A novel neuro-fuzzy controller to enhance the performance of vehicle semi-active suspension systems
This paper proposes a neuro-fuzzy (NF) strategy to implement semi-active suspension in passenger vehicles. The proposed method is composed of two parts: a NF controller (NFC), to establish an efficient controller strategy to improve ride comfort and road handling (RCH), and an inverse mapping to estimate the semi-active suspension current. To effectively estimate the current needed to control the semi-active damper, an inverse mapping based on neural network, modified back-propagation (MBP) is presented. The inverse mapping is incorporated into the FC to enhance RCH. Given the relative velocity between the mass and the base and also the absolute acceleration of the mass, the FC computes the optimum damping coefficient. The fuzzy logic rules are extracted based on expert knowledge encapsulated in skyhook and groundhook. A quarter-car model was adopted for the purpose of simulating and experimenting with the proposed NFC. To verify the performance of the FC, two sets of results are reported. First, an experimental analysis was performed to demonstrate the effectiveness of the FC in comparison with the benchmark skyhook and Rakheja-Sankar controllers. Furthermore, a random input was considered to examine the robustness of the NFC in comparison with the other adopted controllers. It was shown that the developed NFC control enhances the performance of the quarter-car system significantly, in terms of both ride comfort and handling characteristics. Second, four FCs with the same control strategies were implemented on a full-vehicle model to demonstrate the effectiveness of the proposed control strategy in reducing the propensity to rollover. It was concluded that the developed FC enhances the RHC and also has the potential to increase the stability of vehicles.