SURC 2025 Student Presentations
SUNY Undergraduate Research Conference Student Presentations

Developing New Machine Learning Classification Models for Next Generation Wireless Systems

Authors: David Serino, Amit Sangwan

SUNY Campus: SUNY Poly

Presentation Type: Poster

Location: Old Union Hall

Presentation #: 27

Timeslot: Session C 1:45-2:45 PM

Abstract: Current wireless systems face significant challenges due to spectrum scarcity, dynamic environments, and higher reliability and efficiency. As the number of connected devices and data traffic continues to grow, traditional static modulation schemes and signal processing techniques struggle to adapt to rapidly changing channel conditions, interference, and diverse use-case requirements. In parallel, Orthogonal Time Frequency Space (OTFS) modulation excels in such conditions by transforming symbols over the delay-doppler domain, offering robustness against rapid channel variations. The ability to quickly switch between modulation schemes based on real-time environmental conditions could significantly enhance spectral efficiency, reduce latency, and improve reliability. Advanced machine learning classifiers enable this dynamic adaptation by identifying optimal modulation schemes and receiver algorithms, ensuring seamless performance across diverse wireless scenarios. To achieve this, we trained a custom classifier model capable of analyzing received input waveforms and dynamically switching receiver algorithms based on the transmitted modulation scheme. Leveraging MATLAB’s deep learning and wireless toolboxes, we designed and trained a model using convolutional neural networks (CNNs) to classify signals from a diverse set of modulation schemes, including BPSK, QPSK, 8PSK, 16QAM, 64QAM, PAM4, GFSK, CPFSK, B-FM, DSB-AM, SSB-AM, and OTFS. The model achieves high accuracy given a synthetic dataset generated under varying channel conditions. By dynamically adapting to specific use-case scenarios, this approach paves the way for more efficient spectrum utilization and enhanced system performance in next-generation wireless networks, such as 5G/6G and IoT applications.