面向全自主智能机器人的无标定图像视觉伺服控制研究

Research on uncalibrated image visual servo control for fully autonomous intelligent robot

摘要:

在利用全自主智能机器人检测、搬运目标物的过程中,为了提高主机元件对机器人的伺服控制效果,设计了一种无标定图像视觉伺服控制方法。首先,根据机器人轨迹的无标定估计结果,计量目标物的位姿点,推导出机器人无标定图像中的目标物检测与追踪条件。按需连接视觉伺服控制器,通过求解机器人无标定运动内积的方式,计算图像视觉特征的伺服选择标准,实现对机器人无标定图像视觉特征的伺服选择。然后,在无标定图像中定义雅克比矩阵,根据图像视觉分割原则,完成机器人图像角点的伺服匹配处理,再通过确定强化控制参数的取值范围,实现无标定图像的视觉伺服控制。根据实验可知,应用该方法可以解决智能机器人难以运动至标定区域的问题,为提高伺服控制指令的执行有效性提供了保障。

Abstract:

In the process of using fully autonomous intelligent robots to detect and track target objects, in order to improve the servo control effect of the host components on the robot, an uncalibrated image visual servo control method was designed. Firstly, based on the uncalibrated estimation results of the robot trajectory, the pose points of the target object are measured, and the detection and tracking conditions of the target object in the uncalibrated image of the robot are derived. Connect the visual servo controller as needed, calculate the servo selection criteria for image visual features by solving the inner product of robot uncalibrated motion, and achieve servo selection for robot uncalibrated image visual features. Then, a Jacobian matrix is defined in the uncalibrated image, and according to the principle of image visual segmentation, servo matching of robot image corners is completed. Then, by determining the range of reinforcement control parameters, visual servo control of the uncalibrated image is achieved. According to the experiment, the application of this method can solve the problem of intelligent robots difficult to move to the calibration area, providing a guarantee for improving the effectiveness of servo control command execution.