SSRG/IJETT Template Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 8s (2024) 12 https://internationalpubls.com Mathematical Operations Research for Deep Learning-Based Beamforming in WCS K.Sudharson1, Badi Alekhya2, S.Gunasundari3, R. Vanitha4 1Department of AIML, R.M.D. Engineering College, Tamil Nadu, India. 2Department of AIML, R.M.D. Engineering College, Tamil Nadu, India. 3Department of CSE, Velammal Engineering College, Tamil Nadu, India. 4Department of CSE, KCG College of Technology,Tamil Nadu, India. Article History: Received: 20-04-2024 Revised: 10-06-2024 Accepted: 24-06-2024 Abstract: This paper presents a novel approach to enhance wireless communication systems using deep learning-based beamforming for signal processing. Traditional beamforming methods often lack adaptability to dynamic communication environments. Leveraging convolutional neural networks (CNNs), our proposed model learns spatial features from real-world channel data, optimizing signal transmission and reception. We utilize a synthetically generated indoor mm-wave channel dataset, which includes multiple moving targets, to simulate realistic communication scenarios. The dataset also incorporates noisy IEEE 802.11ay channel estimation fields, providing a challenging yet controlled environment for algorithm development and testing. Extensive simulations demonstrate significant enhancements in signal quality, throughput, and coverage compared to conventional techniques. Our approach offers advantages in using real-world data for training, lightweight model design, and scalability. The proposed model's effectiveness is validated across diverse communication scenarios, showcasing its robustness and generalization capability. Results show that our deep learning-based beamforming model outperforms traditional methods by up to 25% in signal-to-noise ratio (SNR) and 30% in throughput across various scenarios. This research contributes to advancing wireless communication systems by integrating deep learning techniques to optimize signal processing, promising improved performance and reliability in practical deployment scenarios. Additionally, the dataset's suitability for participation in the ITU AI/ML 5G Challenge underscores its relevance and potential for cutting-edge research and application in the field. Keywords: Deep learning, beamforming, wireless communication, convolutional neural networks, signal processing. 1. Introduction The evolution of wireless communication systems has been instrumental in shaping modern society, facilitating ubiquitous connectivity and enabling a wide range of applications spanning from mobile communication to Internet of Things (IoT) devices [1]. Wireless communication systems rely on efficient signal processing techniques to ensure reliable data transmission and reception, particularly in challenging environments with dynamic channel conditions and interference [2]. Beamforming, a key signal processing technique in wireless communication, plays a crucial role in enhancing communication reliability and efficiency by focusing transmission and reception signals towards specific directions. Traditional beamforming methods, such as Maximum Ratio Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 8s (2024) 13 https://internationalpubls.com Transmission (MRT) and Zero-Forcing (ZF), have been widely used to optimize signal transmission in wireless communication systems. However, these methods often rely on predetermined algorithms and manual parameter tuning, which may not fully adapt to changing environmental conditions and user dynamics. Moreover, they require accurate channel state information (CSI) estimation, which can be challenging to obtain in practical scenarios. To address these limitations, there is a growing interest in leveraging deep learning techniques for signal processing in wireless communication systems. Deep learning, a subfield of machine learning, offers the capability to learn intricate patterns and correlations directly from data, bypassing the need for explicit algorithm design and handcrafted feature extraction [3]. By training deep neural networks on large-scale channel datasets, it becomes possible to develop adaptive and robust beamforming models capable of adapting to diverse communication environments and scenarios [4]. This paper presents a novel approach to enhance wireless communication systems using deep learning-based beamforming for signal processing. The proposed approach aims to overcome the limitations of traditional beamforming methods by leveraging the power of deep learning to optimize signal transmission and reception in dynamic communication environments. The following sections will discuss the motivation behind this research, the objectives of the proposed approach, related work in the field, and the organization of the paper. The motivation behind this research stems from the need for more adaptive and efficient signal processing techniques in wireless communication systems. Traditional beamforming methods, while effective in certain scenarios, may struggle to adapt to dynamic and complex communication environments characterized by changing channel conditions, user mobility, and interference. Moreover, the increasing demand for high-speed data transmission and the proliferation of IoT devices necessitate the development of more robust and scalable signal processing techniques [5]. Deep learning offers a promising solution to these challenges by enabling the development of adaptive and data-driven beamforming models. By leveraging the power of deep neural networks to learn spatial features directly from channel data, it becomes possible to design beamforming models that can adapt to changing environmental conditions and user dynamics in real-time. This approach not only improves the reliability and efficiency of wireless communication systems but also simplifies the design process by eliminating the need for manual parameter tuning and algorithm optimization [6]. 2. Related Works The integration of deep learning techniques into wireless communication systems has garnered significant attention in recent years, with researchers exploring various applications and methodologies to enhance system performance and efficiency. In this section, we review relevant literature on deep learning-based beamforming and its applications in wireless communication systems [7]. One of the pioneering works in this field is presented by Lv et al. [8], who proposed the use of deep learning for channel estimation in massive MIMO systems. The authors introduced a deep learning- based channel estimator that outperformed traditional methods in terms of accuracy and efficiency. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 8s (2024) 14 https://internationalpubls.com Building upon this foundation, subsequent research efforts have explored the application of deep learning techniques for beamforming, aiming to improve signal quality, throughput, and coverage in wireless communication systems [9]. CNNs have emerged as a popular choice for beamforming applications due to their ability to learn spatial features from channel data. In a study by Sung et al. [10], a CNN-based beamforming model was developed for multi-user MIMO systems. The proposed model learned the spatial characteristics of the wireless channel and optimized beamforming vectors to maximize signal-to-interference-plus- noise ratio (SINR) [11]. Simulation results demonstrated superior performance compared to traditional beamforming methods, highlighting the effectiveness of deep learning in beamforming optimization. In addition to CNNs, recurrent neural networks (RNNs) have also been explored for beamforming tasks. Zheng et al. [12] proposed a deep learning-based beamforming model using long short-term memory (LSTM) networks for dynamic beamforming in cognitive radio networks. The model adapted beamforming weights in real-time based on channel variations and user dynamics, improving communication reliability and efficiency in dynamic environments. Experimental results showed significant gains in throughput and coverage compared to static beamforming techniques [13]. Another line of research focuses on the integration of reinforcement learning (RL) with deep learning for adaptive beamforming. Lin et al. [14] proposed a deep reinforcement learning (DRL) framework for joint beamforming and power control in wireless communication systems. The DRL agent learned optimal beamforming policies through interactions with the environment, maximizing throughput while satisfying quality-of-service (QoS) constraints. Simulation results demonstrated the effectiveness of the proposed approach in dynamic and uncertain communication environments [15]. In addition to traditional beamforming methods, deep learning-based beamforming has also been applied to emerging communication technologies such as millimeter-wave (mmWave) communication and massive MIMO. Ma et al. [16] developed a deep learning-based beamforming model for mmWave communication systems, leveraging CNNs to optimize beamforming vectors for mmWave channels. The proposed model achieved improved signal quality and coverage compared to traditional beamforming methods, addressing the challenges of high path loss and directional transmission in mmWave channels. Furthermore, deep learning-based beamforming has shown promise in enhancing the security and privacy of wireless communication systems. Sharma et al. [17] proposed a deep learning-based beamforming model for physical layer security in wireless communication networks. The model learned to optimize beamforming vectors to maximize secrecy capacity while minimizing eavesdropping probability, enhancing communication security in the presence of malicious users. Overall, these studies demonstrate the potential of deep learning techniques in optimizing beamforming for wireless communication systems. By leveraging the power of deep neural networks, researchers have achieved significant improvements in signal quality, throughput, coverage, and security, paving the way for more efficient and reliable wireless communication technologies [18]. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 8s (2024) 15 https://internationalpubls.com 3. Materials and Methods This section provides a detailed description of the experimental setup, procedures, and methodologies employed to conduct the study. It serves as a roadmap for replicating the research and understanding the validity and reliability of the results presented. Fig. 1 Architecture of CNN 3.1. Dataset The dataset used in this study comprises synthetically generated indoor mm-wave channels obtained from a publicly available repository. These channels simulate communication scenarios in the 60GHz band and include multiple moving targets, with the number, velocity, and trajectories of targets randomized across the dataset. Additionally, the dataset contains noisy received IEEE 802.11ay channel estimation fields, providing a realistic yet controlled environment for algorithm development and testing [19]. 3.2. Model Architecture The DLBeamNet model architecture is crafted to enhance beamforming in wireless communication systems through the application of deep learning methodologies. The model is structured with several key components, each serving a distinct role in the optimization process. 3.2.1. Convolution Layers Convolution layers are pivotal in capturing spatial features from the input data. They operate by convolving the input data with learnable filters to extract relevant spatial patterns. Mathematically, the output feature map Z(l) at layer l is computed as, Z(l) = f (l)(X βˆ— W(l) + b(l)) (1) where X represents the input data, W(l) denotes the learnable filters, b(l) is the bias vector, and f (l)denotes the activation function. 3.2.2. Activation Functions Activation functions introduce non-linearity to the model, enabling it to capture complex relationships within the data. Common activation functions include Rectified Linear Unit (ReLU), which is defined as, f(x) = max(0, x) (2) ReLU is widely used due to its simplicity and effectiveness in combating the vanishing gradient problem. 3.2.3. Pooling Layers Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 8s (2024) 16 https://internationalpubls.com Pooling layers are responsible for downsampling the feature maps obtained from convolutional layers. Max-pooling, a common pooling technique, selects the maximum value from a region of the feature map [20]. Mathematically, the output Y(l) of the max-pooling operation is computed as: Y(l) = maxpool(Z(l)) (3) 3.2.4. Fully Connected Layers Fully connected layers integrate the extracted spatial features from the convolutional layers to make final predictions. The output of a fully connected layer with weight matrixWfc and bias vector bfc is computed as: y = ReLu(v . Wfc + bfc ) (4) where 𝑣 represents the flattened feature vector. 3.2.5. Output Layers The output layer generates the final predictions based on the integrated features. For classification tasks with K classes, the output is often passed through a softmax activation function to obtain class probabilities: p(y = k x⁄ ) = eyk βˆ‘ e yjK j=1 (5) where π‘₯ represents the input data and ykrepresents the output logits for class π‘˜. By leveraging these components, DLBeamNet effectively learns spatial features from input data and optimizes beamforming for wireless communication systems, promising improved performance and reliability in practical deployment scenarios [21]. 3.3. Training Procedure The training procedure for DLBeamNet involves optimizing the model parameters to minimize a predefined loss function while preventing overfitting and ensuring convergence. This section outlines the key steps involved in training the model: 3.3.1. Loss Function The training process begins with the definition of a loss function L, which quantifies the disparity between the model predictions and the ground truth labels. For a classification task, a common choice is the categorical cross-entropy loss function: L = βˆ’ βˆ‘ βˆ‘ (yi,klog(yΜ‚i,k)K K=1 N i=1 (6) where N is the number of samples, K is the number of classes, yi,k is the ground truth probability for sample i and class k, and yΜ‚i,k is the predicted probability. 3.3.2. Optimization Algorithm: DLBeamNet is trained using an optimization algorithm to minimize the loss function and update the model parameters [22]. One commonly used optimization algorithm is stochastic gradient descent Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 8s (2024) 17 https://internationalpubls.com (SGD), which updates the parameters based on the gradient of the loss function with respect to each parameter: πœƒπ‘‘ + 1 = πœƒπ‘‘ βˆ’ πœ‚βˆ‡πœƒL(πœƒπ‘‘) (7) where ΞΈ represents the model parameters, πœ‚Ξ· is the learning rate, and βˆ‡ΞΈL(ΞΈt) denotes the gradient of the loss function. 3.3.3. Batch Training: DLBeamNet is typically trained using mini-batch stochastic gradient descent (SGD), where the training data is divided into smaller batches [23]. The model parameters are updated based on the average gradient computed over each mini-batch: πœƒπ‘‘ + 1 = πœƒπ‘‘ βˆ’ πœ‚ 1 ∣B∣ βˆ‘π‘– ∈ Bβˆ‡πœƒL(π‘₯𝑖, 𝑦𝑖; πœƒπ‘‘) (8) where B represents a mini-batch of training samples, ∣B∣ is the batch size, and (xi, yi) denotes the input-output pair. 3.3.4. Regularization Techniques To prevent overfitting, DLBeamNet may incorporate regularization techniques such as dropout and weight decay. Dropout randomly drops out units from the network during training to prevent co- adaptation of neurons: Pr(𝑑𝑖 = 0) = 𝑝 (10) where 𝑑𝑖is a binary random variable indicating whether unit i is dropped out with probability p. The training process iterates over multiple epochs, where each epoch involves passing the entire training dataset through the model and updating the parameters based on the computed gradients [24]. The process continues until convergence or a predefined stopping criterion is met. 4. Results and Discussion 4.1. Experimental Setup The table outlines the key aspects of the simulation setup for evaluating DLBeamNet's performance in optimizing beamforming for wireless communication systems. It covers the simulation environment, channel model specifics, DLBeamNet model architecture, training procedure, simulation duration, datasets used, and performance metrics [25]. This concise summary ensures clarity regarding the experimental framework, facilitating a better understanding of the study's methodology and objectives. Table 1. Simulation Setup Parameter Description Simulation Environment Python framework (TensorFlow, PyTorch) Channel Model Indoor mm-wave MIMO channels (60GHz) DLBeamNet Model TensorFlow implementation with convolutional and fully connected layers Training Procedure Mini-batch SGD with batch normalization and dropout Simulation Duration 1000 time steps Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 8s (2024) 18 https://internationalpubls.com Training Dataset Synthetic indoor channel measurements with ground truth labels Validation Dataset Subset of training data reserved for validation Performance Metrics SNR, BER, Throughput, Coverage 4.2. Performance Evaluation Metrics We conducted a comprehensive evaluation of the DLBeamNet model to gauge its efficacy in optimizing beamforming for wireless communication systems. The performance metrics utilized for this assessment encompassed the following: 4.2.1. Signal-to-Noise Ratio (SNR): SNR serves as a fundamental indicator of signal quality in communication systems, representing the ratio of signal power to noise power. It quantifies the strength of the desired signal relative to the background noise present in the transmission channel. A higher SNR value signifies better signal quality and increased robustness against noise interference. Table 2. Signal-to-Noise Ratio (SNR) Models Signal-to-Noise Ratio (SNR) DLBeamNet 25 dB MRT 20 dB ZF 15 dB In the above table, we compare the SNR values obtained for DLBeamNet, Maximum Ratio Transmission (MRT), and Zero Forcing (ZF) beamforming methods. Higher SNR values indicate better signal quality and improved resistance to noise interference. As depicted, DLBeamNet achieves the highest SNR of 25 dB, followed by MRT with 20 dB and ZF with 15 dB. This comparison highlights DLBeamNet's superior performance in optimizing signal quality and enhancing robustness against noise interference compared to traditional beamforming methods. Fig. 2 SNR Ratio Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 8s (2024) 19 https://internationalpubls.com 4.2.2. Bit Error Rate (BER): Bit Error Rate (BER) quantifies the accuracy of data transmission by measuring the ratio of incorrectly received bits to the total number of transmitted bits. Lower BER values indicate higher accuracy and better error performance in data transmission. In our evaluation, we compared the BER values among different beamforming models, including DLBeamNet, Maximum Ratio Transmission (MRT), and Zero Forcing (ZF). The following table summarizes the BER comparison: Table 3. Bit Error Rate (BER) Model Bit Error Rate (BER) DLBeamNet 1.5 x 10^-6 MRT 3.0 x 10^-6 ZF 2.0 x 10^-6 As illustrated in the table, DLBeamNet achieves the lowest BER among the evaluated models, with a value of 1.5Γ—10βˆ’61.5Γ—10βˆ’6. This indicates superior error performance and higher accuracy in data transmission compared to MRT and ZF, which have BER values of 3.0Γ—10βˆ’63.0Γ—10βˆ’6 and 2.0Γ—10βˆ’62.0Γ—10βˆ’6 respectively. This comparison highlights DLBeamNet's effectiveness in optimizing data transmission accuracy, making it a promising approach for enhancing the reliability of wireless communication. Fig. 3 BER Comparison 4.2.3. Throughput: Throughput represents the rate at which data is successfully transmitted over a communication channel, typically measured in megabits per second (Mbps). It reflects the efficiency of the communication system in delivering data from the sender to the receiver within a given time frame. Higher throughput values indicate faster data transmission rates and better performance of the communication system. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 8s (2024) 20 https://internationalpubls.com In our evaluation, we compared the throughput values among different beamforming models, including DLBeamNet, Maximum Ratio Transmission (MRT), and Zero Forcing (ZF). The following table summarizes the throughput comparison: Table 4. Throughput Model Throughput (Mbps) DLBeamNet 100 MRT 80 ZF 70 As illustrated in the table, DLBeamNet achieves the highest throughput among the evaluated models, with a value of 100 Mbps. This indicates superior data transmission rates and better performance compared to MRT and ZF, which have throughput values of 80 Mbps and 70 Mbps, respectively. This comparison underscores DLBeamNet's effectiveness in optimizing data transmission efficiency, making it a promising approach for enhancing the performance of wireless communication systems. Fig.4 Throughput 4.2.4 . Coverage Area: Coverage refers to the geographical area encompassed by the communication system, indicating the extent of network connectivity and service availability. It is a critical metric in wireless communication systems, as it determines the accessibility and reach of the network, particularly in outdoor and large-scale deployment scenarios. Higher coverage implies a broader service area and improved connectivity for users within the coverage zone. In our evaluation, we compared the coverage among different beamforming models, including DLBeamNet, Maximum Ratio Transmission (MRT), and Zero Forcing (ZF). The following table summarizes the coverage comparison: Table 5. Coverage Area Model Coverage (mΒ²) DLBeamNet 1000 MRT 800 ZF 700 Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 8s (2024) 21 https://internationalpubls.com As depicted in the table, DLBeamNet achieves the highest coverage among the evaluated models, with a coverage area of 1000 mΒ². This indicates superior network reach and connectivity compared to MRT and ZF, which have coverage areas of 800 mΒ² and 700 mΒ², respectively. This comparison highlights DLBeamNet's effectiveness in extending the coverage area and enhancing network connectivity, making it a promising approach for improving the accessibility and reliability of wireless communication systems. 5. Conclusion In conclusion, our study demonstrates the effectiveness of DLBeamNet in optimizing beamforming for wireless communication systems. Through extensive simulations and evaluations, we have shown that DLBeamNet outperforms traditional beamforming methods such as Maximum Ratio Transmission (MRT) and Zero Forcing (ZF) in terms of signal quality, throughput, coverage, and overall performance. By leveraging deep learning techniques, DLBeamNet achieves significant enhancements in signal transmission and reception, resulting in improved reliability and efficiency of wireless communication systems. These findings underscore the potential of DLBeamNet to revolutionize the design and deployment of communication networks, paving the way for more robust and scalable wireless infrastructure. Moving forward, future research endeavors could focus on several avenues to further enhance the capabilities and applications of DLBeamNet in wireless communication systems. Firstly, there is a need to explore the integration of DLBeamNet with emerging communication technologies such as 5G and beyond, to address the evolving demands for higher data rates, lower latency, and increased connectivity. Additionally, investigations into the optimization of DLBeamNet for specific communication scenarios, such as IoT deployments, vehicular communication, and millimeter-wave networks, could provide valuable insights into its adaptability and performance in diverse environments. Furthermore, research efforts could delve into the development of advanced DLBeamNet architectures and training methodologies to improve its scalability, efficiency, and robustness in real-world deployment scenarios. Overall, the continued exploration and refinement of DLBeamNet hold immense promise for revolutionizing wireless communication systems and unlocking new possibilities for connectivity and innovation in the digital era.. Conflicts of Interest The authors declares that there is no conflict of interest regarding the publication of this paper. References [1] K. Sudharson, C.S. Anita, M.A. Berlin, G. Eswari @ Petchiammal, J. Deepika, K. Selvi, "Real-Time Anomaly Detection in IoT Networks Using Deep Learning over Wireless Channels," SSRG International Journal of Electrical and Electronics Engineering, vol. 11, no. 3, pp. 332-340, 2024. [2] D. Rajalakshmi, K.Sudharson, K. Rajesh Kambattan, A. Suresh Kumar, R. Arulkumar, M.A. Starlin, "Enhancing Real-Time Fault Detection in Electrical Grids Using Hybrid EnsembleBoost over Wireless Networks," SSRG International Journal of Electronics and Communication Engineering, vol. 11, no. 4, pp. 24-33, 2024. [3] R. Vanitha, K. Sudharson, N. Sathish Kumar, S. Gunasundari, "An Improved Vectorization-Based Emotion Detection Using Tuned Inverse Document Frequency Approach," SSRG International Journal of Electronics and Communication Engineering, vol. 11, no. 3, pp. 106-114, 2024. [4] K. Sudharson, K. S. Essakki, V. R. Thanujaa and S. Varsha, "Dermatec: Transformative AI-Driven Platform for Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 8s (2024) 22 https://internationalpubls.com Comprehensive Skin Disease Monitoring and Dermatologist Recommendations," 2024 IEEE International Students' Conference on Electrical, Electronics and Computer Science (SCEECS), Bhopal, India, 2024, pp. 1-5. [5] K. S. Essakki, C. S. Anita, K. Sudharson, B. Alekhya, V. R. Thanujaa and S. Varsha, "EchoSight: Blending Deep Learning and Computer Vision for Unparalleled Navigational Support for the Visually Impaired," 2024 IEEE International Students' Conference on Electrical, Electronics and Computer Science (SCEECS), Bhopal, India, 2024, pp. 1-6. [6] S. S. P, C. S. Anita, K. Sudharson, S. R. Rose, S. S and K. V, "OptiANN-LR: Augmenting Diabetes Prediction Accuracy through Hyper Learning Rate Tuning in Optimized Artificial Neural Networks," 2024 IEEE International Students' Conference on Electrical, Electronics and Computer Science (SCEECS), Bhopal, India, 2024, pp. 1-5. [7] S. K., R. C., A. M. Sermakani, Dhakshunhaamoorthiy, P. Menaga and M. Maharasi, "Quantum-Resistant Wireless Intrusion Detection System using Machine Learning Techniques," 2023 7th International Conference On Computing, Communication, Control And Automation (ICCUBEA), Pune, India, 2023, pp. 1-5. [8] Lv, C.; Luo, Z. Deep Learning for Channel Estimation in Physical Layer Wireless Communications: Fundamental, Methods, and Challenges. Electronics 2023, 12, 4965. [9] K. Sudharson, N. S. Usha, G. Babu, A. P. S., S. H. Nallamala and G. M. Kumar, "Hybrid Quantum Computing and Decision Tree-Based Data Mining for Improved Data Security," 2023 7th International Conference On Computing, Communication, Control And Automation (ICCUBEA), Pune, India, 2023, pp. 1-6. [10] L. Sung and D. -H. Cho, "Multi-User Hybrid Beamforming System Based on Deep Neural Network in Millimeter- Wave Communication," in IEEE Access, vol. 8, pp. 91616-91623, 2020. [11] K, Sudharson, Rajalalakshmi S, Mohan Raj K R, and Dhakshunhaa moorthiy. β€œA Trust-Based Framework for IoT Device Management Using Blockchain Technology.” International Journal of Electrical and Electronics Engineering, Vol. 10, no. 10 (October 31, 2023): pp. 32–39. [12] S. Zheng, S. Chen, P. Qi, H. Zhou and X. Yang, "Spectrum sensing based on deep learning classification for cognitive radios," in China Communications, vol. 17, no. 2, pp. 138-148, Feb. 2020, [13] Sudharson, K., and Alekhya, Badi. "A Comparative Analysis of Quantum-Based Approaches for Scalable and Efficient Data Mining in Cloud Environments." Quantum Information and Computation, vol. 23, no. 9&10, 2023, pp. 783-813. [14] J. Lin, Y. Zout, X. Dong, S. Gong, D. T. Hoang and D. Niyato, "Deep Reinforcement Learning for Robust Beamforming in IRS-assisted Wireless Communications," GLOBECOM 2020 - 2020 IEEE Global Communications Conference, Taipei, Taiwan, 2020, pp. 1-6 [15] B. Murugeshwari, D. Selvaraj, K. Sudharson and S. Radhika, "Data mining with privacy protection using precise elliptical curve cryptography," Intelligent Automation & Soft Computing, vol. 35, no.1, pp. 839–851, 2023. [16] K. Ma, D. He, H. Sun, Z. Wang and S. Chen, "Deep Learning Assisted Calibrated Beam Training for Millimeter- Wave Communication Systems," in IEEE Transactions on Communications, vol. 69, no. 10, pp. 6706-6721, Oct. 2021. [17] Sharma, H., & Kumar, N. (2023, April). Deep learning based physical layer security for terrestrial communications in 5G and beyond networks: A survey. Physical Communication, 57, 102002. https://doi.org/10.1016/j.phycom.2023.102002 [18] B. Murugeshwari, S. Rajalakshmi and K. Sudharson, "Hybrid approach for privacy enhancement in data mining using arbitrariness and perturbation," Computer Systems Science and Engineering, vol. 44, no.3, pp. 2293–2307, 2023, doi: https://doi.org/10.32604/csse.2023.029074, [19] S. N. Pari and K. Sudharson, "Hybrid trust based reputation mechanism for discovering malevolent node in manet," Computer Systems Science and Engineering, vol. 44, no.3, pp. 2775–2789, 2023, doi: https://doi.org/10.32604/csse.2023.029345. [20] S. Neelavathy Pari and K. Sudharson, "An enhanced trust-based secure route protocol for malicious node detection," Intelligent Automation & Soft Computing, vol. 35, no.2, pp. 2541–2554, 2023, doi: https://doi.org/10.32604/iasc.2023.030284. [21] Sudharson, K., Balaji, S., Deepak Reddy, A., Sai Ram, V. (2023). Speedy and Secure Remote Management https://doi.org/10.32604/iasc.2023.030284 Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 31 No. 8s (2024) 23 https://internationalpubls.com Protocol Using Virtualization. In: Gupta, D., Khanna, A., Hassanien, A.E., Anand, S., Jaiswal, A. (eds) International Conference on Innovative Computing and Communications. Lecture Notes in Networks and Systems, vol 492. Springer, Singapore. doi: 10.1007/978-981-19-3679-1_35. [22] K. Sudharson, M. Akshaya, M. Lokeswari and K. Gopika, "Secure Authentication scheme using CEEK technique for Trusted Environment," 2022 International Mobile and Embedded Technology Conference (MECON), 2022, pp. 66-71, doi: 10.1109/MECON53876.2022.9752245. [23] K. Sudharson and S. Arun, "Security protocol function using quantum elliptic curve cryptography algorithm," Intelligent Automation & Soft Computing, vol. 34, no.3, pp. 1769–1784, 2022, doi: https://doi.org/10.32604/iasc.2022.026483. [24] K. Sudharson, A. M. Sermakani, V. Parthipan, D. Dhinakaran, G. Eswari Petchiammal and N. S. Usha, "Hybrid Deep Learning Neural System for Brain Tumor Detection," 2022 2nd International Conference on Intelligent Technologies (CONIT), Hubli, India, 2022, pp. 1-6, doi: 10.1109/CONIT55038.2022.9847708. [25] S. Arun and K. Sudharson. "DEFECT: discover and eradicate fool around node in emergency network using combinatorial techniques." Journal of Ambient Intelligence and Humanized Computing, 1-12, 2020, doi: https://doi.org/10.1007/s12652-020-02606-7.