A Signal Amplitude-Insensitive Triboelectric Touch Panel with a Significantly Reduced Signal Channel and Deep-Learning-Enhanced Robustness

The self-powered triboelectric touch panel has garnered considerable research attention due to its potential to reduce system energy consumption and its applications in human-machine interfaces, e-skin, and the Internet of Things. Current methods for achieving triboelectric-based touch positioning in an M x N detection pixel array typically require signal amplitude comparison across at least M + N signal channels, thereby limiting lightweight design possibilities. In contrast, our novel "resistor ladder" approach necessitates only 4 signal channels for touch positioning. This method leverages a lookup table correlating touch positions with amplitude ratios from different channels, rendering it insensitive to signal amplitude and significantly enhancing robustness. We fabricated a transparent touch panel using PET tribomaterial, where the surface roughness was enhanced through plasma treatment. The panel successfully demonstrated touch positioning for 128 taps within a 4 x 4 pixel detection array and sliding positioning using a predefined lookup table. To further enhance device robustness, a 2D convolutional neural network was implemented, which achieved an impressive touch positioning accuracy of 97.7% even under artificially introduced signal defects. This study represents an initial exploration of amplitude-insensitive touch and sliding positioning methods, significantly reducing the number of required signal channels and enhancing the robustness of triboelectric touch panels.