The Bouncing Ball Apparatus as an Experimental Tool

Abstract

The bouncing ball on a sinusoidally vibrating plate exhibits a rich variety of nonlinear dynamical behavior and is one of the simplest mechanical systems to produce chaotic behavior. A computer control system is designed for output calibration, state determination, system identification, and control of a new bouncing ball apparatus designed in collaboration with Magnetic Moments. The experiments described here constitute the first research performed with the apparatus. Experimental methods are used to determine the coefficient of restitution of the ball, an extremely sensitive parameter needed for modeling and control. The coefficient of restitution is estimated using data from a stable one-cycle orbit both with and without using corresponding data from a ball map. For control purposes, two methods are used to construct linear maps. The first map is determined by collecting data directly from the apparatus. The second map is derived analytically using a high bounce approximation. The maps are used to estimate the domains of attraction to a stable one-cycle orbit. These domains of attraction are used to construct a chaotic control algorithm for driving the ball to a stable one-cycle from any initial state. Experimental results based on the chaotic control algorithm are compared and it is found that the linear map obtained directly from the data not only gives a more accurate representation of the domain of attraction, but also results in more robust control of the ball to the stable one-cycle.