Investigation of Vehicle Speed Prediction from Neural Network Fit of Real World Driving Data for Improved Engine On/Off Control of the EcoCAR3 Hybrid Camaro Colorado State University

Format:

Conference/Event

Author/Creator:

Baker, Baker, author.

Contributor:

Asher, Zachary

Bradley, Thomas

Conference Name:

WCX 17: SAE World Congress Experience (2017-04-04 : Detroit, Michigan, United States)

Language:

English

Physical Description:

1 online resource

Place of Publication:

Warrendale, PA SAE International 2017

Summary:

The EcoCAR3 competition challenges student teams to redesign a 2016 Chevrolet Camaro to reduce environmental impacts and increase energy efficiency while maintaining performance and safety that consumers expect from a Camaro. Energy management of the new hybrid powertrain is an integral component of the overall efficiency of the car and is a prime focus of Colorado State University's (CSU) Vehicle Innovation Team. Previous research has shown that error-less predictions about future driving characteristics can be used to more efficiently manage hybrid powertrains. In this study, a novel, real-world implementable energy management strategy is investigated for use in the EcoCAR3 Hybrid Camaro. This strategy uses a Nonlinear Autoregressive Artificial Neural Network with Exogenous inputs (NARX Artificial Neural Network) trained with real-world driving data from a selected drive cycle to predict future vehicle speeds along that drive cycle. Various prediction windows are analyzed and compared to quantify tradeoffs between prediction window size and speed prediction error for a given drive cycle. To investigate the fuel economy (FE) improvement potential of this new control strategy, a high fidelity model of a Toyota Prius, developed by Colorado State University, is used. An optimal dynamic programming (DP) engine controller is implemented in the Prius model. Several exemplar controllers are studied for the specified drive cycle: the model baseline controller, a DP derived engine controller using NARX Artificial Neural Network speed predictions, and a DP derived engine controller using a 100% accurate speed prediction. These simulations allow for investigation into the tradeoffs between different prediction window sizes. Additionally, the results provide insight into what FE benefit can be expected from speed prediction compared to baseline and idealized conditions. This potentially achievable FE benefit is used as motivation to develop a predictive controller that can be implemented in real-time on the supervisory controller of CSU's Plug-in Hybrid Electric (PHEV) Camaro

Notes:

Vendor supplied data

Publisher Number:

2017-01-1262

Access Restriction:

Restricted for use by site license