Wheeler: A Three-Wheeled Input Device for Usable, Efficient, and Versatile Non-Visual Interaction

TL;DR

Wheeler tackles the core accessibility challenge of navigating complex app UIs without vision by introducing a three-wheel, stationary input device that supports hierarchical (H-nav) and 2D (2d-nav) navigation. The approach combines independent wheel cursors mapped to UI hierarchy levels, tactile/auditory feedback, and a focused design process informed by blind users, yielding a 40% reduction in navigation time over keyboard+screen reader in a study with 12 participants. The contributions span hardware design (3D-printed chassis, encoders, Arduino HID firmware), two interaction modes (H-nav and 2d-nav with 2d-T-nav teleportation), and extensive evaluation including a diary study that documents real-world usability and serendipitous benefits like mixed-ability collaboration. The results demonstrate Wheeler’s potential to increase productivity for blind users and enable new modes of remote collaboration and interaction with partially accessible interfaces, while outlining pathways for open-source release and broader applications in VR and data visualization.

Abstract

Blind users rely on keyboards and assistive technologies like screen readers to interact with user interface (UI) elements. In modern applications with complex UI hierarchies, navigating to different UI elements poses a significant accessibility challenge. Users must listen to screen reader audio descriptions and press relevant keyboard keys one at a time. This paper introduces Wheeler, a novel three-wheeled, mouse-shaped stationary input device, to address this issue. Informed by participatory sessions, Wheeler enables blind users to navigate up to three hierarchical levels in an app independently using three wheels instead of navigating just one level at a time using a keyboard. The three wheels also offer versatility, allowing users to repurpose them for other tasks, such as 2D cursor manipulation. A study with 12 blind users indicates a significant reduction (40%) in navigation time compared to using a keyboard. Further, a diary study with our blind co-author highlights Wheeler's additional benefits, such as accessing UI elements with partial metadata and facilitating mixed-ability collaboration.

Figures (18)

• Figure 1: Demonstration of Wheeler's H-nav mode. (a) Multi-level menus in NetBeans; (b) nested structures in ribbons in MS Word; (c) a sample 3-level tree hierarchy to represent menus in either (a) or (b). For example, $1^{st}$-level elements, {a.1, a.2, ...} could represent the top-level menu items like {File, Edit, View, ... Profile} in (a) or {Home, Insert, ..., Help} in (b). Assuming a.1 equals Home, the $2^{nd}$-level menu items, {b.1, b.2, ...} will be {Clipboard, Font, ...} in (b). Similarly, assuming b.1 equals Clipboard, the $3^{rd}$-level menu items, {c.1, c.2, ...} will be {Paste, Cut, ...} in (b). In H-nav mode, Wheel-1 is always mapped to $1^{st}$-level menu items (i.e., a.*), Wheel-2 is mapped to $2^{nd}$-level menu items (i.e., b.*), and Wheel-3 is mapped to $3^{rd}$-level menu items (i.e., c.*).
• Figure 2: The red arrows highlight the path ( c.2 -> b.1 -> a.1 -> a.2 -> b.4 -> c.7 -> c.8) a blind user would have to take using a combination of keyboard and screen reader when going from c.2 to c.8 in the hierarchy from Fig. \ref{['fig:hnav_illustration']}c. There are six steps in total, and at least six keystrokes would be required.
• Figure 3: Navigating from c.2 to c.8 in the hierarchy from Fig. \ref{['fig:hnav_illustration']}c using Wheeler's H-nav mode. (a)-(d) shows the positions of the three Wheeler cursors mapped to its three wheels at different stages of the navigation. Only three rotations are required in total.
• Figure 4: Illustrations of Wheeler's 2d-nav mode. (Top) A blind user moves the cursor from the lower-left corner to the upper-right corner of a 2D screen with 12 UIs organized in a $3\times4$ grid. (Bottom) shows the sequence of wheel operations the user performs at different steps to achieve this goal. Each step (i) is marked in both the top and the bottom figures. The green-colored wheel indicates which wheel the user rotated at step (i). The user rotates Wheel-1 or Wheel-2 to move the cursor horizontally or vertically and adjusts the speed of the cursor by rotating Wheel-3. Note that the Euclidean distance between the source and the destination is $d$, which sighted users usually take using a mouse. In contrast, blind users take the Manhattan distance between two nodes (e.g., $d_1$ along X, plus $d_2$ along Y) in 2d-nav mode.
• Figure 5: Several low-fidelity prototypes during ideation.