MULTICRITERIA ANALYSIS OF VISUAL SERVOS VIA BIQUADRATIC LYAPUNOV FUNCTIONS

This paper attempts to propose a generic approach to the muticriteria analysis of position-based and image-based visual servocontrol schemes in robotics. Several criteria can simultaneously be taken into account, such as convergence, avoidance of actuators’ and sensors’ saturations, and 3D constraints. Position- based and image-based servos are unified into a rational systems framework, which is turned into a differential algebraic representation. The constraints are incorporated to the definition of a biquadratic Lyapunov function through the S-procedure and the Finsler lemma, leading to Linear Matrix Inequalities. The method is illustrated on a case study.