MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202641110988 A) filed by Pragati Engineering College on September 16, 2026, for Human-Posture Ai Guard For Agricultural Workers.
Inventors include Dr. A. Radha Krishna; Dr. Saroja Rani; Mr. G. Vijaya Kumar; and Mr. Janardhana Rao Addanki.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: HUMAN-POSTURE AI GUARD FOR AGRICULTURAL WORKERS Agricultural workers are exposed to physically demanding tasks involving repetitive bending, lifting, twisting, stooping, and squatting, leading to a high prevalence of musculoskeletal disorders (MSDs). Studies have documented that approximately 58% of agricultural workers suffer from lower back pain, significantly higher than the general population. Existing ergonomic wearables are predominantly designed for office workers and are trained on sedentary posture datasets, rendering them inaccurate for agricultural movement patterns. Furthermore, these devices rely on cloud connectivity, are expensive, and lack durability for harsh farm environments. The present invention provides a clip-on wearable device that addresses these limitations through a dedicated agricultural posture monitoring solution. The device comprises an MPU6050 inertial measurement unit (IMU) sensor with three-axis accelerometer and gyroscope for capturing body orientation and movement data; an nRF52840 microcontroller with onboard flash memory and integrated Bluetooth Low Energy; a TinyML posture classifier stored in said memory and trained on a custom labeled dataset of Indian agricultural worker postures; a haptic motor for vibration feedback; and a rechargeable battery power source. The TinyML classifier is trained on a comprehensive dataset including bending, lifting, twisting, stooping, squatting, side bending, and neutral postures collected from agricultural workers performing actual farm tasks. The microcontroller acquires IMU data at 50 Hz, preprocesses it through filtering and normalization, applies the classifier to detect harmful postures, measures the duration of sustained harmful postures, and activates the haptic motor when said duration exceeds a user-configurable threshold of 1 to 10 seconds. The device operates entirely locally without cloud dependency, making it suitable for remote rural areas. Optional BLE connectivity enables data logging for occupational health research and workplace interventions. The invention provides real-time haptic feedback that creates immediate awareness of harmful postures, enabling workers to correct their posture before injury occurs. The cost-effective component selection and ruggedized design ensure affordability and durability for agricultural environments. This solution fills a critical gap in occupational health technology for the agricultural sector.
Disclaimer: Curated by HT Syndication.
