【作者】 杨波；

【导师】 蒋体钢；

【作者基本信息】 电子科技大学 ， 通信与信息系统， 2014， 硕士

【摘要】 实时跟踪捕获人体的动作在惯性导航、生物医学、虚拟现实、人机控制、体育运动等领域,是一项重要的科研课题。目前的动作捕捉系统主要有光学动作捕捉系统、磁性动作捕捉系统、机械动作捕捉系统、惯性动作捕捉系统等,惯性传感设备能够克服其他传感设备存在的抖动、延迟干扰、缓慢漂移、限制运动范围的问题。动作捕捉系统对人体动作的捕捉主要分为2个方面:人体动作姿态的估计和人体的轨迹追踪定位。轨迹追踪系统是动作捕捉系统的补充,主要是为了研究轨迹追踪算法使用。动作捕捉系统的一般性结构主要分为三个部分:数据采集、数据传输、数据处理。本文首先对动作捕捉系统和轨迹追踪系统的整体框架进行了设计,从硬件选型和节点功能模块的设计论证了系统的可行性。然后对现有的姿态估计算法进行了分析和研究,从算法的复杂度和执行效率方面考虑,选取Mahony滤波算法作为本文姿态估计的算法。然后对动作捕捉系统中数据采集节点、中继节点、服务器端进行了软件功能上的实现。本文研究了用于人体动作捕捉系统中人体模型建立的基本方法,在服务器端建立人体三维骨骼棍状模型,模型的控制以OpenGL三维渲染为基础,基于Visual Studio 2010 MFC平台,采用C++语言实现。在实现动作捕捉系统之后,测试者穿戴MEMS惯性传感设备进行了手臂动作的测试实验,实验结果表明在服务器端能够正确显示人体手臂的动作,实现了人体动作捕捉系统的基本功能。轨迹追踪是对动作捕捉系统的补充,首先对现有的基于MEMS惯性传感器的轨迹追踪算法进行了研究和分析,本文提出了一种改进的轨迹追踪算法,实验仿真结果表明,改进后算法的平均误差仅为2.22%,因此,该算法更能适合实际的应用场景。更多还原

【Abstract】 Real-time tracking of human motion capture in education, human-computer interaction, video animation, medical research, competitive sports, games, sports, military, scientific and other fields, is an important research topic. With the continuous development of science and technology, efficient, fast, and accurate capture human motion behavior has become a reality, the current motion capture system has optical motion capture system, magnetic motion capture systems, mechanical motion capture system, inertial motion capture system, sensing device is able to overcome the inertia of his presence sensing device jitter, delay interference, slow drift problem of limiting the range of motion. This paper mainly discusses the MEMS-based inertial sensing equipment related to human movement capture technology and system design and implementation trajectory tracking, inertial sensors and wireless transmission systems based on simple structure, low cost, easy to operate, Easy to wear, high-precision, real-time advantages.Firstly, the motion capture system and the overall framework of trajectory tracking system has been designed, from the design of hardware and node selection function module demonstrates the feasibility of the system. Then the existing pose estimation algorithm analysis and research, from the complexity and execution efficiency of the algorithm, the selected Mahony complementary filter algorithm as herein pose estimation algorithm. Then the motion capture system, data acquisition nodes, relay nodes, server-side implementation of software functionality. The basic method used to study the human body motion capture system model, according to the basic dynamic theory to establish control stick like a three-dimensional model of the human skeleton model on the server side with OpenGL 3D rendered, based on Visual Studio 2010 MFC platform using C++ language. After the realization of the motion capture system, testers wear MEMS inertial sensing devices were tested experimental arm movements, the experimental results indicate that the server is able to display the correct human arm movements to achieve the basic functions of the human body motion capture system.Trajectory tracking is a supplement to the motion capture system, the first track on the existing MEMS inertial sensor tracking algorithm based on research and analysis, this paper presents an improved trajectory tracking algorithms, simulation results show that the improved algorithm Average error of only 2.22%.更多还原