853 R², 0.035 RMSE, and 0.0251 Mean Average Error, The insights are validated using cross-validation procedures, Root Mean Square Error and Mean average error. This effort aims to help farmers address agricultural yield-related problems by using the crop selection approach.

9. Ultrasound Nerve Segmentation and Injury Detection Using Deep Learning

   Parige Deepthi Priya, Kagitha Karthik, Vinjamuri Satya Sri Madhurya, Nakshatra Sriramoju, P. Chandini, J. Adilakshmi

   Abstract

   Modern-day curative field of medication has been going through interesting progresses, giving wherewithal which in past times were thought to be extraordinary. Healthcare went through several rounds of innovation to address the complicated demands of patients; everything from the clever use of redistribution strategies, to the accuracy of healing methods, to the demonstrative power of X-Indications. However, amid the stomps, some things remain and, in particular care settings serving the aging state, place issues like pain, pain management, and individualized care plans will always have to be thoughtful and cultured. Ultrasound technology emerges as a valuable tool that addresses this set of challenges by providing physicians with a non-invasive way to explore the complexities of the human body in shapes. Due to the adaptive structure of ultrasound, it can reveal and exotic and pathologies into the conditions of tissues that offers invaluable idea into the health of the individuals. The aim of this work is to take advantage of the promise of ultrasound electronics to target the discovery faults and nerve damage vicious problems. Then by using ultrasound scanning, healthcare professionals can catch problems at a very early stage enabling early interventions and embodied location designs The capability of recognizing and intervening on these problems prior to time not only refines patient upshots but again highlights the basic law of eager healthcare management. By combining ultrasound Image Checking technology and being able to clearly present it, doctors are able to administer increasingly particular and patient-centred care in improving healthcare standards and promoting overall health.

10. Harmonize: Pioneering the Music Industry’s Evolution with Block-Chain, LLMs and Emerging Technologies

    K. Eshwari, Manchi Sarapu Abhinav Chandra, Manchi Sarapu Sharat Chandra, Krishna Priya V. S. Garimella

    Abstract

    This research paper explores the transformative potential of blockchain, Large Language Models (LLMs), and other digital technologies in revolutionizing the music industry. Titled “Harmonize: Revolutionizing the Music Industry through Blockchain, LLMs, and Emerging Technologies,” it presents a comprehensive analysis of how these technologies collectively aim to decentralize the industry, empower artists, and redefine the consumer experience. The advent of blockchain technology introduces a decentralized framework, enabling direct artist-fan interactions, transparent royalty distributions, and efficient copyright management through smart contracts.

    Further, the integration of LLMs and other AI technologies offers innovative solutions for music creation, personalized user experiences, and data-driven insights, facilitating new avenues for creative expression and audience engagement. These technologies not only empower artists by automating administrative tasks and enhancing creative processes but also provide platforms for collaborative, cooperative, and collective business models that emphasize shared ownership and democratic governance.

    However, the transition towards a more inclusive and equitable digital music ecosystem is fraught with challenges, including scalability, legal and regulatory compliance, and the cultural acceptance of these new paradigms. This paper discusses potential strategies to overcome these obstacles, advocating for collaborative efforts among stakeholders, continuous technological innovation, and the development of supportive regulatory frameworks.

    “Harmonize” posits that the confluence of blockchain, LLMs, and digital technologies heralds a new era for the music industry, characterized by decentralization, artist empowerment, and enhanced consumer experiences. By addressing the current challenges and capitalizing on the opportunities these technologies present, the paper envisions a harmonized music industry that values creativity, fairness, and community above all.

11. AI’s Impact on Marketing Success

    L. Sampath, Hema Neelam, S. Vasundara

    Abstract

    Following the introduction of artificial intelligence, significant changes in consumer behaviour and marketing strategies are expected. This article introduces a comprehensive framework for assessing the impacts of artificial intelligence (AI) that is based on extensive engagement with real-world applications and previous scholarly investigations. The analysis takes into account multiple factors, including the integration of artificial intelligence into a robot, different levels of intelligence, and various types of tasks. The investigation progresses by outlining a research strategy that includes essential ethical considerations, privacy implications, and bias considerations, as well as upcoming marketing techniques and consumer behaviours. Prior studies indicate that enhancing the effectiveness of artificial intelligence by combining it with human administrators is more favourable than relying solely on autonomous operations.

12. Design and Simulation of a High-Speed 16-Bit Successive Approximation Register ADC (SAR ADC) with Capacitive-DAC Implementation Using Verilog

    M. Keerthana, K. Ragini, D. Anjali, M. Harshitha, S. Akshitha

    Abstract

    SAR ADCs have emerged as popular solutions for applications requiring high accuracy, moderate speed, compactness, and cost-effectiveness. Extensive research has been dedicated to enhancing the precision of commercial successive approximation SAR ADCs, owing to their crucial role in delivering accurate analog-to-digital conversions. However, achieving this precision necessitates precise internal digital-to-analog converter (DAC) circuitry, typically implemented using capacitors (C-DAC) to optimize power consumption.

    In this article, our focus is on implementing a capacitive DAC for a 16-bit SAR ADC using Verilog. Through comprehensive simulation and verification processes, we evaluate the functionality, performance, and reliability of the designed C-DAC, considering factors such as power consumption, area utilization, and conversion accuracy. Ultimately, this article contributes to advancing the field of SAR ADC design by providing a practical implementation of a key component essential for achieving superior analog-to-digital conversion performance.

13. Analysis of DDoS Attacks Using Machine Learning Technique

    T. Mallika Devi, A. Durga Bhavani, B. Chaitanya, B. Vijayalaxmi

    Abstract

    DDoS (Distributed Denial of Service) attacks pose a serious challenge to online security by overwhelming servers with excessive traffic, making them inaccessible to legitimate users. The goal of such an attack is to overload the server's capacity, causing disruptions and service outages. These attacks are dangerous because they can be carried out with little effort and do not require advanced tools. A large number of infected devices, often referred to as bots, are remotely controlled by a single operator (botmaster), using fake IP addresses to avoid detection. This study focuses on evaluating different machine learning (ML) and deep learning (DL) methods for the detection and analysis of DDoS attacks. It will also highlight the differences between these two approaches, helping to determine the best scenarios for their use in addressing such threats.

14. Privacy-Preserving Federated Machine Learning on Block Chain: A Comprehensive Review of Homomorphic Encryption Techniques

    S. Vijaya Lakshmi, R. Rajasekhar

    Abstract

    The increasing convergence of privacy-preserving machine learning and block chain technology has given rise to ground breaking approaches for collaborative model training. This paper meticulously reviews recent advancements in privacy-preserving federated machine learning, with a primary focus on the seamless integration of homomorphic encryption within block chain networks. Delving deep into the intricacies of secure data collaboration, each participant securely maintains their encrypted datasets within the decentralized block chain framework. Our study thoroughly investigates the practical application of homomorphic encryption, carefully examining its computational intricacies and the ongoing efforts in efficiency optimization. Through a comprehensive survey, we critically evaluate the security implications inherent in this integrated framework and present a discourse on potential enhancements to existing cryptographic protocols. The synthesis of our findings aims to contribute nuanced visions into the present state-of-the-art practices, illuminating the challenges faced, and proposing directions for future advancements in privacy-conscious collaborative machine learning within the dynamic context of block chain environments. This research serves as a foundational exploration of the symbiotic relationship between privacy-preserving techniques and block chain, offering valuable guidance for researchers, and others in the field.

15. An Assessment of an Ideal Huffman Coding Related Approach for the Better Lossless Compression of Images Using IWT

    G. Divya, S. Suma, K. Srinivas, S. Krishnaveni, Y. Aruna Suhasini Devi, K. L. S. Soujanya

    Abstract

    In today's world, there are more and more images being generated as data, making it difficult to store and send them. Lossless compression of images is important for certain industries that need high-fidelity images since it can shrink the amount of data associated with images with no sacrificing quality. We recommend a better lossless picture by combining Huffman coding, integer wavelet transform (IWT), and linear prediction, this compression algorithm theoretically provides a nearly quadrupling of compression to address the challenge of increasing the ratio of image compression with minimal loss. The key aspect of this approach is a new hybrid transform that uses an entirely novel prediction template and an IWT coefficient processing step. The results of the experiments on three different image sets show that the suggested strategy is effective works better than cutting-edge algorithms. The compression ratios were enhanced by 6.42% to 73.26%, at the very least. Having a respectable compression performance, our technique is better suited to compressing photos through complicated texture and greater resolution.

16. Dual-Axis Solar Tracker Using Arduino and LDRs

    Kalagotla Chenchireddy, Gouse Basha Mulla, Shabbier Ahmed Sydu

    Abstract

    This paper offers Dual-Axis Solar Tracker Using Arduino and LDRs. The aim of the proposed paper is to enhance the proficiency of solar energy harvesting by developing an intelligent solar tracking system. This system employs Light Dependent Resistors (LDRs) as sensors to detect ambient light levels, enabling precise adjustments of solar panels along both azimuth and elevation axes. The Arduino controller helps as the brain of the system, orchestrating the synchronized movement of dual-axis servo motors to align solar panels optimally with the sun's position throughout the day. The core functionality of the solar tracker involves real-time monitoring of LDR readings to calculate the solar azimuth and elevation angles. These angles are then used to position the solar panels dynamically, ensuring they are constantly oriented towards the sun for maximum energy absorption. The enactment of the dual-axis solar tracker using Arduino and LDRs offers several advantages, plus increased energy output, improved system efficiency, and a reduction in dependency on fixed solar installations. The low-cost and adaptable nature of the proposed system makes it suitable for various applications, such as residential solar installations, off-grid power systems.

17. Pioneering Multi-fusion Approach of Enhancing Underwater Image Quality

    Y. Sunanda, Y. Pavan Kumar Reddy, K. Shankar, J. Raja Kullayappa, M. Ravi kishore

    Abstract

    Enhancing the visual clarity of underwater images is a significant challenge due to the adverse impacts of water and lighting conditions. Color imbalances, low contrast, and limited visibility of details can hinder the interpretation of underwater images. The proposed approach employs a fusion strategy characterized by multiple weights and granularity levels, effectively integrating diverse image processing methods to tackle the unique constraints inherent to underwater imaging. The core of our method involves addressing color imbalances through precise white balance correction applied to input underwater images. Notably, adaptive histogram equalization with limited contrast is anticipated to play a pivotal role in enhancing image contrast while preserving essential features. A critical phase of this investigation entails evaluating the proposed technique using established quality metrics like UIConM, UIQM, and UCIQE. These metrics are poised to provide valuable insights into the efficacy of the technique in elevating the quality of underwater images. By elevating image quality and enhancing clarity, the proposed approach is poised to facilitate more precise analysis and informed decision-making within the challenging context of underwater environments.

18. Agri Sense: Leveraging Machine Learning to Forecast Crop Yield and Market Prices with Meteorological Data

    M. Aparna, Soumya Vulli, Vaishnavi Munigala, Akshaya Rajana

    Abstract

    An essential component of a nation's economic growth is its agriculture sector. In developing nations like India, agriculture is one of the major sources of income for many people. Indian farmers face immense challenges due to unpredictable weather patterns, leading to difficult decisions about crop cultivation and tragically contributing to a rising number of suicides. This not only threatens individual livelihoods but also poses significant economic risks, as agriculture is a cornerstone of the Indian economy and a major employer. Additionally, the use of excessive fertilizers to meet food demand risks soil degradation and contamination, further exacerbating agricultural sustainability issues. Addressing these challenges requires a holistic approach, including support for farmers, sustainable farming practices, and policies to mitigate climate impacts. Protecting farmers and ensuring agricultural sustainability is crucial for India's prosperity and stability. This paper recognizes the transformative potential of technology and aims to leverage it for the betterment of Indian agriculture. So, we came up with the approach which helps the farmer by giving crop predictions based on analysis of the factors like temperature, Rainfall, Crop production, yield and prices by using Machine Learning Algorithms like – Logistic Regression, Random Forest and Decision Tree. By leveraging these Machine learning algorithms and analysing factors, you can create a robust crop prediction system that assists farmers in making data-driven decisions to optimize their yields and profitability. Additionally, it's essential to continuously evaluate and refine the models based on new data to ensure their accuracy and relevance over time.