T-Force: Exploring the Use of Typing Force for Three State Virtual Keyboards

Abstract

Three state virtual keyboards which differentiate contact events between released, touched, and pressed states have the potential to improve overall typing experience and reduce the gap between virtual keyboards and physical keyboards. Incorporating force sensitivity, three-state virtual keyboards can utilize a force threshold to better classify a contact event. However, our limited knowledge of how force plays a role during typing on virtual keyboards limits further progress. Through a series of studies we observe that using a uniform threshold is not an optimal approach. Furthermore, the force being applied while typing varies significantly across the keys and among participants. As such, we propose three different approaches to further improve the uniform threshold. We show that a carefully selected non-uniform threshold function could be sufficient in delineating typing events on a three-state keyboard. Finally, we conclude our work with lessons learned, suggestion for future improvements, and comparisons with current methods available.

In Summary

We explore the use of force as a metric to differentiate resting and typing behaviour on a pressure sensing flat surface keyboard (i.e., a pressures sensing touch screen). By first expanding our limited knowledge of how force plays a role within typing, we then calculate thresholds and propose and test force threshold functions which delineate typing and resting events. Through our work, we envision a broader use of force for flat surface typing.

Key Findings

• While resting forces are relatively consistent; the force exerted by people when typing is however not uniform across a virtual keyboard
• The difference in the maximum force exerted on the keyboard surface varies the most across columns, but also across rows
  • Importantly, different users have different levels of average maximum force exertions and even individual keystroke force profiles
• A uniform force threshold function is much preferred and performs better than a dynamic force threshold function
  • For future work, a solution to improve the threshold functions proposed in our work would be a combination of the three different approaches; where the maximum and minimum values used in the non-uniform threshold approach are scaled and each column follows a dynamic approach, both based on personalized data

Contributions

Our contributions are threefold:

• C1: We present three user studies exploring force characteristics when resting and typing on a virtual keyboard. Through detailed analysis of captured data, we show that a threshold function based only on force could be sufficient to define a three-state virtual keyboard. We further demonstrate that the force exerted by typing on a virtual keyboard is not uniform across the keyboard or among participants.
• C2: We propose three approaches for modifying the threshold function to account for such variation and conduct a fourth user study for validation.
• C3: We conclude with discussion of lessons learned through our approach, comparisons to current methods, and how future work can further improve on our approaches to further enhance the user experience. Overall, our results aim to provide insight into how future three-state virtual keyboards with force-sensitive surfaces can be designed.

In More Detail

Please review our full paper (linked above) for an abstract, study details, methodologies, and complete results.