Research On Cancellation Of Interference From Nb-IoT To LTE System

Due to the scarcity of spectrum,many existing and emerging communication systems are deployed very close to each other,and even spectrum overlap occurs,which will undoubtedly cause strong interference between the two systems.Narrow Band Internet of Things(NB-Io T)is a 5G technology suitable for large-scale machine type communication.Its bandwidth is much smaller than that of Long Term Evolution(LTE),so the interference from NB-Io T to LTE can be regarded as a kind of Narrow Band Interference(NBI).For the harmonious coexistence of heterogeneous networks,the impact of NBI must be eliminated or reduced.The research of this article is mainly focused on proposing some methods to eliminate NBI,and the specific contents are as follows.Firstly,in order to get the block sparse vector of NBI,this paper proposes two algorithms based on Block Sparse Bayesian Learning: Partition Estimated Block Sparse Bayesian Learning(PE-BSBL)and Informative Block Sparse Bayesian Learning(I-BSBL).The biggest difference between these two algorithms is that I-BSBL does not require pre-estimation of block division.Secondly,this paper proposes a novel framework,i.e.,sparse relative entropy minimization,based on sparse machine learning and a sparse combination optimization,which is used to accurately eliminate NBI.The algorithm obtains the relative entropy through the non-zero vector set distribution and the residual error norm,and uses the gradient descent method to derive the minimum value of the relative entropy to solve the sparse combination optimization problem.Simulation results show that the algorithm proposed in this paper is effective in eliminating NBI interference.In addition,compared with traditional compressed sensing based algorithms,the algorithm proposed in this paper performs better in terms of mean square error and recovery probability,and can accurately eliminate NBI with a large sparse level based on limited measurement data.