Simulating robotic operations where the robot interacts with its environment remains a challenging task because of the difficulties involved in contact modeling. At present, the basic methodologies for modeling contact are well established and have been integrated into many existing multibody dynamics formulations and software. One popular approach involves modeling the normal contact force as a continuous function of deformation, according to a particular constitutive relation. These models are usually simple in form and are easily integrated into any multibody framework because they provide an explicit relationship between the normal contact force and a geometric penetration variable, with the aid of appropriate contact parameters. One issue that remains ambiguous, however, is the choice of the contact parameters to be ‘fed’ into the force-deformation law. In this paper, the problem of contact parameter estimation is addressed in the context of a nonlinear contact force model proposed by Hunt and Crossley. An offline parameter estimation algorithm is developed which effectively transforms the nonlinear estimation problem into a linear one, it in turn solved using a multi-pass recursive technique. Results of application of the algorithm to simulated and experimental data are presented, the latter obtained with a six-DOF robotic manipulator and a variety of payload materials and geometries. Comparison of the proposed method to the nonlinear curve fitting algorithm from MATLAB demonstrates some advantages and limitations.

This content is only available via PDF.
You do not currently have access to this content.