Fault Diagnosis of the Vehicle Tire Pressure Using Bayesian Networks with Real-Time ROS Applications

In today's engineering applications, model-based fault diagnosis methods are widely used to reduce costs. This study continues previous work[1,2] on model-based fault diagnosis by integrating the residual value structure of tire pressure into an existing Bayesian network, aiming for more accurate fault detection.A residual value is modeled using the pressure values of a vehicle's four tires, and the Bayesian network is updated accordingly, enabling a stochastic rather than deterministic approach. The updated method is first modeled and tested in the Matlab/Simulink environment. Following this, the algorithm and resolution procedures for obtaining tire pressure values from the vehicle are updated in the ROS environment. The method is then validated through real vehicle tests.During these tests, the tire pressures are deliberately reduced to create a fault scenario. A car lighter pump is used to lower the tire pressure, and the updated Bayesian network is tasked with detecting and identifying the faulty tires. The detected faults are displayed on the Human-Machine Interface (HMI) in real-time, providing feedback on tire pressure status.This integration of the Bayesian network with the residual value structure allows for more accurate and reliable tire fault detection, enhancing both safety and efficiency. The study highlights the importance of combining model-based methods with practical testing to validate diagnostic algorithms. The successful verification of the designed method through real vehicle tests marks a significant advancement in automotive fault diagnosis.

[1] T.Bodrumlu, M.M.Gozum, and Batıkan Kavak, “Enhanced Fault Detection of Vehicle Lateral Dynamics Using a Dynamically Adjustable Bayesian Network Structure and Extended Kalman Filter”, ASME International Mechanical Engineering Congress and Exposition, 2023, V009T14A024.

[2] M.F Yalcin, T. Bodrumlu; M. M. Gozum ; E. Ates. Dinamik Bayes Ağ Yapısı ve Genişletilmiş Kalman Filtresi Kullanılarak Gerçek Zamanlı ROS Uygulaması ile Otonom Bir Araçtaki Yanal Dinamiklerdeki Arıza Tespitinin Gerçeklenmesi