Comparison of two feedforward design methods aiming at accurate trajectory tracking of the end point of a flexible robot arm

This paper discusses the design and properties of two trajectory tracking controllers for linear time-invariant systems, and compares their implementation and experimental results on a flexible one-link robot equipped with a velocity-controlled actuator. High positioning accuracy and low tracking errors within a specified bandwidth are their performance specifications. Both controllers use the same state feedback controller, but have a different feedforward design approach. Both feedforward methods design stable prefilters which approximate the unstable inverse system model. The first method designs a stable prefilter using the extended bandwidth zero phase error tracking control (EBZPETC) method [11], which is an extension of Tomizuka’s well-known zero phase error tracking control (ZPETC) method [10]. The second feedforward method adds delay to the inverse model and then uses common filter design techniques to approximate this delayed frequency response.