Dynamic Pose-Estimation for Visual Control

Abstract

In this chapter, we change fromthe image-based approach to estimate the position and orientation of the camera-robot system in order to regulate the pose in the Cartesian space. This provides the benefits of including a kind of memory in the closed loop, which reduces the dependence of the control on the visual data and facilitates the planning of complex tasks. The camera-robot pose is recovered using a dynamic estimation scheme that exploits visual measurements given by the epipolar geometry and the trifocal tensor. The interest of the chapter is a novel observability study of the pose-estimation problem from measurements given by the aforementioned geometric constraints, as well as the demonstration that the estimated pose is suitable for closed loop control. Additionally, a benefit of exploiting measurements from geometric constraints for pose-estimation is the generality of the estimation scheme, in the sense that it is valid for any visual sensor obeying a central projection model. The effectiveness of the approach is evaluated via simulations and real-world experiments.