Predictive control architecture for real-time image moments based servoing of robot manipulators

Abstract

This paper presents a real-time architecture for visual servoing of robot manipulators using nonlinear based predictive control. In order to increase the robustness of the control algorithm, image moments were chosen to be the visual features which describe the objects from the image. A visual predictive architecture is designed to solve tasks addressed to robot manipulators with an eye-in-hand configuration. An implementation of the proposed architecture is done so that the capabilities of a 6 d.o.f robot manipulator are extended. The results of different experiments conducted with two types of image moments based controllers (proportional and predictive with reference trajectory) are presented and discussed

Development of a neural network structure for identifying begin-end points in the assembly process

“…In assembly process analysis, human detection and object detection methods can be useful. Image analysis of components in the production process can be performed using different types of visual features, which are often divided into: point features (centroids or corners), line or ellipse features, and generic descriptors (image moments) [14,15]. Image moments represent generic visual features that can be computed easily from a binary or a segmented image, or from a set of point features [14] [16].…”

Integrating Features Acceleration in Visual Predictive Control

“…Regarding the weight associated to the features, we adopt the diagonal weighting matrix Q k = α k I, α being a tunable parameter, similarly to what was done in related works. In particular, in [18], [19] it was proposed to use α = 1/e, corresponding to a decreasing weight of features samples. This was justified by the fact that in these works, the authors were also integrating an on-line trajectory generator, and the decreasing factor ensured better tracking of the reference.…”

MPC Control and LQ Optimal Control of A Two-Link Robot Arm: A Comparative Study

“…The robot's trajectory was calculated on-line by using the model predictive method. In [21], a nonlinear predictive control structure in real-time for the visual servoing of a manipulator robot was proposed. The proposed control approach was designed to solve tasks assigned to robot manipulators with an eye-in-hand configuration.…”