Handling visual servoing schemes through rational systems and LMIs

Abstract

This paper proposes a general framework for the analysis and the synthesis of multicriteria visual-based servocontrol schemes. The approach allows to consider position-based or feature-based controllers which can be static, dynamic or gain-scheduled. The visual servoing problem is stated under a state-space representation which is further turned into a rational system and then embedded in a Structured Norm-bounded Linear Differential Inclusion. Many visual servoing requirements may be dealt with, e.g. global stability, avoidance of actuators' saturation, 3 D and image constraints. They often lead to Linear Matrix Inequalities.