主動式履帶機器人應用於連續樓梯攀爬與避障策略之研究
No Thumbnail Available
Date
2015
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
本論文主要是設計並開發自主式履帶機器人之軟體、硬體架構，能於室內未知的環境中自主式的執行連續樓梯攀爬與避障任務。履帶式機器人使用Xtion感測器取得深度影像資訊來辨識環境中的目標物，不需要經過複雜的影像運算，即可辨識出不同環境下目標物的位置和距離，並且在夜晚也能正確地辨識目標物；透過加裝的自製雲台組，Xtion可以依據實驗的需要進行角度調整，輔助履帶式機器人完成不同環境的任務。履帶式機器人使用單板電腦UDOO作為控制核心，使得體積更小、開發上有更多彈性，能依據不同實驗的需要加裝額外的感測器。本論文提出了一種影像避障的方式，依據Xtion感測器提供的障礙物距離資訊計算出機器人左、右履帶速度差與機器人旋轉半徑之間的關係，讓機器人能持續行走且順利的避開障礙物，並且設計了一系列的動作模式，來解決履帶式機器人連續樓梯攀爬的問題，讓機器人能夠在室內環境中任意的移動，最後經由實驗結果證實了自製履帶式機器人能夠實現跨樓層的樓梯攀爬與避障等任務。
In this thesis, we mainly design and create software and hardware about an autonomous tracked robot, which is able to continuously climb stairs and avoid obstacles. The tracked robot uses a depth sensor, Xtion, to obtain data from depth image, and distinguishes targets in the environment. Moreover, without complicated computing, the tracked robot distinguishes targets' position and distance not only in different environment but also in dark conditions. By equipped the self-made cradle head, the sensor’s angle can be modified by users to help the tracked robot complete missions in different environments. The tracked robot’s core system is to use a single board computer, UDOO, whose characters are smaller volume and are convenient for all users to add extra sensors in different kinds of experiments. In this thesis, we announce a method about obstacles avoiding by using image processing, which accords obstacles’ distance data from the sensor and calculates the correlation between differential velocity of the tracked robot’s wheels and its rotation radius. This method let the tracked robot keep moving and avoiding obstacles successfully. Also, we design a series of action models to solve the problem when the tracked robot climbs stairs continuously, and make the tracked robot moves randomly in the environment. In the conclusion, experimental results show the tracked robot is able to climb stairs among different floors continuously and avoid obstacles when moving.
In this thesis, we mainly design and create software and hardware about an autonomous tracked robot, which is able to continuously climb stairs and avoid obstacles. The tracked robot uses a depth sensor, Xtion, to obtain data from depth image, and distinguishes targets in the environment. Moreover, without complicated computing, the tracked robot distinguishes targets' position and distance not only in different environment but also in dark conditions. By equipped the self-made cradle head, the sensor’s angle can be modified by users to help the tracked robot complete missions in different environments. The tracked robot’s core system is to use a single board computer, UDOO, whose characters are smaller volume and are convenient for all users to add extra sensors in different kinds of experiments. In this thesis, we announce a method about obstacles avoiding by using image processing, which accords obstacles’ distance data from the sensor and calculates the correlation between differential velocity of the tracked robot’s wheels and its rotation radius. This method let the tracked robot keep moving and avoiding obstacles successfully. Also, we design a series of action models to solve the problem when the tracked robot climbs stairs continuously, and make the tracked robot moves randomly in the environment. In the conclusion, experimental results show the tracked robot is able to climb stairs among different floors continuously and avoid obstacles when moving.
Description
Keywords
履帶式機器人, 樓梯攀爬, 避障, 深度感測器, tracked robot, stair climbing, obstacle avoidance, depth sensor