This study aimed to explore non-visual drawing for blind and partially sighted individuals by limiting sighted bias through co-design with two blind and partially sighted drawers. The study involved participatory research through the stages of induction, iterative co-design, prototyping and evaluation.

The induction phase enabled me to develop an understanding of challenges facing blind and partially sighted drawers when they are using currently available tools. Video observations informed my inferences about the needs of blind drawers, revealing the importance of continuous perceptual feedback through bimanual manipulation of 3D objects to scaffold their construction of mental models.

The literature review revealed a need for affordable and intuitive drawing tools for non-visual drawers. It showed how most advances are in 2D (such as raised line) and 2.5D (bas-relief) technologies. Absent from the literature was work in 3D construction kits for 3D drawing, the possibility suggested through my video observations. Semi-structured interviews with both participants revealed their preferences for tactile feedback and their desires for capabilities to share drawings digitally.

The process of iterative co-design and prototyping with both participants informed my understanding of their needs for 3D objects, and the role these performed for their construction of 3D mental models. I observed how this was enabled through continuous bimanual haptic feedback.

Prototype 3 consisted of analog components of the toolkit that both participants evaluated. It consisted of 3D-printed custom geometric shapes with magnetic edges for assembling more complex objects from the geometric shapes, custom wax sticks and off-the-shelf components such as Play-Doh, Velcro tabs and loop scissors.

• Specific Insights from Evaluation

During the evaluation phase, I found that the 3D drawing toolkit could be used by P1 and P2 to create both simple and complex line-based and 3D drawings. Overall, I found that a drawing toolkit for blind and partially sighted individuals needs to provide opportunities for 3D spatial along with 2D raised line drawings. Future directions would consist of adding an additional variety of shapes and sizes in the toolkit and testing the prototype with a larger sample size. Specific findings follow:

• Strengths: Geometric tiles and wax sticks

P1 found geometric tiles easiest to work with when creating block-like structures and liked the flexibility of the wax sticks for more free form shapes such as those of hair represented with wax sticks in portrait drawing. P2 found wax sticks easiest to work with due to their familiarity with them, followed by Play-Doh and geometric tiles.

• Limitations: Not enough variety in geometric shapes

P1 and P2 both listed limitations when creating free form structures using pre-set geometric shapes. P2 suggested that further variety in shapes is needed to create complex structures. Both P1 and P2 suggested a need to add a cylindrical shape to the toolkit.

• Raised lines versus 3D

Preferences for line-based versus 3D components may depend on the type drawing being created (relative to its purpose) as well as flexibility and structure afforded by the materials. P1 and P2 both found line-based drawing limiting when creating certain structures (such as in the foreground) and leaned towards materials that provided the opportunity for volumetric spatial representation (e.g., P1 used Play-Doh to form nose, lips, and eyes for a 2.5D structure; P2 started making an airplane with wax sticks but changed their approach to make it with Play-Doh). P1 found raised lines to be effective for background scenes.

• Digital interface for social distancing and remote collaboration

The Covid-19 pandemic highlighted the importance of accessible virtual environments for education and work. Due to added barriers of virtual environments for blind and partially sighted individuals, it is critical to consider methods to create and share digital models of drawings and designs that work for BPSI. Although the scope of this project did not allow for creating digital models from all the components developed in

Prototype 3, a scanning application was developed as a starting point to create simple 3D structures with the use of geometric shapes, QR codes and augmented reality. By testing the feasibility of Prototype 4, I found that QR codes and AR may provide opportunities to create and share simple 3D models. However, this approach may be limited in sharing complex structures such as the ones created by P1 and P2. Advances in 3D scanning technologies have allowed for scanning 3D models. However, since these applications may be limiting for BPSI due to the significantly visual nature of their design i.e., continuous visual feedback needs to accurately scan an object, future directions may include developing an accessible 3D scanning application which can be used with the 3D drawing toolkit as well as other materials. Future directions would include building a prototype of an accessible scanning app. The app will allow users to create digital representations of 3D structures using a variety of materials beyond geometric shapes.