Adaptive Neuro-Fuzzy Control of Active Vehicle Suspension Based on H2 and H∞ Synthesis

Abstract

This paper addresses the issue of a road-type-adaptive control strategy aimed at enhancing suspension performance. H2 synthesis is employed for modeling road irregularities as impulses or white noise, minimizing the root mean square (RMS) of performance outputs for these specific road types. It should be noted, however, that this approach may lead to suboptimal performance when applied to other road profiles. In contrast, the H∞ controller is employed to minimize the RMS of performance outputs under worst-case road irregularities, taking a conservative stance that ensures robustness across all road profiles. To leverage the advantages of both controllers and achieve overall improved suspension performance, automatic switching between these controllers is recommended based on the identified road type. To implement this adaptive switching mechanism, manual switching is performed, gathering input–output data from the controllers. These data are subsequently employed for training an Adaptive Neuro-Fuzzy Inference System (ANFIS) network. This elegant approach contributes significantly to the optimization of suspension performance. The simulation results employing this novel ANFIS-based controller demonstrate substantial performance enhancements compared to both the H2 and H∞ controllers. Notably, the ANFIS-based controller exhibits a remarkable 62% improvement in vehicle body comfort and a significant 57% enhancement in ride safety compared to passive suspension, highlighting its potential for superior suspension performance across diverse road conditions.