A novel force-velocity field for object manipulation with a model-free cooperative controller

Abstract

When robots grasp a rigid object, homogeneous holonomic constraints give rise to stiff nonlinear constrained dynamics. As a result, a smooth motion is preferred when grasping an object in a tight but gentle manner, making it a difficult control problem that increases when uncertainties and unmodelled dynamics exist. In this paper, a fuzzy design is proposed by exploiting a physics-based orthogonalization of contact mechanics to produce desired velocity and force fields. The proposed scheme enables the regulation of the velocity component in order to navigate smoothly, while the force field enforces grasp at contact in the normal direction of the cooperative joint-velocity field. Then, the robust model-free controller is designed to track such orthogonal fields while enforcing cooperation among all robots, even under uncertainties and unknown dynamics. A representative simulation study is discussed to show the feasibility of the proposal.

Robust contour tracking of nonholonomic mobile robots via adaptive velocity field motion planning scheme

“…This guarantees smooth contour tracking to avoid the radial reduction problem, extending 4,5 to the robust command of wheel velocities, enforcing the asymptotic convergence toward the desired spatial contour. This guarantees smooth contour tracking to avoid the radial reduction problem, extending 4,5 to the robust command of wheel velocities, enforcing the asymptotic convergence toward the desired spatial contour.…”

“…To enforce a given velocity field, an adaptive scheme to deal with kinematic uncertainties is proposed in this paper, whose parametric adaptation contributes to point toward the contour, while dynamic uncertainties are compensated with a robust controller. This guarantees smooth contour tracking to avoid the radial reduction problem, extending 4,5 to the robust command of wheel velocities, enforcing the asymptotic convergence toward the desired spatial contour. Thus, in contrast to the inspiring recent contributions, [12][13][14] our contribution amounts for an adaptive kinematic controller that computes the velocity vector field, such that the desired contour becomes asymptotically attractive despite kinematic uncertainties.…”

“…Recently, velocity fields have been explored to produce operational velocities at each position, 4,5 pointing toward a fixed contour, provided that differential kinematics is known; otherwise, the field is misdirected. Nevertheless, conventional trajectory tracking schemes induce undesirable effects, such as radial reduction, and a lack of coordination among the degrees of freedom of the robotic system since the robot tracks a time-varying point rather than a desired spatial contour.…”

“…Nevertheless, conventional trajectory tracking schemes induce undesirable effects, such as radial reduction, and a lack of coordination among the degrees of freedom of the robotic system since the robot tracks a time-varying point rather than a desired spatial contour. Recently, velocity fields have been explored to produce operational velocities at each position, 4,5 pointing toward a fixed contour, provided that differential kinematics is known; otherwise, the field is misdirected. Velocity fields are considered to solve the problem of radial reduction associated with conventional trajectory tracking schemes, as well as to provide high coordination among all degrees of freedom of the robotic system.…”

A generalised proportional-derivative force/vision controller for torque-driven planar robotic manipulators

“…Over the years, the robotic tasks have been evolving into more complex areas and this causes that robotic systems require of different kind of sensors, as a way to obtain information of their surroundings and their own condition [1,18,24,27,28,29]. Particularly, robotic systems that work in unstructured environments require information that allows them to navigate and interact with the environment.…”