A Family of Novel Fast Algorithms to Improve Computing Efficiency of Set-Based Direct Visual Servoing

Visual servoing (VS) is a technique which uses feedback information obtained from a camera to control the motion of a robot. With respect to approaches that rely on feature-based perception processes to recover the relative pose between the target and the robot, direct visual servoing (DVS) does not require pose estimation, nor featur extraction, tracking and matching, but uses the images as a whole feature directly. Set-based DVS (SDVS) algorithms are based on theory of mutation analysis of shapes or sets and these algorithms have been proposed by researchers recently. However, the computational efficiency for SDVS is relatively slow. In this paper, we proposed a family of novel fast algorithms to improve computing efficiency of SDVS. Simulation results demonstrate the validation of our algorithms as well as the computation efficiency advantage over other SDVS work.