AI-Driven 5G Exploring Machine Learning Models for Enhanced Coverage Prediction

Abstract

In the era of 5G technology, predicting coverage areas is crucial for optimizing network performance and ensuring reliable connectivity. This study presents a comprehensive analysis of various machine learning algorithms for predicting 5G coverage based on the RF Signal Data. The target column, Band Width, is used to gauge prediction accuracy across different models. Traditional methods such as Logistic Regression, K-Nearest Neighbors (KNN), Naive Bayes, Random Forest, Support Vector Machine (SVM), XGBoost, LightGBM, AdaBoost, Bayesian Network Classifier, Multi-Layer Perceptron (MLP), and Long Short-Term Memory (LSTM) are evaluated against proposed advanced techniques like Stacking and Voting Classifiers, and Convolutional Neural Networks (CNN). The objective is to identify dominant feature parameters that significantly influence 5G coverage prediction. By implementing a diverse array of models, this research aims to benchmark the performance and accuracy of these algorithms. The comparative analysis highlights the strengths and limitations of each approach, providing valuable insights for network engineers and researchers. The findings suggest that ensemble methods, particularly Stacking and Voting Classifiers, along with CNN, offer superior prediction accuracy and robustness, thereby serving as promising tools for enhancing 5G network planning and deployment.

Introduction

Overview:

This study focuses on predicting 5G network coverage using advanced machine learning (ML) techniques and a large dataset of 164,160 observations across 27 key parameters (e.g., Frequency, Signal Strength, Modulation, Bandwidth). As 5G deployment expands, accurate coverage prediction is crucial for efficient planning, resource allocation, and reliable connectivity across varied environments.

Motivation:

• Traditional coverage planning struggles with accuracy, especially in high-frequency 5G bands.

• The study aims to enhance 5G performance and optimize network deployment using improved predictive models.

Techniques Used:

1. Stacking Classifier: Combines multiple models through a meta-classifier to improve prediction accuracy.

2. Voting Classifier: Aggregates predictions from diverse models (SVM, Decision Tree, etc.) to enhance robustness.

3. Convolutional Neural Networks (CNNs): Used to explore spatial patterns in signal data and improve feature extraction.

Literature Review Highlights:

• Coverage issues and challenges with small-cell deployment in high-frequency bands.

• ML-based indoor planning models (e.g., Genetic Algorithms) outperform traditional path loss methods.

• Studies validate ML's potential in signal strength prediction (e.g., Gaussian Process Regression, Random Forest).

• New frameworks like MLOE (Machine Learning Online Estimator) demonstrate improved accuracy and scalability in network planning.

Proposed System:

• Integrates stacking, voting classifiers, and CNNs to create a robust, multi-model system for 5G coverage prediction.

• Focuses on model training, validation, and feature importance analysis.

• Aims to identify the best-performing models for real-world deployment.

Methodology:

• Stacking Classifier: Multi-stage model with base learners and a meta-learner.

• Voting Classifier: Uses majority or average prediction from multiple models.

• CNN: Extracts spatial and feature-level insights from structured signal data.

Results:

• The Voting Classifier achieved the highest accuracy (99%), outperforming all other individual ML models.

• Its ensemble structure makes it more resilient to biases and more effective for real-world applications.

Conclusion

In conclusion, this study has demonstrated the effectiveness of ensemble methods such as Stacking and Voting Classifiers, alongside Convolutional Neural Networks, in predicting 5G coverage. Through analysis of 27 parameters across diverse locations, including Frequency, Signal Strength, Modulation, and Bandwidth, we identified critical features influencing coverage efficacy. The findings highlight the importance of integrating multiple data modalities to enhance prediction accuracy, crucial for optimizing 5G deployment strategies. By refining predictive models, this research contributes to more efficient network planning and management, offering valuable insights for future advancements in telecommunications infrastructure.

References

[1] C. Sudhamani, M. Roslee, J. J. Tiang, and A. U. Rehman, ‘‘A survey on 5G coverage improvement techniques: Issues and future challenges,’’ Sensors, vol. 23, no. 4, p. 2356, Feb. 2023, doi: 10.3390/s23042356. [2] M. M. Ahamed and S. Faruque, ‘‘5G network coverage planning and analysis of the deployment challenges,’’ Sensors, vol. 21, no. 19, p. 6608, Oct. 2021, doi: 10.3390/s21196608. [3] Y. H. Santana, R. M. Alonso, G. G. Nieto, L. Martens, W. Joseph, and D. Plets, ‘‘Indoor genetic algorithm-based 5G network planning using a machine learning model for path loss estimation,’’ Appl. Sci., vol. 12, no. 8, p. 3923, Apr. 2022, doi: 10.3390/app12083923. [4] M. F. A. Fauzi, R. Nordin, N. F. Abdullah, and H. A. H. Alobaidy, ‘‘Mobile network coverage prediction based on supervised machine learning algorithms,’’ IEEE Access, vol. 10, pp. 55782–55793, 2022, doi: 10.1109/ACCESS.2022.3176619. [5] M. F. A. Fauzi, R. Nordin, N. F. Abdullah, H. A. H. Alobaidy, and M. Behjati, ‘‘Machine learning-based online coverage estimator (MLOE): Advancing mobile network planning and optimization,’’ IEEE Access, vol. 11, pp. 3096–3109, 2023, doi: 10.1109/ACCESS.2023.3234566. [6] H. Chiroma, A. Y. Gital, N. Rana, S. M. Abdulhamid, A. N. Muhammad, A. Y. Umar, and A. Abubakar, ‘‘Nature inspired meta-heuristic algorithms for deep learning: Recent progress and novel perspective,’’ in Advances in Computer Vision, K. Arai and S. Kapoor, Eds. Cham, Switzerland: Springer, 2020, pp. 59–70. [7] 2018 ITU Kaleidoscope Machine Learning for a 5G Future (ITU K), IEEE Staff, Piscataway, NJ, USA, 2018. [8] M. Soltani, V. Pourahmadi, A. Mirzaei, and H. Sheikhzadeh, ‘‘Deep learning-based channel estimation,’’ IEEE Commun. Lett., vol. 23, no. 4, pp. 652–655, Apr. 2019, doi: 10.1109/LCOMM.2019.2898944. [9] T. Gruber, S. Cammerer, J. Hoydis, and S. T. Brink, ‘‘On deep learningbased channel decoding,’’ in Proc. 51st Annu. Conf. Inf. Sci. Syst. (CISS), 2017, pp. 1–6, doi: 10.1109/CISS.2017.7926071. [10] H. Ye, G. Y. Li, and B.-H. Juang, ‘‘Power of deep learning for channel estimation and signal detection in OFDM systems,’’ IEEE Wireless Commun. Lett., vol. 7, no. 1, pp. 114–117, Feb. 2018, doi: 10.1109/LWC.2017.2757490. [11] Khan, M. Q., Gaber, A., Parvini, M., Schulz, P., & Fettweis, G. (2023). A Low-Complexity Machine Learning Design for mmWave Beam Prediction. arXiv. DOI: 10.48550/arXiv.2310.19323 [12] Shi, Y., Lian, L., Wang, Z., Fu, L., Bai, L., Zhang, Y., Zhang, J., & Zhang, W. (2023). Machine Learning for Large-Scale Optimization in 6G Wireless Networks. arXiv. DOI: 10.48550/arXiv.2301.03377 [13] Almasan, P., Suárez-Varela, J., Lutu, A., Cabellos-Aparicio, A., & Barlet-Ros, P. (2023). Enhancing 5G Radio Planning with Graph Representations and Deep Learning. arXiv. DOI: 10.48550/arXiv.2309.03603 [14] Perifanis, V., Pavlidis, N., Koutsiamanis, R.-A., & Efraimidis, P. S. (2022). Federated Learning for 5G Base Station Traffic Forecasting.arXiv. DOI: 10.48550/arXiv.2211.15220 [15] Yuliana, H., Iskandar, & Hendrawan. (2024). Comparative Analysis of Machine Learning Algorithms for 5G Coverage Prediction: Identification of Dominant Feature Parameters and Prediction Accuracy. IEEE Access, 12, 18939–18956. DOI: 10.1109/ACCESS.2024.3361403

Copyright

Copyright © 2025 Mrs. CH. Priyanka, Jammu Keerthana, Samudrala Sowmya Rani , Battu Ajay Kumar, Pachimadla Sai Kumar . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.