连续隐马尔科夫模型在多基地目标识别中的应用

发布时间：2018-07-28 16:30

【摘要】：多基地声纳组网探测系统是目前大范围水下安保领域的研究热点。综合利用多基地系统中各个声纳节点的信息进行水下目标识别是亟待解决的问题。利用传统的多传感器融合的方法进行多基地水下目标识别,往往忽略了各声纳节点之间的相关性,效果并不理想。针对这一问题,本文提出了利用连续隐马尔科夫模型(CHMM)进行多基地水下目标识别的方法。首先利用RELAX算法提取了目标在不同分置角上回波的强散射点特征,组成观测向量,利用Baum-Welch方法对CHMM参数进行训练,然后计算待识别目标的特征值观测序列在不同模型下的似然概率。对所有目标重复此过程,取概率最大值对应的目标类别为最后的识别结果。在消声水池开展多基地模拟实验,对四类目标进行了识别,利用CHMM方法得到的多基地水下目标融合识别率比多基地声纳下单声纳节点的最高识别率提高了30%。
[Abstract]:Multi-static sonar netting detection system is a hot spot in the field of underwater security. The problem of underwater target recognition based on the information of each sonar node in multistatic system is urgent to be solved. The traditional multi-sensor fusion method for multi-static underwater target recognition often ignores the correlation between the sonar nodes and the effect is not satisfactory. To solve this problem, a method of multi-static underwater target recognition using continuous Hidden Markov Model (CHMM) is proposed. Firstly, the RELAX algorithm is used to extract the characteristics of the strong scattering points of the target echo at different separation angles, and the observation vector is formed, and the CHMM parameters are trained by the Baum-Welch method. Then the likelihood probability of the observed sequence of eigenvalues of the target to be identified under different models is calculated. This process is repeated for all targets and the target category corresponding to the maximum probability is taken as the final recognition result. The multi-static simulation experiment was carried out in the silencing tank and the recognition of four kinds of targets was carried out. The recognition rate of the multi-static underwater target fusion obtained by using the CHMM method was 30% higher than the highest recognition rate of the multi-static sonar node of the order sonar sent out by the multi-static sonar.
【作者单位】：中国科学院声学研究所;中国科学院大学;
【基金】：国家自然科学基金自资助项目(11404365,61471353)
【分类号】：TB566

【相似文献】