A new control method of automobile's pneumatic suspension systems based on machine learning algorithm

Based on control rules of fuzzy-logic-control (FLC) optimised by genetic-algorithm, a new control method of machine-learning-algorithm (MLA) developed by an adaptive-neuro-fuzzy-inference-system (ANFIS) is used to learn these optimal control rules. MLA is then applied to control automobiles' pneumatic-suspension-systems (PSS) to improve ride comfort. MLA's performance is then evaluated under different conditions on rigid roads and deformed terrains. Some outstanding results include: 1) the ride comfort on the deformed terrain is worse than on the rigid road. Thus, the automobile should move at low speed on the deformed terrain to ensure the comfort; 2) with MLA used, its control performance is relatively stable under all different operating conditions of the automobile; and 3) with control data maps of MLA learned from FLC's optimal rules, MLA improves the comfort better than FLC under various simulation conditions. Therefore, this new control method of MLA should be considered for application in automobile suspension systems.