無人載具感測器融合與控制系統設計

Sensor Fusion and Control System Design for an Unmanned Vehicle

蘇軒儀(Shuan-Yi Su)

指導教授 ： 王立昇

共同指導教授 ： 張帆人

國立臺灣大學/工學院/應用力學研究所/碩士(2010年)

https://doi.org/10.6342/NTU.2010.01536

摘要

本研究之主要目的在整合GPS接收機、慣性導航系統(INS)、電子羅盤、雙眼視覺感測器以及超音波測距感測器，以發展一套戶外無人載具系統。該系統包括主控站、參考站與移動載具三部分，採用全球定位系統載波相位差分法(CDGPS)進行全域定位，並整合慣性導航系統高資料率(data rate)之優點，以提高定位頻率。經由定位結果得知載具與終點之相對位置後，由電子羅盤取得載具之姿態角，透過模糊控制器，使載具能追蹤參考路徑往終點前進。
本系統並引入雙眼立體視覺，透過平面視差的觀念，利用左右影像之資訊可偵測前方之障礙物，並以雙眼視覺測距法求得障礙物相對於載具之位置與距離，進行障礙物定位，並在考慮終點位置與載具位置後，利用設定障礙物周圍之路徑點的方式規劃避障路徑進行避障。系統中另利用六個超音波測距感測器偵測載具周遭之環境，判斷障礙物是否存在，並透過即時動態避障模式閃避障礙物。
本研究中發展之戶外無人載具系統成功的整合上述之感測器及避障方式，使載具在不同的環境中能順利應用各種避障模式以及終點模式抵達終點。研究中所進行之各個避障模式實驗之結果顯示，本論文提出之方法確實可行。

關鍵字

無人載具 ； 雙眼立體視覺 ； 平面視差法 ； 雙眼測器法 ； 慣性感測器 ； 全球定位系統 ； 卡爾曼濾波器

並列摘要

The objective of this research is to integrate GPS receiver, inertial navigation system(INS), electronic compass, stereo camera, and ultrasonic range sensors to develop an outdoor unmanned vehicle system. The developed system consists of three parts : the main station, the reference station, and the unmanned vehicle.
The position of the unmanned vehicle is provided by GPS with triple difference carrier phase method. The INS, with its high data rate, is integrated with GPS to increase the positioning rate. With the attitude angle which is obtained from electronic compass and the position of the vehicle, the unmanned vehicle is able to track the reference path by a fuzzy controller.
The obstacles in front of the vehicle can be detected by the stereo camera by using the principle of plane induced parallax. The position and distance of the obstacles with respect to the unmanned vehicle can be determined by the method of stereo triangulation. The waypoints around the obstacle are then selected. With the information of the position of the destination, the unmanned vehicle will choose the shorter path through the waypoints to avoid the obstacles and reach the destination point.
The unmanned vehicle is also equipped with six ultrasonic range sensors to detect the obstacle around it. With the help of ultrasonic sensors, the unmanned vehicle can avoid the collision with the moving obstacles in real time. Using various obstacle-avoidance modes and the goal-approaching mode, the unmanned vehicle can reach the destination point without the collision with the surrounding obstacles. The experimental results show that the proposed integration and control methodologies are feasible.

並列關鍵字

Unmanned vehicle ； Stereo vision ； Plane induced parallax ； Stereo triangulation ； Inertial navigation system ； Global positioning system ； Kalman filter

參考文獻

[13] 吳宣誼，“雙無人載具協同控制與實驗”，台灣大學應用力學研究所碩士論文, 中華民國九十四年七月。

[15] 蘇信豪， ”無人載具避障之導航與控制”，台灣大學應用力學研究所碩士論文, 中華民國九十六年七月。

[16] 單幼玫， “無人車雙眼視覺避障系統之設計”， 台灣大學應用力學研究所碩士論文， 中華民國九十八年七月。

[2] A. Ramírez-Serrano and M. Boumedine, “Ultrasonic Sensing and Fuzzy Logic Control for Navigation in Unknown Static Environments”, IEEE Advanced Mobile Robot, Oct., pp.54-59, 1996.

[3] K.J. Kyriakopoulos, P. Kakambouras, and N.J. Krikelis, “Navigation of Nonholonomic Vehicles in Complex Environments with Potential Fields and Tracking”, IEEE International Conference on Robotics and Automation, Vol. 4, pp.3389-3394, 1996.