Students: Mike Taylor Karl Stensvad Aaron Ferber
Advisors: Ed Colgate Michael Peshkin		This material is based upon work supported by the National Science Foundation under Grant No. 0413204.

Approach 1: Controlled Buckling

Motivation: The goal of this project is to develop a haptic field display based on the concept of varying the compliance of the surface with which the user interacts. This is done in the hopes of improved performance compared to typical haptic field displays that are based on height-varying pin arrays.

Background: Most previous haptic field displays have made use of active pin arrays. These arrays of closely packed pins use actuators to move individual pins up and down. By doing so, the shape of the surface created by the collective ends of the pins changes and conveys information to the user. An example device is shown below. As can be seen, developing this type of device with a large work area would require many pins and thus many actuators. This would cause the haptic field display to be overly large, heavy and to draw excessive power.

A Height Varying Pin Array designed and built by students at Northwestern University		Close up of the surface developed by the pins

Potential Advantages: By varying the compliance of the surface instead of the heigth, it is hoped that a smaller, lighter, and more efficient tactor can be made. Additionally, it may be possible to make the HFD with a much larger work area than current designs.

Technique

The current design replaces the pin and actuator pair used in a typical pin array with small beams of plastic held against a rigid wall. When the user pushes down on the top of the plastic stip, as shown below, the stip buckles. This then feels like a very soft spring.

However, if the plastic strip can be prevented from buckling, it will resist the finger forces applied by the user with very little movement. This will the feel like at very stiff spring or a rigid surface.

The strip is prevented from buckling through the use of electrostatic pressure to hold it against a flat wall. By employing a beam of plastic with a conductive side, in this case 0.002 inch thick Mylar, a voltage can be applied between this conductive side and an electrode on the wall. The resultant electrostatic pressure holds the beam against the wall.

Under normal conditions, the user's finger can easily cause the beam to buckle. This feels like a soft spring.		When a sufficient voltage is applied between the wall electrode and the conductive side of the beam, the electrostatic pressure prevents the beam from buckling. This feels like a rigid surface or a very stiff spring.

Current Design

The current tactor prototype is shown demonstrating basic operation in soft and hard modes.

The beam in the prototype is 0.002 inch thick Mylar and the wall electrode is copper tape with a conductive adhesive backing. The length of the beam in the buckling region is roughly one inch. Polycarbonate was used for the surrounding structure to allow for viewing from all angles. This prototype is made for easy adjustment of design parameters and therefore bears little resemblance to the proposed final design.

Despite the non-optimality of the design, the current prototype shows promise in all four areas discussed above:

	Size: As discussed, the tactor consists only of a wall and a beam, and thus can be made with a very small footprint. While the current prototype is not designed for minimum size, the tactors can be made to be packed with a spacing that is comparable to their vertical travel.
	Weight: As with the size, the weight of the current prototype has not been minimized. The majority of the mass of the current design resides in the admittedly crude pin assembly on top of the beam.
	Efficiency: While this design can not maintain a hard setting without power, the power consumption of the design shown above has been measure on the order of tens of microvolts. The need to repeatedly charge and discharge each beam in order to display changing information during normal operation of the device, however, will increase this power demand.
	Work Space: As discussed above, the current design has a relatively small footprint. Additionally, each tactor is almost completely self-contained, requiring only an electrical connection to the control system to make it complete. Therefore, the total size of the device will only be the sum of the footprints of all the tactors plus minimal support structure. This promises to allow a much greater number of tactors than many previous designs.

Load Bearing Performance

The current prototype is capable of withstanding normal exploratory forces, given a high enough voltage is used. The following video shows the current prototype supporting a 110 gram aluminum cylinder. This is roughly equivalent to a 1 Newton exploratory force.

	This is a view of the latest prototype holding up a 110 gram aluminum cylinder while operating at 800 Volts.

The amount of force that the tacator can hold before buckling is a function of the voltage applied between the beam and the electrode on the wall. While an extremely high voltage would allow the beam to withstand vigorous exploration forces, high voltage is something to be avoided if possible.

We are exploring multiple methods to decrease the operating voltage including thinner dielectrics, exploiting friction and decreasing the tactor footprint. If the tactor footprint were made smaller, the users finger would be supported by multiple tactors, thus lowering the requirements for an individual tactor.

Comparison To Height Varying Displays

Given our relative ease in discriminating changes in surface height compared to our ability to discriminate changes in compliance, it was assumed at the outset that a haptic field display based on variable compliance would suffer some performance loss. Therefore, an initial experiment has been performed to compare the basic concepts of height-varying and compliance varying haptic field displays.

In the experiment, users were asked to determine which primitive shape was being displayed by a pin array. The pin array was able to operate as both a variable height and variable compliance device. During variable compliance, each pin was either rigid (hard) or soft and the device could provide one of three levels of compliance for the soft pins. The average time to shape recognition was recorded for each mode and general comparison was made. The results are compiled in the figure below.

In the figure to the left, the relative time required for users to recognize a primitive shape is displayed. The horizontal location corresponds to the compliance of the soft pins. When displaying shapes using variable heights, the lower pins can be thought of as having zero spring rate. Therefore, the results for this mode are shown at the location for zero spring rate. These initial results indicate that the performance of a variable compliance device is tied to the compliance of the soft mode. If the soft mode can be made sufficiently soft, the device can closely match the performance of a variable height device. The current design of the tactor has a soft mode that more than meets this requirement.

This graph shows the rough results developed during the initial experiments. The red bar corresponds to the height varying mode. The blue bars correspond to the variable compliance mode at different levels of compliance.

Future Work

The main focus of the current work is to maintain the ability of the tactor to hold off a resonable force while lowering the voltage required to do so. To that end, a thinner dielectric is being explored. Results of tests using a dielectric roughly 1/20 as thick as currently used will be available soon. It is hoped that this will allow the tactor to hold off reasonable exploratory forces using less than 100 Volts.

Approach 2: Compliant Bubbles

This approach to HFDs consists of micro-valves controlling the flow from small fluid-filled cavities. Research to date has focused on the feasibility of creating the fluid-filled cavity structure. The prototype structure was envisioned as fluid-filled packaging bubbles, arranged in a closely packed hexagonal pattern. A design study was conducted, revealing that the difficulties in creating this structure dealt with the following: creating the shaped film, joining the shaped film to a permanent backing and completely filling the structure with a fluid. The subsequent text describes the methods developed for the creation of the prototype structure.

Forming the Film

The shaped film was created in both synthetic latex and Low-Density-Polyethylene (LDPE). The synthetic latex shaped film was created by dip molding using an aluminum male mold. This method provided a consistent film approximately 5 mils thick. The drawback of this method was that the material filling the narrow areas between the positive features of the mold proved difficult to remove without tearing. The LDPE shaped film was created using pressure molding. The pressure molding device uses pressurized water, heated to ~205 F, to form the 3 mil thick LDPE film to a vented aluminum female mold.

Joining the Film

The dip molded synthetic latex film was joined with flat synthetic latex using ÒBlack MaxÓ by Loctite. This cyanoacrylate based adhesive forms a rigid bond along the contacting surfaces. This rigid bond provides limited reliability when used with an elastic film such as synthetic latex. The adhesive did sufficiently seal the small tears created when removing the synthetic latex from the mold. The formed LDPE was joined to the flat LDPE film by heating the samples while pressing them together. Experiments were conducted to determining that the LDPE films were joined together at 220 F under the applied pressure.

Filling the Bubbles

For simplicity, water was selected as the working fluid. The synthetic latex pieces must be joined absent of water. As a result, syringes were used to remove the existing air and fill with water. This method creates holes in the flat portion of the structure, requiring an additional piece of synthetic latex to be adhered to the bottom to seal these holes. The process of individually filling the volumes with water is time consuming and difficult, therefore additional work towards the development of a filling tool may be reasonable prior to additional production. The LDPE films were joined in the presence of water, eliminating the need of filing individually after joining. As a result of the high temperatures needed to join the LDPE pieces, the pressure applied must be sufficient to keep the water in liquid form