應用類神經網路建立一個具有自然使用者介面之觸覺回饋裝置

Abstract

Virtual Reality (VR) and Augmented Reality (AR) technology are increasingly used in complex engineering applications; however, VR and AR systems provide limited haptic feedback in these applications. In order to solve this problem, a method for providing realistic haptic feedback was designed in this research. A recording device combined accelerometer and force sensor was designed to record the real texture, and a simulation device combined accelerometer, force sensor, and piezo actuator was designed to simulate the actuator haptic feedback. The recorded data from the recording device was processed to be the simulation data using a neural network. The neural network was used to train an approximate function. This function was programed into the simulation device, so that the simulation device can create corresponding tactile feedback which was recorded from the recording device. With our method, we can provide realistic haptic feedback and perform real-time feedback in natural operation.