US 11,946,515 B1
Real-time machine learning and physics-based hybrid approach to perform eLSD torque estimation
Arash Hashemi, Waterloo (CA); Sresht Gurumoorthi Annadevara, Toronto (CA); Naser Mehrabi, Richmond Hill (CA); SeyedAlireza Kasaiezadeh Mahabadi, Novi, MI (US); and Lapo Frascati, Dicomano (IT)
Assigned to GM GLOBAL TECHNOLOGY OPERATIONS LLC, Detroit, MI (US)
Filed by GM Global Technology Operations LLC, Detroit, MI (US)
Filed on Sep. 18, 2023, as Appl. No. 18/469,009.
Int. Cl. F16D 48/06 (2006.01); F16H 48/34 (2012.01); F16H 48/20 (2012.01)
CPC F16D 48/06 (2013.01) [F16H 48/34 (2013.01); F16D 2500/10425 (2013.01); F16D 2500/3115 (2013.01); F16D 2500/3117 (2013.01); F16D 2500/70647 (2013.01); F16D 2500/70668 (2013.01); F16H 2048/204 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for determining, in real-time, an electronic limited-slip differential (eLSD) clutch torque, comprising:
receiving vehicle data in real-time, wherein the vehicle data includes a torque request;
determining a preliminary eLSD clutch torque using a neural network and the vehicle data;
determining clutch torque bounds of the eLSD using a physics-based model;
determining whether the preliminary eLSD clutch torque is outside the clutch torque bounds of the eLSD;
in response to determining that the preliminary eLSD clutch torque is outside the clutch torque bounds of the eLSD, adjusting the preliminary eLSD clutch torque using the clutch torque bounds to determine a final clutch torque of the eLSD; and
commanding, in real-time, the eLSD to apply the final clutch torque to a clutch of the eLSD.