MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202641084801 A) filed by Gitam Deemed To Be University on July 10, 2026, for Real-Time Adaptive Body Bias Voltage Control System For Subthreshold Vlsi Circuits Using Reinforcement Learning.
Inventor includes Masood Ahmad Mahammad.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: Exemplary embodiments of the present disclosure are directed towards a real-time adaptive body bias voltage control system for subthreshold VLSI circuits using reinforcement learning (RL) and method employed thereof. The real-time adaptive body bias voltage control system includes a real-time monitoring sensor module is configured to collect real-time operational data such as propagation delay, temperature, supply voltage, and leakage current, and to generate feedback signals representing a current operating state of a subthreshold VLSI circuit block. A reinforcement learning (RL) agent is configured to receive the feedback signals as state inputs, determine one or more body bias voltage settings based on the current operating state, and generate and continuously update a policy that maps observed states to corresponding body bias voltage actions in order to maximize a predefined reward function associated with power efficiency and circuit performance. A reward evaluation module is configured to evaluate a power–performance trade-off corresponding to an applied body bias voltage and to generate a reward signal indicative of circuit performance improvement. A body bias voltage generator module is configured to generate and apply the selected body bias voltages to the subthreshold VLSI circuit block to dynamically adjust transistor threshold voltages and associated circuit characteristics in response to varying operating conditions. Fig. 1
Disclaimer: Curated by HT Syndication.
