本論文以身心殘障者「行」的能力為出發點，基於路徑規劃、Wi-Fi定位技術、及機器人控制等之整合，發展出一智慧能力之輪椅機器人室內即時導航系統。藉由無所不在的Wi-Fi無線網路佈建作為定位的基礎，具有較不拘泥於空間條件的優勢，再取其佈建價廉無所不在的特性，是為定位技術的優秀選擇。本研究以接收訊號強度(RSSI)估算輪椅機器人與障礙物的位置，在已知環境知障礙空間中，擷取特徵地圖，以廣義Voronoi結構劃分(GVD)生成拓樸地圖，再透過D* Lite動態路徑規則演算法在地圖資訊中即時地規則一條可避開動態障礙物的路徑，其具有動態避障及最低成本路徑之特色。最後，修改一現成電動輪椅之電路與控制策略，使其能與場域中之Wi-Fi定位與動態路徑規劃互相連節配合，利用UDP無線通訊協定發送控制指令，令輪椅機器人循所規劃之動態路徑安全地到達目的地。整合工作完成，實驗結果顯示，在三個子系統獨立運作良好，且相互配合，功能符合預期，然因Wi-Fi定位系統精度難以突破，無法精確導航，尚缺臨門一腳。 Evidenced by the importance of the mobility in human-beings'' life, the study sought to develop an intelligent mobile robotic system which could autonomously navigate a wheelchair in various indoor environment. The system, integrated several techniques of robotic path planning, Wi-Fi localization, and robot control together, tries to get aids for human-beings, especially for those difficult in traversing their living space. Based on an ubiquitous Wi-Fi which has been deployed widely in existing infrastructure and is intrinsically RF propagated in the working space, the Wi-Fi localization can be promisingly in the use of indoor navigation to locate the stuffs, including the wheelchair, in the space. As the known configuration of the space, a path planner D* lite together with its prior generalized Voronoi diagram (GVD) tessellation were employed for planning the path. A wheelchair is thus remodeled with a compliant controller as an automated vehicle to be guided to follow the planned path. The application emphasizes that the motion of the wheelchair can dynamically react to the changes of the obstacle configuration even change the corresponded route to reach the goal. In the study, the developed techniques can work well independently with one exception when they integrated together. Due to low precision and slow response of the Wi-Fi localization, there are no suffciently reliable references of the stuffs to recover the environment and configure the space. It finally makes the steps can not be followed up even if they have been well-established.
