A Precise Approach Calibrated by Experimental Data to Make an Artificial Impulse

Abstract

Baraff’s and Mirtich’s approaches are well known so as to generate force and impulsive force artificially at collision between rigid bodies. However, both cannot be unfortunately applied for a practical use such as complex dynamic animation and simple haptic rendering. The reason is as follows: Baraff’s approach deals with many forces at multiple contacts, but does not always calculate a solution under static and dynamic frictions. Moreover, since almost natural contact phenomena should be operated by not force but impulse (a sequence of forces) supervised by time, his approach cannot manage such practical aspects. On the other hand, Mirtich’s approach completely neglects time, e.g., the interval of collision, and consequently make only an impulsive force (does not generate a true impulse as force distribution). Consequently, they cannot be applicable to dynamic animation including simultaneous multiple contacts and haptic rendering. For this reason, we propose a new approach based upon integration of an arbitrarily chosen force distribution over a finite time interval by time. Also, we discuss a new method to calibrate completely the force distribution and uncertain parameters required in our new approach. By the deep calibration, the accuracy of linear tangential and angular impulses in our approach is superior to that of these impulsive forces in Mirtich’s approach.