Biomimetic acoustic perception via chip-scale dual-soliton microcombs

Abstract

Acoustic perception is a fairly basic but extraordinary feature in nature, relying on multidimensional signal processing for detection, localization, and recognition. Replicating this capability in compact artificial systems, however, remains a formidable challenge due to limitations in scalability, sensitivity, and integration. Here, imitating the auditory system of insects, we introduce an opto-acoustic perception paradigm using fully-stabilized dual-soliton microcombs. By integrating digitally stabilized on-chip dual-microcombs, silicon optoelectronics and bionic fiber-microphone arrays on a single platform, we achieve parallelized interrogation of over 100 sensors. Leveraging the low-noise, multi-channel coherence of fully-stabilized soliton microcombs, this synergy enables ultra-sensitive detection of 29.3 nPa/Hz1/2, sub centimeter precise localization, real-time tracking and identification for versatile acoustic targets. Bridging silicon photonics, optical fiber sensing and intelligent signal processing in a chiplet microsystem, our scheme delivers out-of-lab deployable capability on autonomous robotics. This work not only deepens the understanding of frequency comb science, but also establishes a concept of dual-comb-driven sensor networks as a scalable foundation for next-generation opto-acoustic intelligence.