Classification of Tire Pressure in a Semitrailer Using a Convolutional Neural Network
Zygimantas Ziaukas, Alexander Busch, Mark Wielitzka, Tobias Ortmaier, Jan-Philipp Kobler · 2020
Test Online information about states and parameters of passive vehicles (e.g. trailers) is of high importance for the future of automotive driving and has been quite neglected until today. Direct measurements of these states may require costly additional hardware which costumers are not often willing to pay for. Therefore, in this paper a method for the classification of tire pressure for one tire of a commercial vehicle's semitrailer is presented. The classification is based on measurement of the adjoining axle's vertical acceleration and the wheel speed using a Residual Neural Network (ResNet). The tire pressure is divided into three classes of 8.5 bar, 7.0 bar and 5.5 bar. The experimental results show accuracies beyond 90% for the test case.