Development of a Capacitive Measurement Setup for a Non-Contact Control System

A mobile demonstrator was developed for contactless control of a technical system using a capacitive measurement system. The system employs artificial intelligence to interpret data from a $4 \times 4$ sensor matrix, calculate the position of the user’s fingertip, and generate control signals. Key components include a computing unit (Jetson Nano), a capacitive measurement system (PSoC 4100S CAPSENSE Prototyping Kit), and a custom-designed sensor board with a $\mathbf{4} \mathbf{x} \mathbf{4}$ sensor matrix featuring circular conductor loops (capacitive sensors). The sensor data is processed by two machine learning (ML) models to first detect a hand and second to calculate the fingertip position; both ML-models are trained on customprepared datasets. To create these training datasets, an additional ML-model was utilized for optical position detection of the fingertip, enabling training of the primary models. This approach highlights the potential of combining capacitive sensing with AI to achieve gesture-based control.