自動停車之主要重點技術內容可分成三部份,第一部分為使用感測器建立車輛與停車空間之相對位置,本研究感測器是採用超音波(SRF10)之反射特性,利用兩個或兩個以上車用超音波量測出其與障礙物之間的距離,再透過幾何關係解析出本車與環境中障礙物的位置。第二部份為車輛最佳停車路徑軌跡規劃,此部分停車模式分別為路邊停車和倒車入庫,路邊停車是採用Triangular Function做設計、倒車入庫則採用四分之一圓加上一條直線做規劃,利用最佳曲線使停車過程中能以最短路徑做完停車動作。第三部份為停車過程之實際路徑追蹤控制,來研究根據人們駕駛經驗,提出智慧型自動停車控制方法,分別對路邊停車與倒車入庫兩種不同情況設計規劃自組織模糊控制器與輸入狀態回饋線性化控制器,並藉由本研究建立自動停車系統之相關策略分析。系統結合這三部份技術,以模型車進行實驗驗證,以期同時能代替駕駛人執行完整路邊停車與倒車入庫之任務。
Aimming the main automatic parking technology can be classified into three parts. The first one is to construct the relation position of the vehicle and the parking space by using various sensors. In the thesis, sensor are adopted ultrasonic(SRF10) as reflect characteristic, making angular analysis by two or more ones, accuracy analysis the distance between the model-car and obstacle, through geometry relation and completed environment information, further to judge the position of obstacle. The second one is the optimal parking trajectory planning and based on the relative parking position, parking situations can be divided into road side parallel parking(PP) and garage parking(GP) two categories, PP can be designed by triangular function; GP can be planned by the quarter circle plus a straight line, making a completed motion that use the shorter trajectory path by the optimal curve in the parking process. The third one is real-time vehicle moving path estimation and parking trajectory tracking control, an intelligent auto parking path tracking controllers will be designed for PP and GP based on driver experience, self-organizing fuzzy controller and input-state feedback linearization controller will be employed to design the path tracking controller. The auto parking system must well combine these technologies.