Implementations of the technology described herein provide a method for detecting gesture commands using an ultrasonic pen system. The system has a pen and a user device. Detection of gesture commands is based on two-dimensional gestures relative to the screen of a user device, three-dimensional gestures relative to the screen of the user device, roll/rotation around a longitudinal axis of the pen body, and micro-twisting around the longitudinal axis of the pen body. The user device receives the gestures and translates them into commands such as UNDO and BACK.