Decoupled Intensity-Based Nonmetric Visual Servo Control

摘要

This brief addresses the problem of vision-based robot control where the equilibrium state is defined via a goal image. Specifically, we consider the class of intensity-based nonmetric solutions, which provide for high accuracy, versatility, and robustness. Existing techniques within that class present a fully coupled translational and rotational control error dynamics, what increases analysis complexity and may degrade system performance. This brief proposes a new intensity-based nonmetric visual servoing technique that decouples the translational control error dynamics, regardless of the observed object characteristics, camera displacements, and their relative poses. The obtained system is thus lower triangular in the general case. For some practical cases, the proposed general technique leads to the Grail of a fully decoupled (i.e., strictly diagonal) linear dynamics. The theoretical analysis of local diffeomorphism, local exponential stability, and those decoupling properties is provided. Improved performances are also experimentally confirmed using a 6-DoF robotic arm in various positioning and tracking tasks.