All different approaches resulted in a tighter fit to the data than the original method, although it must be stressed that the efficiency in time varies greatly. The tightest fit was achieved through the neuronal networks with an additional non-linear layer. The tightness of the fit, however, depends on the temperature and the state of charge.

A blown up view of a fit using local subspace is given, for some drive cycle data, at

 

• The Objective
• The Methods
  
  • Subspace Methods
  • Neural Networks
  • Non-Linear Methods
• The Findings
  
  • The Fit to the Data
  • The Mathematical Nature of the Problem
  • Reproducibility & Applications