Floating Time Algorithm for Time Optimal Control of Multi-Body Dynamic Systems

Abstract

Moving a dynamic system in minimum time from a given initial state to a desired final state on a prescribed path is one of the oldest and most enduring technological dreams of the scientific and industrial communities. In this research, the problem of bounded-input time optimal control for applied multi-body dynamic systems subject to a full nonlinear dynamical model is solved. To solve the problem, an innovative method, called the ‘floating-time’ method is introduced and utilized. Compared to traditional methods, the floating-time method is an applied method not based on variational calculus. It can be applied to the full nonlinear model of the dynamical system and can handle static and dynamic constraints defined by differential or algebraic equations. The problem of time optimal control is as follows. Find the control law of bounded inputs that drive a given multi-body dynamic system (such as the gripper of a manipulator) along a pre-specified trajectory (in either configuration space or generalized coordinate space) from a given initial position to a given final position, minimizing the time of the motion as a performance index. Using variable time increments, the equations of motion of the system will be reduced to a set of algebraic equations. Searching for a set of time increments (floating-times) that make the equations to exert the maximum available effort produces the minimum possible floating-times, and minimizes the total time of motion. The applicability of the method will be shown by using three examples: a point mass sliding on a rough surface, a 2R robotic manipulator, and the well-known Brachistochrone.