Identification of the sensory properties of image-based multi-axis force/torque sensors

A novel image-based sensor concept for the detection of multi-axial forces/torques has been introduced by the authors in previous publications. The sensor concept is capable of detecting multiaxial loads in a single image and uniquely identifying all load components. In the present work, the sensory properties of the image-based multi-axis force-torque sensor are investigated in more detail. On the one hand, it compares the measurement resolution and measurement ranges as essential properties of a force/torque sensor with dominate strain gauge-based ones. Moreover, the measuring dynamics of the sensor are addressed and Deep Learning approaches are investigated to enhance the computationally intensive analysis process of the image data of such sensors.