MUMBAI, India, Oct. 5 -- Intellectual Property India has published a patent application (202641116270 A) filed by Sathyabama Institute Of Science And Technology on September 28, 2026, for Artificial Intelligence-Based Human Behavior Analysis And Prediction System.
Inventors include Dr. Senduru Srinivasulu; Dr. G. Rajeswari; Ms. M. Madhumathi; Ms. U. A. Ranjni; Ms. D. Saral Jeeva Jothi; Ms. Percy Paulin J; Ms. N. Jeenath Shafana; Ms. R. Shanmugaa Priyaa; and Ms. S. Priya Lakshmi.
The application for the patent was published on October 02, 2026, under issue no. 40/2026.
Abstract: The present invention discloses an Artificial Intelligence-Based Human Behavior Analysis and Prediction System configured to collect, process, integrate, and analyze multimodal behavioral information for recognizing current behavior and predicting probable future actions. The system comprises a data-acquisition layer, pre- processing module, feature-extraction module, multimodal fusion engine, artificial intelligence module, behavioral baseline database, prediction module, explain ability module, notification module, and security module. Authorized data may be obtained from cameras, microphones, wearable devices, smartphones, environmental sensors, computing devices, and digital interaction platforms. The collected data may represent facial expressions, speech characteristics, gestures, body posture, movement patterns, physiological signals, activity histories, and contextual conditions. The pre-processing module performs noise removal, normalization, segmentation, synchronization, missing-value treatment, and anonymization. Extracted features are assigned reliability-based weights and combined to generate a unified behavioral representation. Machine-learning and deep-learning models compare the representation with personalized or group-based behavioral baselines to recognize activities, identify patterns, detect deviations, and estimate probable behavioral transitions or risk conditions. Each prediction may include a confidence score, uncertainty value, expected time interval, contributing factors, and an explainable reasoning summary. Alerts or recommendations are generated when configurable thresholds are satisfied and are transmitted to authorized users for human review. The system further incorporates informed-consent controls, encryption, role-based access, audit logging, data-retention management, bias monitoring, and model validation. The invention provides a scalable, adaptive, context-aware, privacy- preserving, and human-supervised decision-support framework suitable for healthcare, education, workplace safety, assisted living, transportation, customer service, security monitoring, and human–computer interaction applications without treating predictions as definitive determinations of individual intention or future conduct.
Disclaimer: Curated by HT Syndication.
