基于粒子滤波的频谱检测技术研究

发布时间：2018-05-01 01:15

本文选题：粒子滤波算法 + 混合滤波算法　；参考：《长春理工大学》2017年硕士论文

【摘要】：本文以隐马尔科夫模型频谱算法为基础,针对现有的频谱检测算法中没有对计算复杂度进行更进一步探讨的现状进行分析的问题,对粒子滤波算法进行改进(结合贝叶斯近似方法),使得在保持粒子滤波算法较高精确度的同时,还减少了相应的运行时间,提升了计算效率。该基于粒子滤波的改进算法(混合滤波算法)集中了贝叶斯近似和粒子滤波的优势,在保证了估计精度的前提下,可以在计算复杂度和时间复杂度方面达到良好的平衡。文章中的仿真模型一方面,只利用隐马尔科夫模型,用于模拟子信带的状态估计,动态估计频带的状态,此时不需要考虑信道的影响。另一方面,在隐马尔科夫模型的基础上,辅助自回归模型构成混合模型来仿真单径衰落信道。仿真结果表明,与粒子滤波算法相比,该改进算法在与其具有相近的精确度基础上,可以在频谱感知过程中比粒子滤波减少运行时间。
[Abstract]:Based on the hidden Markov model spectrum algorithm, this paper analyzes the current situation of the existing spectrum detection algorithm, which has no further research on computational complexity. The particle filter algorithm is improved (combined with Bayesian approximation), which keeps the high accuracy of the particle filter algorithm, and reduces the corresponding running time and improves the computational efficiency. The improved particle filter algorithm (hybrid filtering algorithm) combines the advantages of Bayesian approximation and particle filter, and can achieve a good balance between computational complexity and time complexity under the premise of ensuring the estimation accuracy. On the one hand the simulation model in this paper only uses hidden Markov model to simulate the state estimation of subband and dynamically estimate the state of frequency band without considering the influence of channel. On the other hand, based on the hidden Markov model, the auxiliary autoregressive model is used to simulate the single-path fading channel. The simulation results show that compared with the particle filter algorithm, the improved algorithm can reduce the running time compared with the particle filter in the spectral sensing process on the basis of similar accuracy.
【学位授予单位】：长春理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN925

【参考文献】