MUMBAI, India, Feb. 6 -- Intellectual Property India has published a patent application (202521124630 A) filed by Madake Jyoti; Jungade Tanvee Umesh; Lakhe Prathamesh Shankar; Lambat Akhilesh Vijayrao; Bhatlawande Shripad; and Shilaskar Swati, Pune, Maharashtra, on Dec. 10, 2025, for 'elderly person activity detection false fall avoidance.'
Inventor(s) include Madake Jyoti; Jungade Tanvee Umesh; Lakhe Prathamesh Shankar; Lambat Akhilesh Vijayrao; Bhatlawande Shripad; and Shilaskar Swati.
The application for the patent was published on Feb. 6, under issue no. 06/2026.
According to the abstract released by the Intellectual Property India: "The present invention discloses an AI-enabled wearable system for real-time fall detection, activity monitoring, and preventive health support for elderly individuals. The invention integrates dual MPU6050 motion sensors with machine-learning models to classify daily activities such as walking, sitting, sleeping, and falling, thereby enabling accurate detection of abnormal motion events and timely safety intervention. The system operates through a multi-stage architecture comprising sensor data capture, feature extraction, AI-based activity classification, fall-event recognition, and emergency alerting. The data capture unit continually records synchronized accelerometer and gyroscope signals from sensor modules placed strategically on the user's body. This ensures seamless and continuous acquisition of motion information without requiring cameras or microphones, thereby preserving privacy and enhancing portability. The feature extraction module processes each sensor stream independently. Accelerometer signals are analyzed for linear acceleration, posture transitions, and impact strength, while gyroscope signals capture angular velocity, rotational shifts, and sudden orientation changes. These processed features are then used as inputs to a Feedforward Neural Network (FNN) for multi-class activity recognition and a Random Forest classifier for distinguishing true falls from rapid but non-harmful movements. Each model generates classification outputs that collectively determine the user's current activity state. In the event of a detected fall, the system activates an automated emergency response mechanism that sends alerts via SMS or mobile application to preconfigured caregivers, optionally including GPS location data. This ensures rapid assistance even in situations where the user is unable to respond or where internet connectivity is unavailable. The integrated mobile interface provides real-time visualization of activity logs, posture trends, and fall histories, enabling caregivers to track user behavior and detect risk patterns. The system also incorporates preventive-health functionality, generating posture-correction prompts and personalized exercise recommendations based on clinically supported fall-prevention guidelines. A key aspect of the invention is its privacy-preserving and offline-capable design. Unlike camera-based monitoring systems that require continuous visual recording, the proposed wearable processes all motion data locally on the device and stores only derived numerical features, ensuring user confidentiality and operational reliability in diverse environments. In summary, the proposed Al-enabled Fall Detection and Prevention System provides an advanced, automated, and user-centric solution for elderly safety. It reduces caregiver burden, enhances emergency responsiveness, and offers proactive health recommendations, establishing a strong foundation for intelligent, data-driven elderly care and long-term fall-risk reduction."
Disclaimer: Curated by HT Syndication.
