Understanding the Impact of Machine Learning Models on the Performance of Different Flexible Strain Sensor Modalities

Fibre strain sensors commonly use three major modalities to transduce strain—piezoresistance, capacitance, and inductance. The electrical signal in response differs between these sensing technologies, having varying sensitivity, maximum measurable loading rate, susceptibility deleterious effects like hysteresis drift. wide variety of sensor materials tracking applications makes it difficult choose the best modality for a wearable device when considering quality, cost, difficulty manufacture. samples employing mechanisms are fabricated subjected using tensile tester. Their mechanical properties measured profiles designed exhibit particular shortcomings behaviour. Using data, compared identify technologies well suited different textile motion applications. Several regression models trained validated on random pattern providing guidance pairing each with model architecture that compensates non-ideal effects. A thermoplastic elastomer-core piezoresistive had highest sensitivity (average gauge factor: 12.6) similar construction polyether urethane-urea core largest bandwidth, capable resolving rates above 300% s−1 36% amplitude attenuation. However, both piezoresistve suffered from larger drift than coaxial polymer capacitive mechanism. Machine learning improved sensors’ root-mean-squared error by up 58% while maintaining their high ease interfacing electronically.