基因演算法於車輛懸吊系統最佳化之研究－－以二分之一車模型為例

Optimization of Vehicle Suspension System Using Genetic Algorithm- a Case Study of the Half Car Model

本論文主旨在運用一套針對車輛主動懸吊系統設計的控制方法，其中包含力量消除、天勾阻尼與路面跟隨彈簧、虛擬輪胎阻尼等概念，並以遺傳演算法整合上述概念應用於軍用悍馬車的二分之一車模型，以獲取較佳的乘坐舒適性及操控性。研究內容可分成二部分：首先搜尋懸吊系統最佳被動元件參數，接著根據第一部分所獲致的最佳被動元件參數來搜尋最佳主動控制參數；吾人並將電腦模擬分析之結果與傳統被動式懸吊系統做比較。結果顯示應用上述設計概念所發展出的主動控制方法，在搜尋參數過程中，於設定世代內均可有效率搜尋最佳被動元件參數及最佳主動控制參數，針對兩種路面之模擬結果，在俯仰運動、前輪操控性、以及垂直舒適性等三項平均改善分別達54.58％、43.28％、以及41.60％，僅在前懸吊工作行程一項中稍有犧牲，證明此方法可明顯改善車輛乘載之舒適性及操控性，此研究對於未來車輛懸吊系統之研改工作有相當之參考價值。

In this study, control algorithms are developed for force control in an active vehicle suspension design using genetic algorithms with linear half car model of a M998 High Mobility Multipurpose Wheeled Vehicle. Force cancellation control scheme, skyhook damper and spring, road following spring, and virtual tire damper are employed in the design concepts, which are used to design the force controller to achieve better ride and handling quality. This research is divided into two parts. The first part is to find the best passive suspension parameters. And those passive parameters are used to find the best passive parameters of active system in the second part. It is demonstrated that the active control model converges within specific generations in the parameter searching processes. Computer simulations indicate that the improvements for pitching, front wheel handling and vertical ride quality are 54.58%, 43.28%, and 41.60%, respectively. And only front suspension stroke indicates minor degrading. This study will have potential contribution to future relevant research on vehicle suspension system.