基于支持向量机的跌倒判定研究.docVIP

PAGE \* MERGEFORMAT 2 基于支持向量机的跌倒判定研究 刘洋，数学计算机科学学院 摘 要：跌倒严重影响老年人的身体健康,尽早判定老年人是否跌倒对及时救援和避免并发症至关重要。利用智能手机的加速度计、陀螺仪和旋转矢量传感器等运动传感器采集人体运动信息，提出一种基于支持向量机的跌倒检测方法。该方法利用支持向量机分类方法和LIBSVM开源软件包对由115种运动特征向量构成的训练集和测试集进行训练和测试，实现跌倒判定。实验结果表明这种方法的有效检测率较高，具有一定的实用价值。 关键词：跌倒判定；支持向量机；运动传感器；运动特征向量；LIBSVM Fall Detection Research Based of the Support Vector Machine Liu Yang, College of Abstract：The falls affects the elderly health seriously, the early detection of falls is very important to rescue the patient and avoid badly complication. In this paper, we use accelerometer, gyroscope and rotation vector sensor in smart phone to capture the movement data of human body, and propose a novel fall detection method based on support vector machine. The method utilizes the classification method based on support vector machine and the LIBSVM open source software package to train and test the data set which consists of 115 species of sport feature vectors, and realizes the fall detection. Experimental results show that the valid detection rate using this method is high, and has a certain practical value. Key Words: Fall detection; Support vector machine; Sport sensors; Sport feature vector; LIBSVM TOC \o 1-3 \h \z \u HYPERLINK \l _Toc354328378 第一章 绪论 PAGEREF _Toc354328378 \h 1 HYPERLINK \l _Toc354328379 1.1研究背景 PAGEREF _Toc354328379 \h 1 HYPERLINK \l _Toc354328380 1.2本文研究目标和内容 PAGEREF _Toc354328380 \h 2 HYPERLINK \l _Toc354328381 1.2.1 研究目标 PAGEREF _Toc354328381 \h 2 HYPERLINK \l _Toc354328382 1.2.2 研究内容 PAGEREF _Toc354328382 \h 2 HYPERLINK \l _Toc354328383 1.3相关技术 PAGEREF _Toc354328383 \h 2 HYPERLINK \l _Toc354328384 1.4本文结构安排 PAGEREF _Toc354328384 \h 7 HYPERLINK \l _Toc354328385 第二章 基于SVM的跌倒判定原理 PAGEREF _Toc354328385 \h 7 HYPERLINK \l _Toc354328386 2.1特征值 PAGEREF _Toc354328386 \h 7 HYPERLINK \l _Toc354328387 2.2 SVM跌倒判定原理 PAGEREF _Toc354328387 \h 9 HYPERLINK \l _Toc354328388 第三章 基于SVM的跌倒判定实验方案设计 PAGEREF _Toc354328388 \h 11 HYPERLINK \l _Toc354328389 3.1实验目的 PAGEREF _Toc