基于变长倒立摆模型的双足机器人步态研究与实现

发布时间：2018-03-09 12:34

本文选题：双足机器人　切入点：变长倒立摆　出处：《天津理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：双足机器人的步态控制研究一直是智能控制的热门领域,又由于双足机器人的高自由度、高耦合的特点,成为智能控制领域的难点问题之一。对于双足机器人的研究是一个结合机械制造技术、自动化智能控制技术、信息电子技术、软件控制技术等多种学科,多种技术的融合课题。针对双足机器人斜坡行走以及上台阶等非水平情境下稳定性较差的问题,本文采用NAO双足机器人作为实验平台,在控制算法上提出应用光滑函数拟合的思想,区别于三次样条插值,避免了速度和加速度的突变,保证了双足机器人活动的平稳性,并完成了上台阶步态的仿真实验。本文致力于完成了以下工作:首先,进行基于变长倒立摆模型的建模研究。结合NAO机器人的固有参数,将下肢的大腿、小腿、双脚和髋躯干等效为七连杆模型,依照上台阶步态质心和关节角变化特点,建立变长倒立摆模型以达到物理过程等效、简便计算的目的。然后,针对应用三次样条插值算法引起加速度突变的问题,本文提出采用光滑函数拟合的方法。本方法是对踝关节轨迹进行规划,并结合运动学和逆运动学原理,求解得到上台阶过程中下肢的各个关节角序列。再次,对NAO机器人行走的ZMP稳定性进行验证。分别通过实验阐释本课题选用光滑函数拟合与三次样条插值算法对踝关节进行规划在运动稳定性上的差异,光滑函数你和方法提高了ZMP的稳定裕度与行走的稳定性。最后,搭建NAO深层开发环境,完成NAO连续上台阶步态实验。本文应用Matlab、VS2010、Cmake编译器、Webots PRO以及Choregraphe等相关软件,搭建了双足机器人的深层开发环境,利用NAO的控制台程序完成模拟机和样机的步态控制与实现。
[Abstract]:The research on gait control of biped robot has always been a hot field of intelligent control, and because of the characteristics of high degree of freedom and high coupling of biped robot, The research of biped robot is a combination of mechanical manufacturing technology, automatic intelligent control technology, information technology, electronic technology, software control technology and so on. Aiming at the problem of poor stability of biped robot in non-horizontal situations, such as slope walking and up steps, the NAO biped robot is used as the experimental platform in this paper. In the control algorithm, the idea of smooth function fitting is put forward, which is different from cubic spline interpolation, avoids the mutation of velocity and acceleration, and ensures the stability of biped robot. The following work is done in this paper: firstly, the modeling based on variable length inverted pendulum model is studied. Combined with the inherent parameters of NAO robot, the thigh and lower legs of the lower limbs are studied. The feet and the hip trunk are equivalent to the seven-bar model. According to the characteristics of the gait center of the upper step and the angle of the joint, an inverted pendulum model with variable length is established to achieve the purpose of physical process equivalence and simple calculation. In order to solve the problem of acceleration catastrophe caused by cubic spline interpolation algorithm, a smooth function fitting method is proposed in this paper. This method is used to plan the ankle trajectory and combine the kinematics and inverse kinematics principles. The angle sequence of the lower extremity in the process of stepping up is obtained. Again, The ZMP stability of NAO robot walking is verified. The difference of motion stability between smooth function fitting and cubic spline interpolation algorithm for ankle joint planning is explained by experiments. The smooth function can improve the stability of ZMP and the stability of walking. Finally, the deep development environment of NAO is built, and the step up gait experiment of NAO is completed. In this paper, we use Matlab / VS2010 / C make compiler, Webots PRO and Choregraphe, etc. The deep development environment of biped robot is built, and the gait control and realization of the simulator and prototype are accomplished by using the NAO console program.
【学位授予单位】：天津理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【相似文献】