In this paper, we present an experimental study concerning the gravity compensation of flexible link arms based on linear springs. In the field of flexible link robotics, the gravity compensation based on counterweights has been successfully applied in the past, but little effort has been made to examine the potential benefits and difficulties of using spring-based compensation mechanisms. This paper focuses on the modeling and identification of a single link flexible arm compensated with a spring based mechanism. As modeling approach, we followed the lumped-mass methodology to develop a model capable of reproducing the first vibrational frequency of the flexible link arm. Keeping in mind the forces that interact with the flexible link, a combination of sensors is suggested in order to measure and estimate the most important variables of the system. Subsequently, a very simple and reliable identification method based on the time and frequency response of the system is proposed. Finally, the results of the modeling and identification are validated on our experimental platform.

Despite the benefits that the spring based gravity compensation mechanism has brought to the field of rigid robotic manipulators, there have been no substantial efforts toward transferring these developments to the field of flexible link robotics. In this paper, we present an input state feedback linearization controller for the tip positioning of a flexible link arm with a gravity compensation system based on springs. The controller is implemented into a double loop control scheme, in which the inner loop addresses the motor position control in presence of joint friction, and the outer loop deals with vibration cancellation and the tracking of fourth-order trajectories for the tip position of the flexible arm. Taking into considerations the interacting forces between the flexible link and the gravity compensation mechanism, and also the characteristics of the control law, we propose a sensory system to measure all the relevant signals. The proposed controller is tested on an experimental prototype built in our laboratory.