Improving the robustness to input errors on touch-based self-service kiosks and transportation apps

Abstract

Text input is cumbersome on public self-service kiosks and apps. Applications often provide prefix-based text suggestions to speed up input, but the input must be correct. However, application specific lists are usually limited in size. For example, the Norwegian rail network comprise 324 stations. This can be exploited in the input process. Several approaches are explored herein for more flexible text input that are robust to errors.