基于混沌PSO算法优化LS-SVM的惯导系统测试  

Monitoring of INS based on chaos PSO algorithm optimization parameters of LS-SVM

作　　者：王成[1] 郝顺义[1] 翁大庆[2] 冯文[1] 

机构地区：[1]空军工程大学工程学院 [2]中国人民解放军94371部队装备部

出　　处：《传感器与微系统》2011年第2期125-128,共4页Transducer and Microsystem Technologies

摘　　要：基于混沌PSO算法优化最小二乘支持向量机(LS-SVM)实现惯导系统初始对准测试。通过小波包分解消除陀螺漂移数据的噪声,获取LS-SVM的训练与测试样本。针对LS-SVM解决大规模数据样本回归问题时所出现的训练时间长、收敛速度慢等缺点,提出了混沌PSO算法优化LS-SVM的模型参数。该方法不仅克服了传统PSO算法早熟、容易陷入局部最小值等缺点,同时显著提高了LS-SVM的预测能力。将一般LS-SVM和GM(1,1)模型的预测结果与本算法预测结果进行对比,验证了本方法在预测精度上具有明显优势。Based on chaos PSO algorithm optimization parameters of LS-SVM,monitoring of INS＇ s initial alignment is realized. Noises in INS＇ s error data are eliminated by wavelet decomposition,learning and testing samples for LS- SVM are also acquired. Aimed at LS-SVM solving large scale data regression led to long training time and slow convergence speed, chaos PSO algorithm optimization parameters of LS-SVM is proposed. The disadvantages of earliness and tending to get into local solution in traditional PSO algorithm are overcomed by this method. It also remarkably improves forecasting ability of LS-SVM. The results of general LS-SVM and GM（ 1,1 ） forecasting model is compared with the results of this article,it proves this method has a transparent superior in forecasting precision.

关 键 词：最小二乘支持向量机 混沌粒子群优化算法 惯导系统 初始对准 

分 类 号：V249.32[航空宇航科学与技术—飞行器设计]