MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641099384 A) filed by Dr. K. Balamurugan on August 17, 2026, for A Machine-Learning Battery Health Estimator With Synchronized Multi- Channel Signature Acquisition.
Inventor includes Dr. K. Balamurugan.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: A battery health estimation apparatus includes an isolated measurement interface, a bidirectional perturbation circuit, a simultaneous- sampling converter, an edge processor, memory and a battery-control interface. The perturbation circuit applies a bounded pseudo- random current waveform to a rechargeable battery while voltage, current, cell-tap and temperature channels are acquired from a common sampling clock. The processor derives instantaneous resistance, polarization resistance, relaxation time constant, impedance magnitude and phase, and cell-voltage dispersion from aligned response windows. A stored two-branch machine-learning model processes a temporal waveform tensor and an engineered-feature vector and fuses corresponding estimates of state of health and remaining useful life. An uncertainty engine suppresses an estimate when synchronization, sensor integrity or model-domain criteria are not satisfied. When confidence and safety criteria are satisfied, the control interface transmits a charge-current or discharge-current derating command to a battery controller. The apparatus converts synchronized measurements into a physical battery-control action while preventing unsafe excitation. (Accompanied Figure No. 1 to 5)
Disclaimer: Curated by HT Syndication.
