仿人机器人手臂动作模仿系统的研究与实现

Research and Implementation of Humanoid Robot Arm Movement Simulation System

为实现最直观的人机交互方式,与仿人机器人的动作模仿,建立了仿人机器人系统。利用Kinect体感摄像机的骨骼追踪技术,采集示教者的重要关节骨骼点三维空间坐标,对采集到的数据进行霍尔特指数平滑滤波处理。基于骨骼数据建立机器人手臂重要关节的数学模型,将骨骼数据转换成相应伺服电机控制信息。并发送到仿人机器人主控制器驱动各关节伺服电机转动,实现动作模仿。实验结果表明,所述建模方法适用于仿人机器人手臂动作模仿系统,能够实现实时动作模仿,且运动平滑无抖动,动作模仿效果较好。

In order to achieve the most direct human-computer interaction and humanoid robot movement simulation,establishes a humanoid robot system. Three-dimensional coordinates of important joints of the instructor are acquired by skeletal tracking technique of Kinect camera,the collected data are processed by Holt smoothing filter. The mathematical model of important joints of the robot arm is established using the skeleton data and the skeleton data is converted into the corresponding servo motor control information. Then sent to the humanoid robot main controller to drive the joint servomotor,realizing the movement simulation. The experimental results illustrate that the modeling approach described in this article is applicable to humanoid robot arm simulation system,which can simulate real-time motion with smooth and stable motion trajectory. And this motion simulation has good effect.