Computationally efficient parametric analysis of discrete-time polynomial based acceleration-deceleration profile generation for industrial robotics and CNC machinery

摘要

Accurate position control and its related dynamics have been under research for industrial robotics and CNC machinery applications. Several works have developed techniques to fulfil the trajectory planning and profile generation from the CADCAM point of view, however, there is a lack in the mathematical foundations on computationally efficient profile generation for implementation in digital systems. This paper shows a novel approach for acceleration-deceleration profile generation based on polynomials at the discrete-time domain which is computationally efficient and easy to implement in most digital system available technologies. Results in the form of simulations and a servomotor driving the axis position on a CNC milling machine are presented to show the procedure efficiency and jerk reduction. It is shown that the proposed parameterisation of polynomial based profile generation requires half the operations to be computed, compared with the currently available methodologies for direct polynomial evaluation. Furthermore, the proposed procedure is multiplier free which means that almost every low-cost processor or embedded digital system can easily perform the generation task.