Impact Localization in Inkjet-Printed Tactile Grid Sensor with Echo State Network

Tactile sensors with impact localization are becoming an essential part in automotive, aerospace, and civil engineering for damage assessment, safety assurance and structural monitoring. Inkjet printing is on rise for its eco-friendliness, cost-efficiency, low power consumption and quick design iteration ability. However, its minimal fabrication process results in operational challenges. These challenges can be mitigated by integrating artificial intelligence with inkjet printed sensors to enhance their performance. Among all artificial intelligences, echo state networks are gaining recognition for their low computational demands and hardware compatibility. This study developed an inkjet-printed tactile grid sensor with an echo state network for impact localization. The sensitivity of sensor was assessed through a pencil drop experiment, with data transformation across time and magnitude domains to improve network adaptability. Hy-perparameters of the model were fine-tuned through sequential search. Developed echo state network with grid tactile sensor demonstrated high accuracy, pinpointing the impact location of pencil drops with an impressive precision rate of 94.89%.