Design, characterization, and testing of skin-stretch feedback integrated into a game controller.

摘要

Haptic feedback in modern game controllers is limited to vibrotactile feedback. The addition of skin-stretch feedback would significantly improve the type and quality of haptic feedback provided by game controllers. Skin-stretch feedback requires small forces (around a few newtons) and translations (as small as 0.5 mm) to provide identifiable direction cues. Prior work has developed skin-stretch mechanisms in two form factors: a flat form factor and a tall but compact (cubic) form factor. These mechanisms have been shown to be effective actuators for skin-stretch feedback, and are small enough to fit inside of a game controller. Additional prior work has shown that the cubic skin-stretch mechanism can be integrated into a thumb joystick for use with game controllers. This thesis presents the design, characterization, and testing of two skin-stretch game controllers. The first game controller provides skin stretch via a 2-axis mechanism integrated into its thumb joysticks. This controller uses the cubic skin-stretch mechanism to drive the skin stretch. Concerns that users' motions of the joystick could negatively impact the saliency of skin stretch rendered from the joystick prompted the design of a controller that provides 2-axis skin stretch to users' middle fingers on the back side of the controller. Two experiments were conducted with the two controllers. One experiment had participants identify the direction of skin stretch from a selection of 8 possible directions. This test compared users' accuracies with both controllers, and with five different finger restraints on the back-tactor controller. Results show that users' identification accuracy was similar across feedback conditions. A second experiment used skin stretch to rotationally guide participants to a randomized target angle. Three different feedback strategies were tested. Results showed that a strategy called sinusoidal feedback, which provided feedback that varied in frequency and amplitude as a function of the user's relative position to the tactor, performed significantly better on all performance metrics than the other feedback strategies. It is important to note that the sinusoidal feedback only requires two 1-axis skin-stretch actuators, which are spatially separated, in order to provide feedback. The other lower performing feedback strategies used two 2-axis skin-stretch actuators.