Abstract
This paper presents a comparison between a kinematic observer approach and a dynamic observer approach for leader-follower synchronization control of mechanical systems formulated in the Euler-Lagrange framework. The kinematic observer utilizes a virtual system that is designed to mimicthe behaviour of the leader to estimate the state vector with a minimum of information available, while the dynamic modelbased observer utilizes a model-based controller as part of the correction term in the observer. The two observers are compared in terms of estimation principle and practical performancefor an underway replenishment operation for surface ships.