- 學位論文
使用多感測器之自主式行動輔助及導航機器人設計與實現
Design and Implementation of Autonomous Mobility Aids and Navigation Guidance Mobile Robot Using Multi-sensors
摘要
本論文針對輪式移動機器人，開發一套智慧型行動導航系統。機器人系統使用兩個模糊控制器，分別利用Kinect感測器及紅外線距離感測器於導航環境中進行障礙物偵測。在導航的行動規劃中，本系統利用RFID感應進行機器人定位決策，並提供導引軌跡，引導機器人繼續前進或是轉向。前方障礙物主要由Kinect感測器進行偵測，擷取影像的深度資訊以判別障礙物與機器人的遠近關係。此外，障礙物的水平位置可以被量測，以至於模糊控制器能使機器人在考慮機器人目前移動速度來進行躲避障礙物。紅外線距離感測器則負責偵測Kinect感測器無法偵測到的底部障礙物，所對應的模糊控制器就可以進行底部的障礙物閃避。機器人具備避障及循牆功能後，再搭配RFID定位的設計，進一步利用B-Spline曲線理論規劃機器人移動軌跡，可完成自動導航的目的。為了整合避障及導航功能，本論文設計行為監督者來協調所需要的機器人移動控制。實驗結果顯示，即使在未知的環境條件下，本論文所開發的機器人系統能順利躲避多個障礙物，並且能成功到達所規劃的目的地。
並列摘要
This paper develops an intelligent navigation system for wheeled mobile robot. The robot has two fuzzy controllers which respectively utilize the Kinect sensor and infrared distance sensors to detect the obstacles in the navigation environment. In the navigation strategy, RFID sensing is used to provide the robot with localization data, and motion guidance such as moving forward or turning. The obstacles are mainly detected by the Kinect sensor, which intercepts the depth information of fetching the image in order to differentiate the relative distance between the obstacles and the robot. Besides, the horizontal location of obstacles can be obtained such that the fuzzy controller can drive the robot to dodge the obstacles while considering the current robot moving speed. The infrared distance sensors are responsible to detect the downside obstacles that cannot be sensed by the Kinect sensor, so that the corresponding fuzzy controller can avoid the downside collision. With the capability of obstacle-avoidance and wall-following, the RFID-based localization strategy is carried out by using B-Spline curve theory to generate a motion trajectory for robot navigation. A fuzzy supervisor is designed to compromise the motion control from the obstacle-avoidance and the navigation capability. The experimental results show that the multiple obstacles can be avoided even under conditions of unknown environments, and the robot system can achieve the required destination.