Learning-Based WiFi Traffic Load Estimation in NR-U Systems

The unlicensed spectrum has been utilized to make up the shortage on frequency in new radio (NR) systems. To fully exploit advantages brought by bands, one of key issues is guarantee fair coexistence with WiFi reach this goal, timely and accurate estimation traffic loads an important prerequisite. In paper, a machine learning (ML) based method proposed detect number users bands. An unsupervised Neural Network (NN) structure applied filter detected transmission collision probability spectrum, which enables NR precisely rectify measurement error estimate active users. Moreover, NN trained online related parameters rate are jointly optimized adaptively high accuracy. Simulation results demonstrate that compared conventional Kalman Filter detection mechanism, approach lower complexity can achieve more stable estimation.