Adaptive Driving Situation Characterization for Predicting the Driving Load of Electric Vehicles in Uncertain Environments

Javier A. Oliva and Torsten Bertram
Submission Type: 
Full Paper
phmce_14_022.pdf1.07 MBMay 26, 2014 - 5:26am

Battery powered electric vehicles (EVs) have emerged as a promising solution for reducing the consumption of fossil fuels in modern transportation systems. Unfortunately the battery pack has a low energy storage capacity, which causes the driving range of the EV to become very limited. It is therefore essential to properly characterize the different driving situations of the vehicle in order to better predict the driving load along the road ahead and to better estimate the remaining driving range (RDR). However, this prediction cannot be achieved straightforward due to sources of uncertainty introduced by the randomness of the driving environment. In this paper a novel approach for characterizing driving situations and for predicting the driving load of an EV is presented. The prediction of the driving load occurs in a model-based fashion, where the model input variables are modeled as discrete-time Markov processes. An approach for estimating the transition probabilities between Markov states in the presence of sparse driving data is introduced. Furthermore, to capture the changes in the driving environment a Bayes-based methodology for recursively updating the established transition probabilities is presented. The validity of the proposed approach is illustrated through simulation and by a series of experimental case studies.

Publication Year: 
Publication Volume: 
Publication Control Number: 
Page Count: 
Submission Keywords: 
model based prognostics
Markov chain
Uncertainty Quantification
Bayesian updating
Submission Topic Areas: 
Model-based methods for fault detection, diagnostics, and prognosis
Modeling and simulation

follow us

PHM Society on Facebook Follow PHM Society on Twitter PHM Society on LinkedIn PHM Society RSS News Feed