This paper deals with road recognition by the use of neural networks on vibration signals. In particular, an accelerometer sensor is used and it acquires the vibrations transmitted in the contact between the tyres and the ground. Twelve different types of road have been tested at different speed of the car. Based on these data a neural network has been proposed to correlate a set of suitable characteristics of the vibration signal with the type of road. The effectiveness of the resulting neural network has been proved on a control set of data. Moreover the paper reports a sensitivity analysis of the neural network in order to minimize the number of inputs needed and to make it rugged.
- Design Engineering Division and Computers in Engineering Division
Use of Neural Networks in Road Recognition by Vibration Data
- Views Icon Views
- Share Icon Share
- Search Site
Reggiani, G, Cocconcelli, M, Rubini, R, & Lolli, F. "Use of Neural Networks in Road Recognition by Vibration Data." Proceedings of the ASME 2010 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference. Volume 4: 12th International Conference on Advanced Vehicle and Tire Technologies; 4th International Conference on Micro- and Nanosystems. Montreal, Quebec, Canada. August 15–18, 2010. pp. 345-349. ASME. https://doi.org/10.1115/DETC2010-29083
Download citation file: