Real-time zone-adaptive control of SCARA-type robot arm

Abstract

This paper presents a piecewise-linear control strategy applied to a two degree-of-freedom (DOF) manipulator performing under variablaload and variablepath conditions. Each axis of the robot arm is modelled as a separate SISO system with a single nonlinear parameter, inertia, which varies according to the load carried and position of the manipulator. The trajectory of the robot arm is divided into zones. Within each zone, an estimate of the inertial load for each axis is obtained. The inertial value for each axis is then assumed to be constant throughout the zone. A PID controller is designed within each zone for each axis which minimizes an ITAE performance index for the estimated value of inertia The result of this algorithm is a robust controller which tunes its parameters at specific times and can withstand a wide range of variations of the payload and path. The controller was implemented on a two DOF manipulator and was found to perform as theoretically predicted. It was also able to reduce manipulator vibration due to mechanical resonance.

Industrial robot track modeling and vibration suppression

“…Feed‐forward approaches typically make use of input and model information to generate control output and to make the plant follow the predefined vibration free trajectory (Piazzi and Visioli, 2000; Kim and Hong, 2004). The open‐loop approaches have to work with close‐loop control to achieve other control objectives such as steady‐state accuracy, system stability and robustness against disturbances and uncertainties, etc.Close‐loop approaches, such as conventional PID control, adaptive PID control, model‐based adaptive control (McEver and Leo, 2001; Coyle‐Byrne and Klafter, 1990; Whitcomb et al , 1993; Book et al , 1976), H ∞ control design (Doyle et al , 1989), etc.The trajectory smoothing approach is one of the most popular approaches currently used in robotic industry. This is mainly due to its simplicity, flexibility and universality.…”

“…Close‐loop approaches, such as conventional PID control, adaptive PID control, model‐based adaptive control (McEver and Leo, 2001; Coyle‐Byrne and Klafter, 1990; Whitcomb et al , 1993; Book et al , 1976), H ∞ control design (Doyle et al , 1989), etc.…”