论文部分内容阅读
针对四自由度串并混联拟人空间机械臂本体,构建了基于运动控制卡模式的机器人软件和硬件系统。采用基于位置的控制结构,利用所求得的机械臂逆运动学方程,设计了由图像反馈和末端运动组成的视觉伺服控制系统。提出了一种基于SURF特征的静态目标识别算法,算法首先对目标图像提取SURF特征,并利用欧氏距离实现模板图像与目标图像特征点匹配,然后计算已匹配特征点的质心来获得目标的位置信息,最终实现了稳定的伺服定位。实验结果证明了系统的可靠性,并且证明了四自由度串并混联机械臂本体在实际运动的精确性和稳定性。
Aiming at the four-DOF serial-parallel hybrid humanoid robot arm, a robot software and hardware system based on the motion control card mode is constructed. Based on the position-based control structure, a visual servoing control system consisting of image feedback and end-point motion was designed by using the inverse kinematics equation of the manipulator. A new SURF-based static target recognition algorithm is proposed. Firstly, the SURF feature is extracted from the target image and the Euclidean distance is used to match the template image with the target image feature point. Then the centroid of the matched feature point is calculated to obtain the target position Information, the final realization of a stable servo positioning. The experimental results prove the reliability of the system and prove the accuracy and stability of the four degrees-of-freedom serial and hybrid manipulator body in actual motion.