Investigating Precise Control in Spatial Interactions: Proxemics, Kinesthetics, and Analytics

Abstract

Augmented and Virtual Reality (AR/VR) technologies have reshaped the way in which we perceive the virtual world. In fact, recent technological advancements provide experiences that make the physical and virtual worlds almost indistinguishable. However, the physical world affords subtle sensorimotor cues which we subconsciously utilize to perform simple and complex tasks in our daily lives. The lack of this affordance in existing AR/VR systems makes it difficult for their mainstream adoption over conventional $2D$ user interfaces. As a case in point, existing spatial user interfaces (SUI) lack the intuition to perform tasks in a manner that is perceptually familiar to the physical world. The broader goal of this dissertation lies in facilitating an intuitive spatial manipulation experience, specifically for motor control.

We begin by investigating the role of proximity to an action on precise motor control in spatial tasks. We do so by introducing a new SUI called the Clock-Maker's Work-Space (CMWS), with the goal of enabling precise actions close to the body, akin to the physical world. On evaluating our setup in comparison to conventional mixed-reality interfaces, we find CMWS to afford precise actions for bi-manual spatial tasks. We further compare our SUI with a physical manipulation task and observe similarities in user behavior across both tasks.

We subsequently narrow our focus on studying precise spatial rotation. We utilize haptics, specifically force-feedback (kinesthetics) for augmenting fine motor control in spatial rotational task. By designing three kinesthetic rotation metaphors, we evaluate precise rotational control with and without haptic feedback for 3D shape manipulation. Our results show that haptics-based rotation algorithms allow for precise motor control in 3D space, also, help reduce hand fatigue.

In order to understand precise control in its truest form, we investigate orthopedic surgery training from the point of analyzing bone-drilling tasks. We designed a hybrid physical-virtual simulator for bone-drilling training and collected physical data for analyzing precise drilling action. We also developed a Laplacian based performance metric to help expert surgeons evaluate the resident training progress across successive years of orthopedic residency.