摘要
本文的目的主要在研製一具影像伺服割草機器人，並發展一套簡單且有效率的路徑規劃方法，來提高割草工作的實際覆蓋率。此割草機器人由一台兩輪驅動的自走車及一具背負式引擎割草機所組成，自走車包含四個系統，分別是嵌入式影像感測器、Arduino微控制器、馬達驅動系統及障礙物偵測系統。本文提出一種新的螺旋路徑規劃方法，結合影像感測器及適應模糊控制理論，透過一系列目標物影像追蹤控制，可以消除往復式路徑規劃方法所造成的累積定位誤差，也因為不須偵測方位及草坪邊緣，感測控制系統的成本及初始設定工時都比一般割草機器人低。 為了驗證本文所提路徑規劃方法的有效性能，吾人設計了一個客觀比較實驗，分別讓螺旋路徑規劃方法及往復路徑規劃方法操作在完全覆蓋的情況下，並比較兩者的重複率。實驗結果顯示，本文所設計的方法可以降低20%的重複率，對於節能及環保貢獻顯著。
並列摘要
In this study, a robot lawn mower with vision-servo system is developed to achieve higher coverage rate and lower overlap rate by using a simple area-coverage path planning algorithm. The hardware of robot lawn mower is composed of an autonomous two-wheel mobile robot and a grass mowing machine. The autonomous mobile robot includes four parts, embedded vision sensor, Arduino-based microcontroller, motor controller and obstacle-detect system. With a spiral path, the proposed novel algorithm can eliminate the accumulation of errors, which occurred frequently in the reciprocal path planning process, by utilizing vision sensor and adaptive fuzzy sliding mode control (AFSMC) to realize a sequence of objects tracking control. In addition, the requirements for robot positioning by using orientation sensor (electrical compass or/and GPS) and virtual wall (a low voltage boundary wire for creating an invisible fence to define the cutting area) can be relieved. For verifying the performance of the proposed algorithm, a comparative experiment was carried out to achieve the same coverage rate (100%) for both the algorithms. The experiment result showed that the proposed algorithm can lower the overlap rate from 30% to 10% compared with classical reciprocal path planning approach. This means that the proposed algorithm can contribute a 20% reduction in energy waste to the environmental protection.