Simulation and Optimisation of a Tendon-Based Stewart Platform
This paper deals with a special type of Stewart Platform. The usual rigid legs are replaced by tendons. By this means, larger scalability and a lighter-weight kinematic structure is made possible. The following disadvantages face this advantages: tendons can transfer only pulling forces and in general a redundant kinematic structure is required. In this paper a dedicated kinematic model is presented, as well as a tool for computing the controllable workspace. Various approaches are described, which yield a better relation of workspace to package space.
Thank you for your interest in tendon-based Parallel Kinematic Machines (You are absolutely right, this one is the correct term. But sometimes Stewart Platform is used in the wider sense as a synonym for PKM and so we used the term Stewart Platform in the title and explained the differences in the text.)
Workspace in our definition is the set of poses where a given range of external forces / moments can be hold by a set of 8 hauling forces of the drives, limited by the moment of the motors. In the examples in the text we considered 10 N in -Z direction i.e. gravitation. In the center of the controllable workspace the range of holdable external forces is usually larger, but at the edge of the controllable workspace a larger external force leads to a indetermind position. In that sense the PKM is stiff in all directions, given the range of external forces / moments does not exceed the specified range.The influence of the stiffness of the tendons, of the platform and of the basis on positional accuracy will be investigated in an ongoing research project.
We have some applications in mind (machine tools, manipulators, construction site equipment) but without a deeper understanding of the limitations of a tendon-based PKM they are pure speculation. So I want to answer that question later.
The position-force control is necessary because in a given pose we need not only a set of determined tendon-lengths but also a set of determined hauling forces. A pure position control results in vanishing forces when the target position is reached. So we investigate different types of forward control in order to meet the exact tendon-length with given non-zero hauling forces.
Thank you for your very clear reply. We wish you well with your project and look forward to reading about your new results at the IPROMS 2006 conference!
Thank you for a most interesting paper. As far as I know, the mechanism as originally described by Stewart had six rigid legs. Your mechanism has 8 tendons. Therefore, is it really a Stewart Platform? Also, please comment on the lateral stiffness of your mechanism and the accessibility of its workspace. What applications are envisaged for the mechanism? Why would force-position control be needed? In general, what is your view of parallel manipulators? What applications are they most suitable for?