Intelligent Position, Pressure and Depth Sensing in a Soft Optical Waveguide Skin

In this paper, we develop and validate a simple processing method suitable for position, pressure and depth sensing for a simple Soft Optical Waveguide Skin (SOWS). The soft skin consists of an elastomeric sensitive area (69 cm2and 2 mm thick) free of any other material pattern except for infrared emitters and photodetectors embedded at its periphery. When sensing area is touched, the photodetectors experience a loss in light intensity, and if a proper processing algorithm is developed, spatial information can be retrieved. Using adaptive boosting and decision trees as base learners, we develop classification and regression models to map this light intensity variation to our variables of interest: position, pressure and depth of indentation. This simple approach achieves high accuracy in predicting touch position with a spatial resolution of 9 mm, while simultaneously estimating with high precision the pressure level and depth of indentation.