A Spectrum Detection and Unknown Noise Variance Tracking Estimation Method and Device

Abstract

Aiming at the problem of spectrum detection in the case of unknown noise variance dynamics in practical applications, the present invention first proposes a dynamic state space system model to reflect its internal mechanism: the authorized user state and the time-varying noise variance are regarded as two hidden states, and two states are adopted respectively. A first-order Markov and autoregressive model models its dynamic migration properties. On this basis, a novel spectrum sensing method device is designed and proposed. The method is rooted in Bayesian statistical reasoning theory, and adopts marginalized particle filter technology (see attached figure), which can realize joint estimation of authorized users and noise variance. In particular, the present invention proposes a two-stage adaptive prediction coefficient adjustment method for marginalized particle filtering, which can fully utilize the time-varying characteristics of noise variance and can achieve accurate tracking of noise variance. Using the obtained noise variance information, the algorithm device can be extended to (but not limited to) a single-node single-antenna sensing system, and obtains good spectrum sensing performance.

Description

technical field [0001] Aiming at the problem of frequency spectrum detection under the condition of time-varying unknown noise variance, the present invention first designs and proposes a dynamic state-space system model (Dynamic State-space Model, DSM), which is based on two-state Markov state probability transition model and autoregressive (Autoregressive) model respectively. -regressive, AR) model, which regards the authorized user state and time-varying noise variance as two hidden states (Hidden States); on this basis, a novel spectrum detection method is proposed. This method is based on Bayesian theory and uses Marginal Partical Filtering (MPF) technology, which can realize accurate detection of authorized user status and real-time tracking of unknown noise variance at the same time, and significantly improve the frequency spectrum under the condition of noise variance dynamic unknown. perceived performance. belongs to the field of communication. Background technique...

AI Technical Summary

Problems solved by technology

In practical applications, due to the uncertainty and time-varying nature of noise, it is difficult for cognitive users to obtain accurate real-time noise variance, which leads to a significant decline in the detection performance of existing detection methods

Images