應用ANFIS於漫遊式機器人室內定位系統之設計與建構

Using the ANFIS to Construct and Design the infra-red Positioning System for the Autonomous wheel-based Robot in the Indoor Environment

謝智成(Chih-Cheng Hsieh) ； 季永炤(Yeung-Jaw Jih)

《國立虎尾科技大學學報》 30卷1期 (2011/03) Pp. 13-23

https://doi.org/10.6425/JNHUST.201103.0013

摘要

本文探討如何運用紅外線系統設計漫遊式機器人之室內定位系統，使其能在室內環境下自我重新校正其目前所在位置。對於漫遊式機器人之定位系統設計主要探討的問題有兩大類，其一爲如何解決環境對於紅外線系統之影響與如何避免紅外線發送源訊號重疊時產生訊號接收錯誤之問題。其二在於避免使用『三角函數』繁瑣的公式推論過程，使用模糊推論系統爲核心設計機器人之模糊定位推論系統。機體方面設計一個紅外線接收塔並搭配位置伺服馬達回饋其接收角度資訊。在紅外線接收器方面加裝光遮罩的方式一方面避免接收訊號重疊產生訊號接收錯誤。另一方面使紅外線接收器具有較好的指向性功能使接收角度資訊能夠較爲準確。在紅外線發送器方面以訊號編碼方式使接收器能夠過濾錯誤訊息的接收，並增加發送光源辨識能力。本文爲了使用模糊定位推論系統，首先在電腦平台建立反三角函數推論系統，利用此推論系統建立機體接收塔角度與車體實際位置之對應表，之後應用對應表經由MATLAB軟體中的適應性網路模糊推論系統(ANFIS)模組作訓練，調整出合適的模糊定位推論系統之模糊推論關係及歸屬函數，最後再將ANFIS訓練得到的模糊定位推論系統植入微控器中。在實驗方面首先分析紅外線接收器與發送源之間的距離對接收器接收角度與接收訊號源資料準確率之影響。最後再以隨機擺放車體實際位置與定位系統推估位置之實驗做比較了解其定位誤差，依照實驗結果探討此定位系統的可行性。

關鍵字

漫遊式機器人 ； 室內紅外線定位 ； ANFIS ； 三角函數推論

並列摘要

The using of the infra-red sensors to design a positioning system for the autonomous wheel-based robot for the indoor environment is studied in this article. The position system is to enable the robot to perform the self calibration and updates its present position. There are two major problems has been mostly discussed in the positioning system for the robot. One is the method to deal with the influence of the environment to the infra-red sensors system and to distinguish the source of the signal if more than one is detected. Another is avoiding the use of the trigonometric functions in the mathematical deduction. In here, the Fuzzy inference is used as the main algorithm to generate the updated position of the robot.An infra-red torrent is mounted on the top of the robot as the infra-red signal detecting sensors system. The sensor is used to read the coded signal emitted by the landmark placed in the known position of the area of roaming. The directionality of the sensor in enhance by the tube mounted in the front of the sensor. This also reduces the chance of reading more than one signal source. The Fuzzy-based position inference system which trained by ANFIS with data from using trigonometric method is implant into the microcontroller of the robot. The influence of the distance to the rate of data accuracy and the directional sensitivity is studied and analyzed through the field test. Finally, the robot is randomly placed in the area and the positioning error is studied for accessing the characteristics of the system.

並列關鍵字

Autonomous robot ； infra-red positioning system ； ANFIS

參考文獻

Siciliano, B.,Khatib, O.(2008).Springer Handbook of Robotics.New York:Springer.

Giuffrida, F.,Morasso, P.(1996).Active Localization Techniques for Mobile Robots in the Real World.IEEE/RSJ International Conference on Intelligent Robots and Systems.3,1312-1318.

Josep, M.,Joaqim, A.(2009).Consistent triangulation for mobile robot localization using discontinuous angular measurements.Journal of Automation of Robotics and Autonomous Systems.57(9),931-942.

Esteves, J. S.,Carvalho, A.(2003).Generalized geometric triangulation algorithm for mobile robot absolute self-localization.IEEE International Symposium on Industrial Electronics.1,346-351.

Bing-Fei, W.,Cheng-Lung, J.(2010).Neural Fuzzy Based Indoor Localization by Extending Kalman Filtering with Propagation Channel Modeling.Journal of Computer and Information Science.328-348.