Open Access
Subscription Access
AI-driven Transformer Networks in Shared Spectrum for Enhanced Signal Processing for Nonlinear Receivers
Subscribe/Renew Journal
Communication systems face challenges from high-power adjacent channel signals, or blockers, inducing nonlinear behavior in RF front ends. Ensuring robust performance in the presence of blockers is crucial for IoT and other spectrum-consuming devices coexisting with advanced transceivers. This paper proposes a flexible, data-driven solution using a Deep Belief Network (DBN) to mitigate third-order intermodulation distortion (IMD) during demodulation. Numerical evaluations of AI-enhanced receivers employing DBN as an IMD canceler and demodulator show significant improvements in bit error rate (BER) performance. The effectiveness of DBN varies with RF front end characteristics, notably the third-order intercept point (IP3).
Keywords
Deep Belief Network, IMD Cancellation, Nonlinear Receivers, RF Front End, Bit Error Rate
Subscription
Login to verify subscription
User
Font Size
Information
Abstract Views: 132