Static shape and vibration control of flexible payloads with applications to robotic assembly

This paper addresses the simultaneous control of vibration and static shape deformation of arbitrarily shaped thin-walled flexible payloads. During robotic assembly of these thin-walled parts, gravity and inertial forces acting on the parts may induce both static shape deformation and vibrations in the part sufficiently large that accurate and high speed assembly of these parts is hindered. Static deformations, which arise due to deformation of the part caused by its own weight under the influence of gravity, lead to misalignment of mating points of the parts. Unwanted vibrations, arising from inertial forces acting on the thin-walled parts as they are positioned for assembly, must damp out before parts can be mated, further hindering the process. In this work, a smart gripper with actuated contact points to grasp the flexible thin-walled parts is proposed to solve this problem. The smart gripper is capable of both part reshaping and active damping of unwanted vibrations of the part. It is fixed to a robotic manipulator and is comprised of multiple linearly actuated fingers with laser-based noncontact proximity sensors, and associated signal processing and controllers. In this paper, a simultaneous vibration and static shape controller is developed. The proposed controller is a composite modal controller in conjunction with a quasi-static modal filter and a bias Kalman filter, which is synthesized based on the reduced-order dynamic model of the flexible payload. A near industrial practice demonstration of the feasibility of the proposed approach is carried out using a proof-of-concept smart gripper to manipulate an automotive fender. Experimental results indicate that unwanted vibrations are successfully damped out, allowing faster cycle times for an assembly process, and static shape deformations are corrected, allowing accurate positioning of parts for assembly.

I. Introduction

Interest in robotic manipulation of flexible payloads has grown more recently due to its potential applications in many assembly processes in industry. Assembly of thin-walled parts, with robots replacing costly fixtures, has great potential for cost savings in the aircraft and automotive industries [1]. In these industries, thin-walled parts are traditionally held in costly hardware fixtures that accurately locate parts prior to mating with other parts. In a robotic assembly process, when manipulated by robots, these thin-walled parts will deform under their own weight due to gravity and vibrate due to inertial forces that act on them as they are positioned for assembly. When a payload is thin walled (e.g., 1–2 mm in thickness) and its remaining dimensions are sufficiently large, its structural flexibility creates difficulties in ensuring mating surfaces and contact points are properly aligned for successful assembly. Furthermore, because subsequent assembly operations are delayed until the vibration decays, the reduction of vibration amplitude settling time is critical for high-speed assembly. Fig. 1. Smart gripper manipulating flexible payload