مهندسی مکانیک شریف (Nov 2017)
TRAJECTORY OPTIMIZATION OF FLEXIBLE REDUNDANT ROBOT MANIPULATORS BY SEPARATION OF ELASTIC VIBRATION
Abstract
In this article, trajectory optimization of robot manipulators with additional degrees of freedom is studied. In the suggested method, all possible answers are obtained based on rigid conditions and redundancy of robot. Then, by using open-loop optimal control approach, the trajectory which minimizes the objective function will be selected. Due to flexibility, the dimensions of system are unlimited. So, the assumed mode method is used for discretization. By calculating kinetic and potential energy of system and based on Lagrange formulation, dynamic equations of motion are extracted. Given that global optimization is more accurate and efficient in comparison with local optimization, global optimization is used in this research. The objective function is defined as an integral index and will be minimized in the entire path. The objective function and constraints of optimization problem will be selected based on conditions of point-to-point motion or the specified trajectory between two fixed points. In two states, reduction of elastic vibrations of flexible members is considered as one of the main goals of issue. Dynamic equations of system are constraints of optimization problem in point-to-point motion. For condition of motion in the specified trajectory, kinematic equations are added to dynamic constraints. Also, unequal constraints are applied for bounding the velocity and torque magnitudes. By selecting the state and control signal vector which are obtained from assuming rigid motion of robot, the objective function and constraints will be changed to the standard form of an optimization problem. To solve the equations, Pontryagin's minimum principle is used. By using this principle, the equations of the classic form will be changed to a set of first order of differential equations, and finally, to a two-point boundary value problem. Given that the suggested method for point-to-point motion and movement in the specified trajectory is provided; therefore, two examples in simulation are considered for each. Results indicate accuracy and efficiency of the suggested method.
Keywords
