PODs Conceived:

A few of the sketches below appeared in our original proposal to the National Science Foundation. They do not look like machines. These sketches identify some of the kinematic architectures, or mechanical skeletons, for the machines we are investigating. From left to right, these machines are the following: The Spherical Four-Bar (s4R), The Coupler-Driven Spherical Four-Bar (CDs4R), The Prismatically Mounted Spherical Four-Bar (Ps4R), and the The Prismatically Mounted Coupler-Driven Spherical Four Bar (PCDs4R).



s4R
CDs4R
Ps4R
PCDs4R

The Need for PODs:

As a product is assembled in an automated factory, both the product and its individual parts are picked up, reoriented and inserted into subassemblies or fixtures. For a complex product, the number of manipulations could run into the thousands. Parts are picked out of bins and placed into assemblies. Partial assemblies are rotated to allow additional parts to be added. Fasteners are inserted to hold it all together. The machines that perform these tasks are selected for their speed, precision, cost and reliability.

Assembly line designers try to keep the manipulations as simple as possible. Rotations about vertical or horizontal axes are preferred, often of 90° or 180°. These tasks have a well established set of solutions.

However, operations which involve a translation along and a rotation about an axis which is not vertical or horizontal is more challenging to the designer. Additional constraints on the trajectory of the object (e.g. obstacle avoidance, or part meshing) increase the difficulties. One solution is to use devices with a high number of degrees of freedom, such as robots. These mechan­ically complex devices perform the tasks, but at penalties in dollars, setup time and maintenance. A second solution is to use a cascading series of simple manipulators. Creating this manipulation pipeline takes a longer design time, and is often more art than science.

This research offers another solution. Moving an object from one place to another doesn’t require six degrees of freedom if the motion of the mechanism is designed with the task in mind. This research will create low degree-of-freedom machines capable of producing spatial trajectories and rotations. A single “part orienting device” (POD) can be used in an assembly task that might otherwise require a robot or multiple single degree-of-freedom mechanisms.

Spatial devices similar to the proposed PODs exist in research settings, but have yet to be well utilized in a practical environment. The design and manufacturing issues have been too daunting. This research will comprehensively address the design of PODS.
  1. A kinematic synthesis method will be derived to enable the designer to create PODS for a prescribed task.
  2. An interactive design system will be created to allow the designer to navigate a broad array of choices in a goal-oriented, immersive system.
  3. Tools will be created to provide the designer with prototyping and actuating recommendations.

PODs Embodied:

Actually, the PODs above were not the original starting point for the research. Our original concept more closely resembled the machine to the left. Click on it to see the animation of what it does.

Note that the robot moves the device through space while performing an intriguing reorientation of the part in the gripper. The reorientation combines a 90 degree rotation about the axis perpendicular to the floor with a (followed by) a 90 degree rotation about an axis parallel to the long edge of the light brown conveyor (table). By the way, this kind of 90 degree rotation about axis 1 followed by a 90 degree rotation about axis 2 is very common in industry.

This was our original concept because it has only two degrees-of-freedom. That is, we can move and control it with only two motors. This is called a PR open chain. P stands for prismatic or sliding joint. R stands for revolute or hinge joint. In fact, we can design a PR open chain to move a part in a gripper from any start location and orientation to any end location and orientation.

So, why isn't the PR chain a useful POD? Because it offers no control over the initial or final trajectory. Notice how the part slides off of one conveyor and slides on to the other? To be useful, the part needs to move, generally, upward and downward on to or off of the conveyors. The PR chain does not allow for that freedom in its design.

A spherical four-bar mechanism (s4R) is shown to the right. This mechanism can be designed to reach two orientations. Moreover, it can be designed to allow for control over the trajectory into and out of those orientations.

An s4R, on the other hand, cannot be designed to reach two arbitrary locations. (This may be a good time to review the spatial assembly challenge.) There has been a significant amount of research into the theoretical aspects of s4R mechanisms. There is far less research regarding their practical design and implementation. Designing and building usable s4Rs is critical to this research.

What are the challenges to practical implementation and design?
· They simply cannot be designed on a piece of paper. Hence, designers need sophisticated software.
· The mechanical design of complex spatial devices with low degrees-of-freedom has not been performed enough to have a set of useful heuristics.
· The internal forces encountered in prototypes seem restrictively high.

The figure to the left shows some early work in developing heuristics for the mechanical design of s4Rs. From top to bottom, they represent progressively easier contructs for making one link.

This "tinker toy approach" conceives of pieces of the mechanism in as standardized a way as possible. Each concept, ultimately, does require some custom machining. Also worth noting is that the closer the link resembles a circular arc, the easier it is to design a more compact mechanism. So, the link at the top, although requiring the most machining, may lead to the most compact design. Again, the heuristics are in development.

The figure to the right shows an s4R constructed using one of these tinker toy approaches. These methods provide a quick and dirty tool for assembling and playing with s4R devices, but they are by no means a functioning machine!

Implementing successful machines is a problem we are tackling on all fronts, from the generation of new theory and models, to the creation of new design software, to simply designing and building new machines. Please take the time to explore the world of PODs. Finally, don't hesitate to offer suggestions and ask questions: murray@udayton.edu.
The Spatial Assembly Challenge:

A triangular widget is shown to the left. The widget has an intial location and a final location. Note that the way the triangular widget nests with its environment is different at the initial and final locations.

The spatial assembly challenge is to design a machine that can grasp the widget at the initial location, move it through space while flipping it around, and insert the widget into the final location. Performing this operation quickly with a modest number of degrees-of-freedom is the goal of PODs research.

In more technical terms, the spatial assembly challenge is a two pose spatial synthesis problem with constraints. On the first pose, there is an exit trajectory indicating, roughly, the direction in which the part should move. On the final pose, there is an entry trajectory indicating, roughly, the direction from which the part should arrive. These are rough constraints as the practical situation almost never demands a specific direction of motion. Implicit in these trajectories is that the angular velocity of the part will be small at the intial and final poses. Finally, as indicated by the gray arrow, there may exist a rough notion of the path the part should follow. Again, the path is considered approximate as the practical situation almost never demands a specific path for the part.

 

This material is based upon work supported by the National Science Foundation under Grant No. 0422731 & 0422705.

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.