How to Reach the Dynamic Limits of Parallel Robots? An Autonomous Control Approach

Abstract

Based on closed kinematic chains, parallel robots obtain favorable dynamic properties as well as high stiffness. Hence, their application can significantly enlarge the productivity of automated production processes. A control concept for tapping the high potential concerning low cycle times and high path-tracking accuracy is presented. The proposed approach adapts autonomously to changing dynamic parameters as varying payload. The autonomous behavior is achieved by combining an adaptive control approach with an adaptive, time-optimal trajectory planning concept and an online-trajectory adaption mechanism. Extensive experimental results prove the performance of the proposed approach.Note to Practitioners-Many applications in the field of production automation (material handling, assembly, etc.) require high operating speeds and accelerations. During the past years, parallel robots proved to be an efficient and suitable supplement to serial robots. Unfortunately, the promising possibilities of parallel robots often cannot yield profit because their dynamic potential is still not fully exploited. The payload/robot mass ratio of parallel structures is even higher compared to serial robots, where the influence of the payload on the impedance of the robot is negligible. By use of direct drives the influence of a variable payload cannot be ignored. A modified adaptive control concept, which adapts autonomously to changing dynamic parameters-as varying payload due to diversity of assembly processes-guarantees high tracking accuracy and therefore better process quality as well as accurate estimates of changing dynamic parameters and therefore better process quality. In addition the productivity of the process can be enlarged, if the full drive power can be used at each point on the path. Thus, a new adaptive time-optimal trajectory planning algorithm is used to exploit the dynamic potential of the direct drives and consequently to shorten the cycle times. The aim of time-optimal trajectory planning, as it is commonly understood, is the determination of the maximum velocity profile along a given path that complies with all given dynamic and kinematic robot constraints like limited drive forces/torques, limited path and/or drive velocities and limited path jerk. Combining the adaptive control scheme and the adaptive, time-optimal trajectory planning algorithm with an online trajectory adaption mechanism, a control concept is realized, which autonomously adapts to changing dynamic robot behavior. Using this new approach, the advantages of parallel robots-as well as serial robots with direct drives-can better be utilized. This is a necessary prerequisite for a larger extension of PKMs for industrial applications.Index Terms-Robot control (RC), parallel robots, time-optimal trajectory planning. Fig. 1. Serial (left) and parallel machine structure (right).

Practical Model-based and Robust Control of Parallel Manipulators Using Passivity and Sliding Mode Theory

“…This does not only cause a severe computational problem but yields high noisy estimation of the velocities and accelerations of the end-effector, even for very small termination conditions of the direct kinematics and especially for the rotational DOFs. The use of the Jacobian and its time derivative to calculate the velocities and accelerations may yield modest or acceptable results for lower-mobility manipulators, like reported in Cheng et al (2003); Pietsch et al (2005); Ren et al (2005) and Vivas and Poignet (2005). The results are however not acceptable for accurate tracking of 6 DOFs mechanisms.…”

“…A second crucial issue for the control of parallel robots is the high complexity of their dynamics, that compromises the real-time implementation of u C . Thus, many researchers suggest the simplification of the dynamics model (Caccavale et al, 2003;Lee et al, 2003;Pietsch et al, 2005;Vivas and Poignet, 2005;Wang et al, 2007) to ensure real-time ability. This will increase the uncertainties to be counteracted by using higher feedback action.…”

Smooth and time-optimal trajectory planning for robot manipulators

“…They determine smooth and time-optimal trajectories by solving the optimization problem numerically, using cubic splines to parametrize the trajectory inside the admissible motion space, and the flexible tolerance method. Pietsch et al [6], [7] provide an algorithm for time-optimal adaptive trajectory planning for parallel robots, which considers an explicit limitation of the pseudo-jerk.…”

Kinematic control of constrained robotic systems

“…The kinematic and dynamic control problems for parallel robots are considered in (Kövecses et al, 2003;Cheng et al, 2003;Rosario et al, 2007) and an autonomous control approach to reach the dynamic limits of parallel mechanisms is presented in (Pietsch et al, 2005).…”