MUMBAI, India, April 17 -- Intellectual Property India has published a patent application (202641022182 A) filed by Prathyusha Engineering College, Tiruvallur, Tamil Nadu, on Feb. 25, for 'ai based micro stress detector using finger touch pressure variations.'
Inventor(s) include Ms H Kezia; Mr Aadithiyan R; and Mr Bhuvanesh A.
The application for the patent was published on April 17, under issue no. 16/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to an artificial intelligence-based micro stress detection system and method configured to identify subtle physiological and behavioural stress variations through analysis of dynamic finger touch pressure behaviour combined with multi-sensor physiological signal monitoring. The invention utilizes a force-sensitive sensing interface to capture finger touch pressure signals and derives stress-related descriptors from pressure variability characteristics including micro fluctuations, variance behaviour, pressure instability patterns, stability metrics, and pressure transition dynamics. Unlike conventional stress detection systems relying on absolute pressure magnitude or rigid threshold mechanisms, the present invention performs stress inference based on dynamic pressure behaviour modelling. The system further incorporates physiological sensing modules configured to acquire heart rate and body skin temperature parameters, thereby enabling multi-dimensional stress assessment. Sensor-derived features are processed through noise filtering, normalization, and feature extraction mechanisms, wherein behavioural and physiological descriptors are combined using feature-level fusion to enhance prediction robustness and reduce sensitivity to individual sensor noise or variability. Stress classification is performed using a supervised Random Forest-bast::d machine learning engine configured to model nonlinear relationships among pressure variability descriptors and physiological indicators. Real-time prediction is achieved through continuous sensor streaming, temporal buffering, dynamic feature computation, and machine learning inference, enabling stable and noise-resilient stress state estimation. The invention provides a non-intrusive, cost-effective, and scalable architecture for real-time micro stress detection, suitable for applications including behavioural monitoring, physiological assessment, cognitive state analysis, human-machine interaction, and adaptive intelligent systems. 'Micro Stress Detection Capability: Enables identification of subtle stress variations not captured by conventional systems. Hynamic Pressure Behaviour Modelling: Improves inference accuracy through variability-based interpretation. Robust Stress Prediction: Multi-sensor fusion enhances prediction stability and reliability. Noise-Resilient Classification: Random Forest ensemble improves robustness under signal variability. Cost-Effective Implementation: Utilizes low-cost sensors and scalable system architecture."
Disclaimer: Curated by HT Syndication.
