Improved Six Rules-Based Fuzzy Compensation for Time-Varying Control Cycle in Active Powertrain Oscillation Reduction

Abstract

Abstract To prevent the degradations of comfort and drivability, this research presents an improved fuzzy-logic-based active oscillation control algorithm for an automotive powertrain with a time-fluctuated control cycle limitation. A sampled-data controller is firstly employed as the base control, and a model-predictive algorithm is implemented to compute the optimal control signals at a periodic time interval. However, the update timing of the real control input fluctuates over time due to the powertrain actuator limitation. The fuzzy reasoning is applied to tackle the time-fluctuated control cycles. In the present method, the update timing is translated into some fuzzy sets. Moreover, the periodic control signals by the sampled-data controller are combined as fuzziness, resulting in linguistic fuzzy sets. Six fuzzy rules are created based on those fuzzy sets to derive feasible control inputs at the fluctuated updating timings. The control strategy is tested via simulations with a powertrain oscillation model. The compensation effect is observed for the two patterns of the fluctuated control cycles. In addition, the test results confirm the improvement of the oscillation reduction over a conventional control algorithm.