SBIR Phase II: AI-Enabled Acoustic Signal Interpretation for Automated, Standardized Quality Control Testing

The broader/commercial impact of this Small Business Innovation Research (SBIR) Phase II project is the advancement of an artificial intelligence–enabled sensing platform that uses audio signals (known as an acoustic fingerprint) to monitor conditions in situations where physical access or visual inspection is difficult or impossible. This technology addresses a longstanding limitation of conventional sensing and signal-processing methods, which fail under extreme noise conditions common in manufacturing, infrastructure, and field operations. By enabling reliable, real-time interpretation of acoustic signals, the innovation creates new capabilities for quality control, system monitoring, and autonomous decision-making across multiple sectors, including advanced manufacturing, robotics and autonomous systems, transportation, and assistive technologies. Potential applications include early fault detection in electric motors and rotating machinery, structural monitoring of energy systems, enhanced situational awareness for robotics, and improved acoustic performance in various devices. The work enhances scientific and technological understanding by integrating artificial intelligence with acoustics to unlock a new class of intelligent sensing capabilities. This Small Business Innovation Research (SBIR) Phase II project advances the state of the art in intelligent sensing by developing a physics-informed, artificial intelligence-enabled acoustic fingerprinting approach for real-time, non-contact quality inspection in harsh industrial settings. The project addresses a fundamental technical challenge: extracting meaningful acoustic information from complex, high-noise settings where traditional sensing and signal-processing techniques are ineffective. The research integrates physics-based acoustic modeling with machine learning methods to preserve signal integrity and isolate defect-related acoustic features prior to data-driven analysis. The Phase II research objectives include designing factory-ready acoustic sensing architectures, developing physics-guided feature extraction techniques, and implementing hybrid artificial intelligence models capable of identifying both 39 known and previously unseen defects. The proposed work emphasizes real-time performance, and scalability, ensuring compatibility with high-speed production situations without disrupting existing workflows. Anticipated outcomes include validated sensing architectures, improved defect detection accuracy under extreme noise conditions, and new scientific insights into combining physical acoustics with artificial intelligence. Together, these results represent a substantial technical advancement in non-contact diagnostics and intelligent quality inspection. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.