【作者】 张晓波；

【导师】 王兴松；

【作者基本信息】 东南大学 ， 机械电子工程， 2015， 硕士

【摘要】 体感网是最近十年才兴起的一个概念,并且以相对于传统无线网络具有的明显优势而引起各科研单位广泛的兴趣。本文依托于体感网,设计出了一套用于人体运动测量和评价的便携式设备,该设备可以用于人体在复杂路面(如上下楼梯,上下坡)进行人体运动学和动力学的测量分析。同时也可以针对各种运动类型进行人体氧气消耗的测量,从而对人体运动的强度进行评价。本论文的主要研究工作如下：1、具体介绍了体感网的基本概念及其组成,以及国内外各研究机构以体感网为平台在几大热门领域所做的研究成果。另外专门针对人体爬楼梯介绍了一些最新的研究成果。2、介绍体感网测量系统的硬件组成部分并对各个节点传感器进行误差分析及其标定。本系统主要由7个9轴惯性传感器,一双用于测量人体步态和足底压力的智能鞋,用于测量耗氧量的传感器以及一个总的控制器。各个传感器与控制器之间用CAN总线连接,控制器通过五路WIFI模块与PC机通信,传输速率达到1OOHz。3、介绍体感网测量系统的软件组成部分。首先,利用传感器的测量模型设计了基于四元数的扩展卡尔曼滤波器,对三轴角速度和四元数进行实时更新。然后利用四元数在空间坐标转换的性质设计了依托于本套实验设备的姿态估计算法,实时捕捉人体运动过程中的空间姿态。另外,采用反正切模糊成员函数对人体在爬楼梯时步态进行识别。4、根据拉格朗日原理,利用慢速行走的准静态特征,建立起人体慢速爬楼梯状态下的七连杆的动力学模型。在该模型中将整个步态行走相分为双脚着地相,左脚着地右脚摆动相,左脚着地右脚虚触地相,右脚着地左脚摆动相以及右脚着地左脚虚触地相的5个子步态相。5、利用设计的体感网人体测量系统,对人体爬楼梯状态过程中的步态进行了识别,进而在0kg,5kg,10kg,15kg的负重下进行人体负重爬楼梯实验,并且将实验结果与经典的基于高速相机的运动捕捉技术进行比较,验证了模型的正确性。其次对负重爬楼梯状态下能量消耗进行了实验,并且同平地行走的能量消耗进行了对比,该结果可以作为不同负重下关节力矩大小的一个直观反映,同时也表明爬楼梯时人体的耗能量远远大于在平地行走时的耗能量。更多还原

【Abstract】 Body Sensor Network is a concept which emerged in the latest decade. Compared with traditional wireless network, it has an obvious advantage leading to wide interests from scientific research. Based on the idea of BSN(Body Sensor Network), this paper designs a wearable device, including IMUs, smart shoes and an oxygen sensor, which is used for human motion measurement and evaluation especially in some complex environment, such as climbing stairs. Through this device, some kinematic and dynamic analysis can be done based on some models. Here is the main research content in this paper:First of all. it introduces the basic concept and composition of BSN and the several major researches based on BSN from research institutions at home and abroad. Besides, it introduces the current research results of climbing stairs.Secondly, in the hardware part, this system is mainly composed of 79-axis inertial sensors, a pair of smart shoes used to measure both the human gait and plantar pressure and the sensor used to measure the consumption of oxygen. In addition, there is one controller connecting to each unit by CAN bus and to the PC by WIFI. The transfer rate is up to 100Hz.Thirdly, in the software part, it mainly uses the sensor measurement model to designs an 7-D Extended Kalman Filter based on quaternion to make the angle velocity and quaternion more accurate. Then it designs a new quaternion algorithm to estimate the human low limbs motion. In addition,.it designs a fuzzy control rule using the arctangent fuzzy function for human body gait identification during climbing stairs slowly.Fourthly, according to the principle of Lagrange, it sets up the 7 connecting-rod model of climbing stairs slowly using the quasi static characteristics. The whole process of walking gait in the model is divided into five phases:DS(double support phase), RSLE(right leg support and left leg redundancy phase), RSLT(right leg support and left leg transfer phase), LSRE(left leg support and right leg redundancy phase) and LSRT(left leg support and right leg transfer phase).At Last, it designs three experiments:The first one is to recognize the gait during climbing stairs to verify the validity of the fuzzy controller. Then the experiment of climbing stairs with loads weighing Okg, 5kg,10kg and 15kg is done and the results is compared with the classic method based on high-speed camera motion capture technology to verify the validity of the model. Last one experiment is about the energy consumption during climbing stairs, which is compared with that walking on the flat floor, and it shows it is more energy-consuming during climbing than that walking on the flat floor.更多还原