中文摘要：

将假肢穿戴者主要跌倒情形分为绊倒与滑倒两大类,对可能作为假肢穿戴者跌倒预警的潜在数据源进行提取,参考人体运动学,排除掉不能作为预警的数据源,设计了跌倒预警系统的结构,运用支持向量机的原理,并结合采集到的有效数据,训练得出相对应的判别式,将其嵌入跌倒预警系统中.该方法可以保证在患者跌倒之前发出警告信号,使得动力型假肢有足够时间完成保护动作.分别进行绊倒与滑倒实验,结果表明应用此系统的平均预警正确率可达97%.

英文摘要：

The stumble of lower limb amputees was divided into two major categories of tripping and slipping,the potential data sources may be used as stumble pre-warning of amputees was extracted, then the invalid data were excluded with the reference of human body kinematics.Followed by the de-sign of the structure of the stumble pre-warning system,using the principle of support vector ma-chine,and combined with the valid data collected,the corresponding discriminant was concluded,then applying it to the stumble pre-warning system.This method can outgoing warning signals before stumble,making the powered prosthesis have enough time to complete protection action.In addition, tripping and slipping experiments were conducted,and the results showed that the warning accuracy can reach 97% by applying this system.