Paper prototyping is a well-established technique in traditional user interface design as it supports design teams in early development phases to communicate and discuss their ideas. Unfortunately, there is no equivalent technique for the development of augmented reality (AR) applications. We developed PapAR, a prototyping technique with the same simplicity as using pen and paper but taking into account important specifics of AR applications. By drawing on two layers instead of one, with PapAR it is possible to represent the dynamic behavior of different content stabilization techniques as well as the interplay of virtual content and the underlying real world (Fig. 1). We used PapAR in the early design phase of a user interface, which is specifically designed to be displayed in a head-mounted display (HMD) in the car. In this paper we first present the PapAR technique itself and then discuss our first experiences of using it.

Head-mounted displays (HMDs) have great potential to improve the current situation of car drivers. They provide every benefit of a head-up display (HUD), while at the same time showing more flexibility in usage. We built an infotainment system specifically designed to be displayed in an HMD. With this system, we then conducted a dual task study in a driving simulation, comparing different techniques of content stabilization (head-and cockpit stabilized visualizations). Interaction with the system took place via a physical input device (rotary controller) or indirect pointing gestures. While cockpit-stabilized content generally resulted in a slightly better driving performance, HMD visualizations suffered from technological limitations, partly reflected in the secondary task performance and subjective feedback. Regarding input modality, we found that horizontal gesture interaction significantly influenced the quality of lane keeping. Apparently, horizontal interaction with the one hand caused unintentional steering with the other.