K-mer-Based Recognition of Multiple Load Conditions and Force Measurement Using Graphene/PVDF Helical Sensor

Advancements in sensor technology have catalyzed the development of flexible sensors capable of detecting diverse stimuli, such as external loads, thermal variations, and human motion. This study developed an innovative helical-shaped piezoelectric sensor incorporating a precisely aligned polyvinylidene fluoride (PVDF)/graphene composite film as the active piezoelectric component. This advanced sensor generates voltage time-series data with distinct waveforms in response to various forces, enabling multidirectional sensing. Through standardized experiments involving external forces (tensile, radial compression, axial compression, bending, and twisting), voltage signal variations were measured under different types and magnitudes of these external forces, and classified based on K-mer frequency. A functional relationship between force magnitude and voltage signal was established to facilitate subsequent force measurements. Using the acquired signals, K-mer-based machine-learning algorithms, including random forest (RF), support vector machine (SVM), and K-nearest neighbors (KNNs), with specialized signal sampling and preprocessing methods were applied to analyze voltage data over specific durations. Ten-fold cross-validation resulted in a classification model with an accuracy rate of 92.82% (±1.07%). Finally, the graphene/PVDF piezoelectric helical sensor was integrated into a wearable device designed with thermoplastic polyurethane (TPU). When worn on the arm, wrist, and finger, this wearable device generated different forces through various postures, resulting in voltage signals that matched standardized patterns. The K-mer-based RF classifier, optimized with 19 200 subsequences, 0.3 k overlap, and smoothing every 100 records, achieved an 83.97% (±0.85%) recognition rate on the hand action dataset. Additionally, force measurements were derived from the functional relationship established between voltage signals and force.