Indirect Tire Pressure Monitoring System based on Deep Learning

In this paper, a means of monitoring tire pressure indirectly based on deep learning is proposed. The proposed indirect tire pressure monitoring system (TPMS) uses the anti-lock braking system (ABS) sensor. By using the frequency characterstics of wheel speed data collected by ABS sensor, the deep neural network (DNN) classify the tire status.The proposed indirect TPMS algorithm shows very robust performance regardless of driving condition of vehicle. The proposed method is expected to overcome the limitation of previous indirect TPMS algorithm. The proposed algorithm is verified using vehicle test. In addition, the performance of DNN classifier is verified using five-fold cross validation, which is well-known validation method in machine learning field. In conclusion, performance of the deep learning based indirect TPMS algorithm proposed in this paper is compared with previous indirect TPMS method.