MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202621048141 A) filed by Indian Institute Of Technology, Bhilai on April 15, 2026, for Hybrid Optimization-Trained Neural Network Estimator For Sensorless Pmasyrm Drives In Light Electric Vehicles.
Inventors include Kumar, Dr. Shailendra; Kumar, Sushant; and Kurm, Shashank.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: The present disclosure provides an optimized neural network-based estimation system (100) for permanent magnet-assisted synchronous reluctance motor (114) drives in solar photovoltaic-battery powered light electric vehicles comprises a solar photovoltaic array (102) connected to a boost converter (104), a battery energy storage system (108) interfaced through bidirectional DC-DC converter (110) to DC bus (106), a three-phase voltage source inverter (112), and a context neural network-based estimator (116). The context neural network-based estimator (116) receives stator voltage and current measurements and simultaneously estimates rotor speed and machine parameters including stator resistance, d-axis inductance, and q-axis inductance. The context neural network is trained using a hybrid particle swarm- gravitational search algorithm combining exploration capability of gravitational search optimization with exploitation strength of particle swarm optimization. A vector control unit with Karush-Kuhn-Tucker-based maximum torque per ampere technique (130) generates optimal current references, maintaining proper field orientation and efficient torque control without mechanical speed sensors. (Figure 1)
Disclaimer: Curated by HT Syndication.
