MUMBAI, India, Oct. 12 -- Intellectual Property India has published a patent application (202641118329 A) filed by Raghav Bansal Rncs; Rns Institute Of Technology, Bengaluru; Ms. Divya M Vassistant Professor; Mohammed Saqib Sk Rncs; Khaled Azim Katingeri Rncs; Dr. Vidya Yassociate Professor; and Ovais Nazir Rncs on October 03, 2026, for Aerodetect — Inmp441-Based Intelligent Acoustic Sensor Product For Real-Time Multi-Threat Audio Detection Using Tinyml On Esp32-S3 Microcontroller.
Inventors include Ms. Divya M Vassistant Professor; Dr. Vidya Yassociate Professor; Mohammed Saqib Sk Rncs; Ovais Nazir Rncs; Raghav Bansal Rncs; and Khaled Azim Katingeri Rncs.
The application for the patent was published on October 09, 2026, under issue no. 41/2026.
Abstract: The present invention discloses a product patent for an acoustic sensor system — the AeroDetect INMP441 Sensor Unit — comprising an INMP441 I2S digital MEMS microphone as the primary acoustic sensing element, interfaced with an ESP32-S3 microcontroller, forming a complete on-device acoustic threat detection product. The sensor unit captures environmental audio at 20 kHz sampling rate and processes it locally using Mel-Frequency Cepstral Coefficients (MFCC) feature extraction. A quantized Convolutional Neural Network (CNN) model is used for classifying multiple acoustic threat events including gunshots, footsteps, fire crackles, and human falls — all within a single on-device pipeline requiring no cloud infrastructure. The model is optimized using TinyML techniques (int8 quantization) to ensure low memory usage under 1 MB and fast inference latency under 200 milliseconds. The invention ensures real-time response, operates without internet connectivity, and preserves user privacy by avoiding transmission of raw audio data. The sensor product is suitable for deployment in rural healthcare, forest monitoring, smart homes, elderly care, and security surveillance applications.
Disclaimer: Curated by HT Syndication.
