Experimental Analysis of Smart Tires

Abstract

A novel smart tire monitoring system was designed and implemented on a fully functional car tire. Polyvinylidene fluoride (PVDF) based piezo-electric sensors were embedded inside rubber tire to measure strain related data. System electronics were implemented inside a robust IP-68 (Ingress Protection) rated enclosure. This enclosure was mounted on a car wheel and successfully recorded sensory data onto an SD card during driving. Data collected from the PVDF sensors were then post-processed in Matlab. An artificial neural network (ANN) was built to correlate the sensor data to the readings given by an industry grade load wheel. Although the correlations are very crude, this study shows a promising way to analyze the strain related information from car tires by using PVDF sensors in conjunction with ANNs. This strain related information can then be used to estimate six different values concerning the tire, namely lateral force (Fy), longitudinal force (Fx), normal force (Fz), aligning moment (Mz), inflation pressure and friction coefficient. All of which are very important parameters for vehicle dynamics. However the estimation of these values is not presented within the context of this work.

Two low cost data acquisition systems were designed in-house with two different Arduino platforms. However these fell short of data acquisition performance requirements required for realistic driving applications. It was seen that the Arduino family, low-end microprocessors, were not the best choice for applications of this nature. Finally electronic improvements such as the usage of field programmable gate arrays (FPGA) is discussed and suggested for future works.