Road grade and vehicle mass estimation for heavy-duty vehicles using feedforward neural networks
Paper in proceedings, 2019

In this paper, a neural network approach is presented for solving the problem of estimating road grade and vehicle mass, for the case of simulated heavy-duty vehicles (HDVs) driving on highways. After training, and using only signals normally available in HDVs, the (feedforward) neural network provides road grade estimates with an average root mean square (RMS) error of around 0.10 to 0.14 degrees, and mass estimates with an average RMS error of around 1%, when applied to two different test data sets (one with synthetic roads and one based on a real road), not used during the training phase. The estimates obtained outperform road grade and mass estimates obtained with other approaches.

Neural networks

Road grade estimation

Mass estimation

Author

Sina Torabi

Chalmers, Mechanics and Maritime Sciences, Vehicle Engineering and Autonomous Systems

Mattias Wahde

Chalmers, Mechanics and Maritime Sciences, Vehicle Engineering and Autonomous Systems

Pitoyo Hartono

Chukyo University

4th International Conference on Intelligent Transportation Engineering, ICITE 2019

316-321 8880261

4th International Conference on Intelligent Transportation Engineering, ICITE 2019
Singapore, Singapore,

Subject Categories

Infrastructure Engineering

Probability Theory and Statistics

Signal Processing

DOI

10.1109/ICITE.2019.8880261

More information

Latest update

12/5/2019