Virtual Keyboard Based on IMU

The development of sensors and artificial intelligence have promoted human-computer interaction. However, as the most widely used human-computer interaction, traditional typing keyboard, has not undergone fundamental changes in the past few decades. In this paper, a system based on the inertial measurement unit and long short term memory network algorithm are designed to recognize keyboard finger movements and achieve efficient text input. The experimental results show that the proposed scheme achieves over $\mathbf{9 9 \%}$ recognition accuracy for both traditional QWERTY layout keyboards and phone numeric keypads. The proposed virtual keyboard scheme not only saves space but also improves the input convenience for mobile device, which offers the valuable exploration in the smart typing and the interaction with virtual world.