Design of Sensor Permanent Fail-safe Algorithm based on Deep Learning for 100Nm class Electro-Hydraulic Power Steering System of Medium and Heavy Commercial Vehicle

• Published in: 2019 International Conference on High Performance Big Data and Intelligent Systems (HPBD&IS) pp. 283 - 288
• Main Authors: Kim, Joo Hyung, Jeong, Seung Yeop, Kim, Hun Mo
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.05.2019
• Subjects: DC motors; Deep learning; EHPS; Fault Tolerance; Gears; Mathematical model; Permanent Fault; sensor Fail-safe; Sensor Fault; Steering systems; Tensor flow; Torque; Wheels
• ISBN: 9781728104669; 1728104661
• DOI: 10.1109/HPBDIS.2019.8735464

In this paper, Deep Learning-based sensor Fail-safe for Electro-Hydraulic Steering Systems (EHPS) of 100Nm commercial vehicle is introduced. There are four types of permanent faults, such as shortage, drifted sensitivity, drifted off-set and fixation. When steering wheel angle sensor and motor angular velocity sensor in steering system failed, permanent faults are recovered by using a Deep Learning-based sensor Fail-safe algorithm. The learning variables are given by Electro-hydraulic motor modeling of steering system and are trained by using Tensor Flow. The learning data are extracted through MATLAB/Simulink and TruckSim. The proposed algorithm is verified through Co-simulation of MATLAB/Simulink and TruckSim. Finally, this paper confirmed that the algorithm proposed recovered the faults of steering system sensors.