基于BP神经网络的VMG预测  

VMG Forecasting of Sailboat Using Back Propagation Neural Network Technique

作　　者：潘明杰[1] 任久春[1] 汪建波 朱谦[1] 

机构地区：[1]复旦大学通信系 [2]上海市水上运动中心

出　　处：《微型电脑应用》2016年第4期45-47,共3页Microcomputer Applications

基　　金：上海市科研计划项目(编号14231202102)

摘　　要：VMG(Velocity Made Good)指帆船于前进过程中船速在风方向上的投影,其体现了帆船在风向上前进的能力,反映了帆船运动员利用风的能力。因此,对运动员而言,若能掌握VMG的变化趋势,即为运动员制定帆船航行方向决策提供了科学依据。基于目前所采集到的有关帆船运动的原始数据,利用BP神经网络建立三层神经网络结构模型,以船速、船向、海风速度以及海风风向等4个指标作为输入样本,对帆船在下一时刻的VMG速度进行预测。实验仿真结果证明了这种研究方法的有效性。VMG（Velocity Made Good） is the projection of sailboat speed in the direction of true wind during the sailing,which shows their ability in sailing against wind as well as athletes＇ ability in making use of wind.Therefore,if we can master the trend of VMG,athletes can make scientific decisions according to the forecast.Based on the raw data which we collected during the athletes＇ daily training,a three-layer neural network structure model was built using back propagation neural network model.Four indexes such as velocity and direction of boat and velocity and direction of wind were used to predict the VMG of next moment.Simulation results have approved the effectiveness of this method.

关 键 词：VMG 时间序列预测 BP算法 神经网络 机器学习 帆船训练 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]