A Cantilever-Based Broadband Acoustoelectric Device with Aligned PVDF-HFP Nanofibers for Self-Powered Sensing and High-Accuracy Speech Recognition

As the adoption of acoustic sensing grows in speech recognition, health monitoring, and the Internet of Things, the limited broadband response of flexible acoustic sensors has become a significant challenge. This study presents a novel acoustoelectric device featuring a cantilever-based electrode structure. By designing an electrospun aligned PVDF-HFP nanofiber piezoelectric active layer alongside a gold-coated film electrode with cantilever architecture, the device achieves a triple coupling mechanism of acoustic, mechanical, and electrical interactions, leading to enhanced performance. Under 110 dB sound pressure, the 12 × 8 mm 2 cantilever-area device demonstrates an extended operating bandwidth of up to 800 Hz, a peak voltage of 33.55 V, and high sensitivity within the 60–110 dB sound pressure range (74.25 mV·Pa –1 at 1000 Hz, 94 dB SPL and 8.13 V·Pa –1 at 170 Hz, 100 dB SPL). It attains a voice capture signal-to-noise ratio of 61.50 dB and a machine learning-based speech recognition accuracy of 98.5%. This work establishes a new paradigm for broadband acoustic device design and advances the practical application of high-fidelity human–computer interaction and intelligent sensing technologies.