A Case Study of Spherical Parallel Manipulators Fabricated via Laminate Processes

TL;DR

This study demonstrates that laminated fabrication techniques can produce spherical parallel manipulators with comparable performance to traditional methods, offering cost and time savings while maintaining acceptable precision.

Contribution

Introduces a laminated 2-DOF spherical parallel manipulator and a neural network-based position compensation method, showing laminated fabrication's viability for robotic performance.

Findings

Laminated manipulators can achieve similar precision to conventional ones.

Neural network models effectively compensate for position uncertainties.

Laminated design reduces manufacturing costs and time.

Abstract

This paper evaluates how laminated techniques may be used to replicate the performance of more traditionally manufactured robotic manipulators. In this case study, we introduce a laminated 2-DOF spherical, parallel manipulator. Taking advantage of laminating techniques in the construction of the robot can result in considerable saving in construction costs and time, but, the challenges caused by this technique have to be addressed. By using stiffer materials in rigid links, the rigidity of the robot is increased to an acceptable level. We discuss one method for compensating position uncertainty via an experimental identification technique which uses a neural network to create a forward kinematic model. Final results show that the proposed mechanism is able to track desired rotation with acceptable precision using open-loop model-based control. This indicates that parallel manipulators fabricated using lamination techniques can provide similar performance with prototypes made in conventional methods.