Computational Design and Fabrication of Wearable Haptics

Abstract

Wearable technologies are increasingly embedded in our lives, from smartwatches on our wrists and augmented reality displays on our heads to more advanced sensing garments that can estimate our pose in real-time. While wearable sensors aim to collect and analyze information from our bodies, wearable haptics aim to physically provide information back to our bodies. The two technologies are potentially complementary, for example, allowing users to physically interact with objects in Virtual Reality (VR). However, while wearable sensing has advanced significantly, wearable haptics remains a challenging problem. The main challenge in designing wearable haptics are the opposing objectives of performance and wearability. Higher performance devices are likely to be bulkier and heavier, thereby reducing comfort. In addition, haptic devices need to be in close proximity to the body, and unlike wearable sensors, must account for the physical interaction with the soft outer layers of the body. Because of these considerations, wearable haptics has largely been confined to basic vibrotactile actuation on the body (e.g. buzzing), which falls far short of fully harnessing our proprioceptive and tactile abilities. We address this challenge in two parts. First, we develop actuators that are designed from the ground-up for compliance - the property of flexible and semi-stretchable soft components that makes them ideal for on-body use. Counter intuitively, the property of compliance also makes actuators incredibly inefficient (as they are poor at force transmission) if not properly designed, placed, and connected to the body. Thus our second idea is to design such wearable haptic systems via computational methods. Specifically, we use physical modeling and simulation in combination with topology optimization to create inverse design methods. This allows for the automatic creation of a variety of garments designed to resist specific motions (kinesthestic garments). We further extend our framework to leverage stateful components (haptic devices that can become stiff on command), and optimize for multiple objectives depending on these states. The resulting wearable haptics are a novel type of soft exoskeleton which are comfortable to wear, and yet can resist a variety of motions on-demand. Going forward, our results based on force characterization and in various VR tasks, show that automatically designed haptic systems have the potential to perform significantly better than manually designed counterparts.