Integrated Position and Force Sensing for Optical Artificial Skin using Machine Learning Methods
| Autor: | Jeremy Scerri, Kris Scicluna, Clive Seguna, Tiziana Borg, Shelley Cardona Mills |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Computer science
Acoustics 010401 analytical chemistry 02 engineering and technology 021001 nanoscience & nanotechnology 01 natural sciences Artificial skin 0104 chemical sciences law.invention Microcontroller law Position (vector) Line (geometry) Slab 0210 nano-technology Robotic arm Light-emitting diode |
| Zdroj: | IECON |
| DOI: | 10.1109/iecon.2019.8927765 |
| Popis: | In this paper a flexible and transparent gel material is employed as an optical artificial skin. The setup involves providing white light from one side of the rectangular gel slab and measuring the amount of light received on the opposite side with phototransistors. It is shown that using two (opposite) sides only is enough to get the x, y position of the pressure points to within 0.57 cm of accuracy and also the force level to within 1.37 N of accuracy. Using only two sides is beneficial for applications such as robotic arms and prostheses as once the gel is wrapped around them, the light source and the sensors would end up placed next to each other, in a line. This maximizes the functional area and facilitates installation. The reflective, dispersive and refractive properties of the layout was captured with adaptive neuro-fuzzy inference system techniques. The resulting model consists of four fuzzy inference systems which, once trained, could be easily embedded in a low-cost microcontroller. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|