Self-powered temperature pressure sensing arrays with stepped microcone structure and Bi2Te3-based films for deep learning-assisted object recognition

Flexible temperature-pressure bimodal sensing arrays can detect multiple types of information, including force and heat, making them crucial for applications such as object classification, human-machine interaction, and artificial intelligence. However, current sensors primarily focus on single-parameter and single-point measurements, while lacking a continuous and stable power supply. This study developed flexible, self-powered temperature-pressure sensing arrays by integrating a stepped microcone structure with thermoelectric materials. This stepped distribution microstructure design enabled effective pressure measurements across a wide range, with high sensitivity and fast response. Temperature-independent measurements were achieved synchronously over a wide temperature range (35-173 °C) by incorporating high-performance Bi₂Te₃-based thermoelectric films. These temperature and pressure sensing units can discern temperature and pressure stimuli without mutual interference. Furthermore, with the assistance of deep learning, these bimodal sensing arrays performed spatial mapping of temperature and pressure simultaneously, demonstrating their ability to identify different types of objects with an accuracy exceeding 98%. Therefore, this study shows promise for advancing human-machine interaction, artificial intelligence, and self-powered electronic skins.