机器人视觉伺服控制研究进展与挑战

The Development and Challenges of Studies on Robot Visual Servo Control

针对机器人视觉伺服研究,分别从基于位置的视觉伺服、基于图像的视觉伺服、混合视觉伺服等对视觉伺服的研究进展进行了回顾.从雅可比在线估计法、雅可比自适应估计法、深度独立雅可比矩阵估计法等多方面详细分析了无标定视觉伺服.在此基础上,就视觉伺服的实时性、可靠性、多传感器融合、多任务控制、基于离散时间域及混杂系统的控制器等,对机器人视觉伺服控制所面临的挑战及进一步研究进行了分析与展望.

The survey on robot visual servo was carried out based on position-based visual servo,image-based visual servo,and hybrid visual servo respectively.Moreover,the analysis of uncalibrated visual servo was conducted including Jacobian on-line estimation method,Jacobian adaptive estimation method,depth independent Jacobian matrix estimation method,and other uncalibrated visual servo methods.Then,the challenges of robot visual servo control were analyzed including real-time performance,the reliability,multi-sensor fusion,multitask control,controller design based on discrete time domain and hybrid system,and so on.