A Machine Learning‐Aided Triboelectric Sensor for Boundary Indication in Volleyball

The present work was conducted to design a triboelectric nanogenerator (TENG)‐based sensor aided by machine learning for boundary indication for volleyball. For this purpose, a TENG‐based sensor was fabricated using polydimethylsiloxane (PDMS) and indium tin oxide (ITO) as the negative and positive active materials, respectively. Signals recorded by the sensor for three different mechanical actions (three classes) were processed to extract features that were used as input in a learning process. Four different algorithms, ada boost classifier (AdBC), decision tree classifier (DTC), random forest classifier (RFC), and extra trees classifier (XTC), were employed in the learning process to build classification models that could be used to distinguish between the three actions. The results revealed that the XTC‐based model shows the best performance with 92% accuracy and this was significant in the first class where an accuracy of 100% was obtained. In addition, it was found that cycle period (CP) is the most important feature in the built model.