Paper Reading #14: TeslaTouch: Electrovibration for Touch Surfaces

Reference Information
TeslaTouch: Electrovibration for Touch Surfaces
Olivier Bau, Ivan Poupyrev, Ali Israr, Chris Harrison
Presented at UIST'10, October 3-6, 2010, New York, New York, USA

Author Bios

• Olivier Bau is a Post-Doctoral Research Scientist in Disney Research in the Interaction Design group. His PhD is in Computer Science and was received from INRIA Saclay.

• Ivan Poupyrev is a Senior Research Scientist in Disney Research in the Interaction Design group and studies interactive technologies and interface design.

• Ali Israr has a PhD in Mechanical Engineering from Purdue and now works in the Interaction Design @ Disney group in Disney Research. His interests are in disseminating haptic technologies in consumer-grade products.

• Chris Harrison is a PhD student at Carnegie Mellon's Human Computer Interaction Institute. He is also a Microsoft Research PhD Fellow and focuses his research on novel interaction techniques.

Summary
Hypothesis
How effective is electrovibration produced by contact with a touch screen as dynamic tactile feedback?

Methods
All tests grounded users.

The first test determined how users perceived the produced sensations. Four frequency and amplitude combinations were used and represented the limits of the range of each. The participants filled out questionnaires that asked them to describe the sensation, the texture of the sensation, and the dimensions of the sensation on a Likert scale.

The second test was used to determine the threshold of voltage for the detection of a sensation. The thresholds were estimated for five frequencies and narrowed down through a series of minor changes both up and down in voltage. After 12 reversals from increasing to decreasing voltage or vice versa, the test was concluded. The amount of change decreased after 3 reversals. Frequency and amplitude thresholds were determined similarly. The absolute detection threshold test presented participants with two screens, only one of which had a tactile screen, and challenged participants to pick the one with the electrode. The absolute discrimination test had another screen and allowed participants as much time as they needed.

Results
Grounding increased the intensity of the tactile sensation. For the first test, low frequencies felt rougher than higher frequencies. The effect of amplitude depended on the frequency. High frequency and increased amplitude increased the smoothness, while low frequency and decreased amplitude increased the stickiness. The participants tended to describe the sensations as combinations of vibration and friction.

The second test found detection threshold levels similar to force detection threshold levels found by another group, suggesting that the two are related. These results were placed onto a chart for use by future designers that show the relationship between frequency and voltage. The relationship between these is not linear. The frequency just-noticeable-differences (JND), which indicate when a user perceived a difference between two surfaces, resembled those produced by mechanical vibration. The amplitude JND was slightly lower than previously found results, but was constant.

Contents
A lack of haptic feedback inhibits user interaction, but most of the research in that field focuses on using electromechanical actuators through either the screen or a pen used to interface with the surface. Electrovibration uses electrostatic friction between the screen and a user's fingers. It is quick, dynamic, uses little power, is scalable, and can be added to current devices fairly easily. Electrovibration is based on Mallinckrodt's 1954 discovery and uses alternating voltage on the surface to produce an intermittent attraction force that modulates friction and creates the feeling of a rubbery surface. Unlike other tactile actuation systems, it does not send charges into the skin as the sensation is produced by a mechanical action and directly actuates the fingers. Electrovibration produces a flat response, is more subtle than traditional mechanical vibration, is entirely noiseless, and is reliable. TeslaTouch delivers its interaction through a clear electrode on the surface of the device.

In this paper, the system used a diffuse illumination multitouch tabletop device. The transparent electrode is excited with a periodic electrical signal that spans the whole surface and is spatially uniform. With sufficient amplitude, the attractive force increases the dynamic friction. Only fingers that are moving feel the sensation since it is through friction. The current is limited to 0.5mA, which is safe. The tabletop surface was built with several panels, tracked interaction through infrared cameras, and ran in real time. The developed technology was deliberately general and can be used in a variety of ways not tested in the paper. Specific sensations can be engineered for a variety of situations, including painting programs, non-visual information layers, conventional GUI widgets, direct manipulation, and rubbing interactions. Multitouch interactions are supported so long as only one finger is moving at a time.

Discussion
The authors created a system that allows for tactile sensations to be perceived through electrovibration. Their extensive user testing convinced by that their results are well-founded. More importantly, they also produced a chart that can be tested later for accuracy.

I could see this being used in a variety of touch devices, as it is less power-hungry than mechanical vibration. The fact that it is scalable makes me think that this is a viable technology on tabletop devices or on ubiquitous computing devices. This is not as useful with the shoe-based UI (though slightly less than impossible), but any surface that requires finger-based interaction might find this a simple and cost-effective way to incorporate tactile feedback.

I was concerned that the entire surface receives the same sensation of friction modulation. If I slightly move the wrong hand with multitouch input, I would not expect to feel the texture that my other hand is feeling. All the possible multitouch ideas presented here rely on the fact that one of the hands is not moving, which dramatically reduces the possibilities for the technology.