Two Examples of Using Machine Learning for Processing Sensor Data of Capacitive Proximity Sensors

Abstract

Recently we have shown developments on capacitive tactile proximity sensors (CTPS) in combination with machine learning techniques to extract further information out of the sensor signals. In this work we summarize two examples of the applications we have presented. In the first approach we have investigated distance classification based on the proximity information of the sensors. In the second approach the possibility of material recognition was investigated. The latter is done by variating the spatial resolution and the exciter frequency of our sensors. For both approaches, distance classification and material recognition, an artificial neural network was set up and fed with various data sets of different electrode combination. The influence of the electrode combinations and shapes on the recognition accuracy was investigated and some promising results could be achieved.