MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202641085358 A) filed by Sri Eshwar College Of Engineering on July 11, 2026, for A Wearable Triple-Gate Sensor Fusion System And Method For Real-Time Open-Water Drowning Detection Using Sequential Validation.
Inventors include Samraj S.; P. Ganeshan; and M. Tamil Selvan.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: FIGURE OF ABSTRACT A WEARABLE TRIPLE-GATE SENSOR FUSION SYSTEM AND METHOD FOR REAL-TIME OPEN-WATER DROWNING DETECTION USING SEQUENTIAL VALIDATION The present invention relates to a wearable drowning detection apparatus configured for open-water environments, comprising a motion sensing module with an Inertial Measurement Unit sampling at 100Hz, a depth sensing module with a barometric pressure sensor, a physiological sensing module with a photoplethysmography sensor for measuring heart rate and blood oxygen saturation, and an embedded controller executing a sequential triple-gate validation process. The apparatus incorporates an on-device machine learning classifier utilizing a Random Forest algorithm with 50 decision trees to distinguish drowning motion signatures from normal swimming using features comprising acceleration variance, gyroscope angular rate energy, body tilt angle, zero-crossing rate, and tilt switch state extracted from 3-second sliding temporal windows of IMU data, wherein the classifier outputs a Swimming (Safe) class which is explicitly suppressed to prevent false alarms, and Pre-Drowning Distress and Drowning classes which trigger progression through the validation pipeline. The apparatus further comprises a personalized baseline calibration engine configured to record physiological and motion parameters during a 30-to-60- minute baseline period, compute statistical descriptors comprising mean and standard deviation, and store said descriptors as a personalized baseline profile in non-volatile memory, enabling individualized threshold adaptation rather than population-wide fixed thresholds. The sequential triple-gate validation process operates as follows: Gate 1 comprises motion anomaly detection requiring acceleration variance exceeding the personalized baseline and vertical body orientation detected by redundant tilt switches for more than 5 seconds; Gate 2 comprises submersion validation requiring hydrostatic pressure corresponding to depth exceeding 0.3 meters and sustained immersion for more than 8 continuous seconds, wherein the 8-second threshold provides a rescue window while avoiding false alarms from brief submersions; and Gate 3 comprises physiological distress validation requiring heart rate below 40 beats per minute and blood oxygen saturation below 90%. Upon successful completion of all three validation gates, the apparatus activates audible, visual, and haptic alerts while the GSM/GPS module acquires GPS coordinates and transmits SOS SMS messages containing location data to predefined emergency contacts every 2 minutes until acknowledgement is received. The apparatus operates independently of internet connectivity and external infrastructure, with all sensor data acquisition and machine learning inference performed locally on the nRF52840 microcontroller without transmitting raw physiological data to external servers, and is configured as a wrist-worn device with an IP68 waterproof enclosure and a rechargeable battery providing at least 72 hours of standby operation, targeting the specific gap of open-water drowning prevention in rivers, lakes, reservoirs, and floodwater where conventional camera-based systems and lifeguard supervision are unavailable.
Disclaimer: Curated by HT Syndication.
