俱目標偵測及扣合功能之全自主行動搬運機器人應用於老人及行動不便人士照顧

Abstract

本論文之研究著重於俱目標偵測及扣合功能之全自主行動搬運機器人應用於老人及行動不便人士照顧。行動搬運機器人為機器人學中廣為運用的研究領域。目前在工業運用的行動式搬運機器人大多採用設置於環境或目標物的電線、磁軌或標誌達到目標辨識、自動導航與路徑規劃等功能。若將此種機器人運用在家庭環境，在不確定因素較大的場合，或是環境無法設置電線或磁軌等裝置的時候，使用上將受到限制。 本論文提出應用於老人及殘障照顧之全自主行動搬運機器人。一位老人或行動不便者坐在一張椅子上，機器人可以將自動偵測環境、辨識椅子，將行動不便者連帶椅子搬運至想到達的地方，完成輔助老年人或殘障者行動之目的。此機器人將可應用在家庭環境，甚至是醫院或養老院中。論文描述機器人之硬體架構、感測器設置、辨識系統、行動路徑規劃，使機器人能夠在不設置標誌或電線的環境中也能達成搬運任務。本論文在Rao-Blackwellised粒子濾波與蒙地卡羅定位法的架構中提出建置地圖與自動導航之系統，且發展演算法處理雷射測距儀之資訊與三微資料的特徵粹取，使機器人能夠準確自主偵測並辨識目標物，並完成目標物之搬運。 本論文包含機器人之地圖建置、定位、目標感測、動作與路徑規劃和自主扣合功能。採用演算法以及感測器中的資料，使行動式搬運機器人能在各種環境中自主運作。論文中以各種演算法、三維感知與自動扣合功能以及實驗結果，驗證行動搬運機器人之概念。

The objective of this paper is to describe the autonomous mobile carrier robot system with perception and docking for elder and handicap care. Mobile carrier robot is one of the most widely used applications in robotics. Nowadays, most of the carrier robots for industrial applications are autonomously guided by signs, magnetic lines or electrical wires marked on the ground. The robots would be unsuitable for domestic use or for highly uncertain labor environment, since it is difficult to set up the wires or signs in every environments where the robot works. The mobile carrier robot system for elder and handicap care is addressed in this thesis. The robot is able to autonomously carry a chair with an elderly person or a handicapped to desired place, aiding the person to move around. This kind of applications is helpful in domestic environment, or even for hospitals or nursing facilities. To reach this goal, problem of sensor perception and motion planning strategy is needed, making the robot without the need to follow markers or wires. In this paper, we developed map construction and navigation system based on Rao-Blackwellised particle filter and Monte Carlo localization. We also developed algorithms processing the signal from laser range finder and feature extraction methodologies of 3D data, so that the robot may detect and recognize the target autonomously with high stability and complete the carrying process. The thesis contains map building, localizing, perception system, motion planning and docking. With the algorithm and the data from both the laser range finder and the Xtion sensor, the carrier robot will be suitable to navigate through more versatile conditions. In this thesis, various algorithms for 3D perception and docking are developed, followed with experimental implementation and successfully demonstrated the proof of concept of the mobile carrier robot.