GL-RBF优化的数据手套手势识别算法  

Data Glove Gesture Recognition Algorithm Based on GL-RBF Neural Network Optimization

作　　者：李东洁[1] 李洋洋[1] 杨柳[1] 

机构地区：[1]哈尔滨理工大学机电控制及自动化技术研究所,黑龙江哈尔滨150080

出　　处：《哈尔滨理工大学学报》2017年第4期7-12,共6页Journal of Harbin University of Science and Technology

基　　金：国家自然科学基金(51105117);黑龙江省自然科学基金(QC2014C054);黑龙江省博士后科研启动基金(LBH-Q13094);黑龙江省高校青年学术骨干支持计划项目(1254G023);哈尔滨理工大学青年拔尖创新人才培养计划资助(201304)

摘　　要：针对5DT数据手套手势识别过程中的精度和实时性问题,提出GA算法和LM算法混合优化的RBF神经网络手势识别方法。该算法首先确定RBF神经网络基函数的中心,根据遗传算法确定神经网络的初值,同时利用LM算法和遗传算法混合训练RBF神经网的权值和阈值,将已经完成训练的神经网络应用到实际数据手套的手势识别过程中。通过实验及仿真结果可知,所提出的设计方法能够有效的提高训练速度,减少训练时间,并且提高了数据手套与机器人操作系统的实时性和控制精度。In order to solve the problem of accuracy and real-time in the process of gesture recognition of 5DT data glove, a hybrid optimization method of RBF （Radial Basis Function） neural network gesture recognition is based on LM （ Levenberg Marquardt） algorithm and GA （ Genetic Algorithm） neural network. First of all, The algorithm determines the center of the RBF neural network basis functions, determining the initial value of the neural network based on the genetic algorithm, at the same time, the LM algorithm and genetic algorithm are used to train the weights and thresholds of RBF neural network, and the trained neural network has been applied to the gesture recognition of real data glove. The experiment and simulation results show that the proposed design method can effectively improve the training speed, reduce the training time, and improve the real-time performance and control accuracy of the data glove and the robot operating system.

关 键 词：手势识别 数据手套 RBF神经网络 GA算法 LM算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]