MUMBAI, India, July 13 -- Intellectual Property India has published a patent application (202641084105 A) filed by Vellore Institute Of Technology on July 08, 2026, for System And Method For Observer-Augmented Reinforcement Learning Control Of A Rotary Flexible Link Apparatus.
Inventors include Vinodh Kumar E; Jitendra Kumar Goyal; Yuvapriya T; Nitin Mutyala; and Chilukuri Venkata Chandrahasa.
The application for the patent was published on July 10, 2026, under issue no. 28/2026.
Abstract: The invention provides a system (100) for observer-augmented reinforcement learning control of a rotary flexible link apparatus (102) having a rotary actuator (114) coupled to a flexible link (116), an angular position sensor (118), and a deflection sensor (120). A processor (106) executes a super twisting observer module (128) to generate a finite-time angular velocity estimate from angular position, and an observation vector generation module (130) to form an observation vector including measured angular position, tracking error, and estimated angular velocity. A soft actor-critic reinforcement learning control module (132), trained using tracking- error, control-effort, and flexible-link deflection terms, generates a continuous motor control signal applied through a power amplifier (112). The system (100) achieves trajectory tracking, vibration suppression, disturbance rejection, and energy-efficient actuation under payload variation, actuator delay, and external disturbance without mathematical modelling or full-state measurement.
Disclaimer: Curated by HT Syndication.
