基于CSAGA-LSSVM算法的坦克驾驶模拟训练数据分类挖掘  

Data classification mining of tank driving simulation training based on CSAGA-LSSVM algorithm

作　　者：邓青 薛青 翟凯 DENG Qing;XUE Qing;ZHAI Kai(Training Center, Academy of Army Armored Forces, Beijing 100072, China;68303 Troops, Geermu, Qinghai 816009, China)

机构地区：[1]陆军装甲兵学院演训中心,北京100072 [2]68303部队,青海格尔木816099

出　　处：《山东科技大学学报（自然科学版）》2022年第1期24-33,共10页Journal of Shandong University of Science and Technology(Natural Science)

基　　金：军内科研基金项目(JY2019C095)。

摘　　要：利用坦克驾驶模拟器进行训练是提高操作技能的重要方法。针对以往驾驶模拟训练采用统计分析方法难以从复杂训练数据中发现知识和规律的不足,提出CSAGA-LSSVM算法对坦克驾驶模拟训练数据进行分析。首先选择关键点快速生成Shapelets,以减少候选Shapelets数量;其次,根据距离和时间间隔对Shapelets进行组合,增强特征辨识能力;然后,设计自适应遗传算法,动态调整交叉、变异概率,寻找最小二乘支持向量机最优参数解,提高分类结果的准确性;与其他分类方法进行实验对比,验证了CSAGA-LSSVM算法的可行性与有效性。最后,将算法应用于某型坦克驾驶模拟器换挡操作数据的分类挖掘,提取不同训练水平人员的操作特征,促进指导个性化训练。Training with tank driving simulators is an important method to improve operation skills.In view of the difficulty to find knowledge and rules from complex training data by statistical analysis method in the past driving simulation training,this paper proposed the CSAGA-LSSVM classification mining algorithm to analyze tank driving simulation training data.Firstly,key points were quickly selected to generate shapelets and reduce the number of candidate shapelets.Secondly,the shapelets were then grouped according to distance and time interval to enhance the ability of feature identification.An adaptive genetic algorithm was designed to dynamically adjust the probability of crossover and mutation to find the optimal parameter solution of least squares support vector machine and improve the accuracy of classification results.Compared with other classification methods,the feasibility and effectiveness of the CSAGA-LSSVM algorithm were further verified.Finally,the algorithm was applied to the classification mining of gear shift operation data from a certain tank driving simulator to extract the operation characteristics of personnel with different training levels and guide personalized training.

关 键 词：坦克驾驶模拟器 支持向量机 Shapelets特征 遗传算法 分类挖掘 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]