Battery SOC Estimation Using Dual-Model Multicore Correction

The patent divides the SOC estimation task into two separate cores: the first core performs rapid but less accurate estimation using the extended Kalman filter with electric circuit model, while the second core performs accurate but computationally intensive estimation using electrochemical model. This segmentation allows each core to specialize in different aspects, with the second core correcting errors from the first core without requiring the entire system to slow down.

The second core acts as an intermediary that periodically corrects the SOC estimation from the first core. Instead of having the first core directly handle all corrections, the second core serves as a mediator that provides accurate reference values to adjust and improve the first core's estimates, thereby reducing error rates while maintaining the first core's rapid response capability.

The patent segments the computation into two distinct parts handled by two different cores. The first core uses simplified electric circuit models for fast computation, while the second core uses complex electrochemical models for high accuracy. This segmentation allows the system to achieve both speed and accuracy by distributing different computational tasks to specialized processing units.

Instead of continuously running the computationally intensive electrochemical model, the patent applies it partially - only periodically from the second core to correct and refine the estimates from the first core. This partial application of the complex model provides sufficient accuracy improvement without incurring the full computational cost continuously, thereby maintaining high productivity.