Systems and methods for accurately predicting state of charge in battery powered systems

A tangible, non-transitory computer-readable medium storing instructions that, when executed, cause a processing device to:

• generate, at a cloud-based computing system, a computer-implemented model comprising a domain adaptation architecture, wherein the computer-implemented model is trained to predict a state of charge of a target battery cell, and the domain adaptation architecture comprises:
• a first private encoder that is pre-trained using first data pertaining to a plurality of battery cells in a fleet,
• a shared encoder that is trained using the first data and second data pertaining to a target battery cell, and
• a computer-implemented predictor that is trained to predict the state of charge of the target battery cell using output from the first private encoder and the shared encoder;
• and
• deploy, to an edge processing device, the shared encoder and the predictor to cause the edge processing device to predict the state of charge of the target battery cell when sensor data pertaining to a battery pack is received by the edge processing device.

A system comprising:

• one or more memory devices storing instructions;
• and
• one or more processing devices communicatively coupled to the one or more memory devices, wherein the one or more processing devices execute the instructions to:
• generate, at a cloud-based computing system, a computer-implemented model comprising a domain adaptation architecture, wherein the computer-implemented model is trained to predict a state of charge of a target battery cell, and the domain adaptation architecture comprises:
• a first private encoder that is pre-trained using first data pertaining to a plurality of battery cells in a fleet,
• a shared encoder that is trained using the first data and second data pertaining to a target battery cell, and
• a computer-implemented predictor that is trained to predict the state of charge of the target battery cell using output from the first private encoder and the shared encoder;
• and
• deploy, to an edge processing device, the shared encoder and the predictor to cause the edge processing device to predict the state of charge of the target battery cell when sensor data pertaining to a battery pack is received by the edge processing device.

A computing device comprising:

• one or more memory devices storing instructions;
• and
• one or more processing devices communicatively coupled to the one or more memory devices, wherein the one or more processing devices execute the instructions to:
• generate, at a cloud-based computing system, a computer-implemented model comprising a domain adaptation architecture, wherein the computer-implemented model is trained to predict a state of charge of a target battery cell, and the domain adaptation architecture comprises:
• a first private encoder that is pre-trained using first data pertaining to a plurality of battery cells in a fleet,
• a shared encoder that is trained using the first data and second data pertaining to a target battery cell, and
• a computer-implemented predictor that is trained to predict the state of charge of the target battery cell using output from the first private encoder and the shared encoder;
• and
• deploy, to an edge processing device, the shared encoder and the predictor to cause the edge processing device to predict the state of charge of the target battery cell when sensor data pertaining to a battery pack is received by the edge processing device.