The Project Virtual Keyboard is a continuation of the Senseboard idea. The fingers position were measured by Electric Field Sensing and the combination with an IMU a learning system (based on ANN – Artificial Neural Networks) could learn your way of typing a text. After typing one page the system could read the typing. The accuracy was, however, only 90-95%.
After analyzing huge data volumes from the tests with the Senseboard-device and combining this with the research being carried out at Mälardalen University, the solution to the low accuracy is a completely new sensor that not only can increase the accuracy of the hand position – it can also measure the position of the finger tip. The old system had only one degree of freedom per finger. Now we have at least two (it can actually be increased to three of even four).
The basic idea is still valid – to measure both hand movement (using both an IMU and the new sensor) and a learning system (based on ANN – Artificial Neural Networks) could learn your way of typing on a keyboard. There will be a standard keyboard that you can use out-of-the-box. By using the learning mode you can actually make your own keyboard. You can easily define a complete new layout. As an example you can movethe ‘a’-key to your upper right corner of your personalized keyboard.