模擬自動駕駛之機器人研究

基於人工智慧的發展，自駕車研發技術已成為新的科技領域之熱門研究議題。本研究將探討自駕車之技術，並且以AutoRace自駕車競賽規則為本研究之議題。結合機器人之視覺系統與光學雷達之定位演算法，於移動機器人TurtleBot3上，達成機器人自駕之目標。TurtleBot3機器人使用Raspberry Pi 3晶片，透過外接RPLidar-A1光學雷達以及攝影機為其視覺與感測系統。為提升整體運算效能，將利用一台高效能電腦使用ROS（Robot Operating System）系統進行物件辨識與智能計算，透過通訊界面傳送指令給TurtleBot3來執行動作，以達成智能自駕車之功能。針對AutoRace自駕車比賽規則，透過物件辨識直線追蹤、感測與定位系統之演算法設計，讓機器人自駕車可以行駛在道路之軌道內，辨識交通號誌並且依照號誌執行動作，如：紅綠燈、鐵路平交道、路口左右轉、停車等。最後機器人自駕車在隧道內無光線處，透過光學雷達即時掃描與定位比對（SLAM，Simultaneous localization and mapping）方法，讓機器人自行避開障礙物與離開隧道，達成自駕車之競賽功能。

關鍵字

自動駕駛；人工智慧； Turtlebot3機器人；路牌辨識； SLAM演算法

並列摘要

Based on the evolution of artificial intelligence (AI), the topic of self-driving car has become popular in the new technology field. This research will explore the technology of self-driving cars, and the research topic is based on the rules of the AutoRace self-driving car competition. Combining the robot's vision system and the positioning algorithm of the optical radar, the mobile robot TurtleBot3 achieves the goal of robot self-driving. The TurtleBot3 robot uses a Raspberry Pi 3 chip and uses an external RPLidar-A1 Lidar and camera for its vision and sensing system. In order to improve the overall computing performance, a high-performance computer will use the ROS (Robot Operating System) system to perform object recognition and intelligent calculation and send commands to TurtleBot3 through the communication interface to perform actions to achieve the function of intelligent self-driving. According to the rules of the AutoRace self-driving car competition, through the design of algorithms such as object recognition, line tracking, sensing, and positioning systems, the robot self-driving car can drive on the track of the road, recognize the traffic signs, and execute actions according to the signs, such as: traffic lights, railway level crossings, turning left and right at intersections, parking, etc. Finally, the self-driving car can avoid obstacles and leave the light-free tunnel, by itself using the optical radar instantly scanning and positioning comparison (SLAM, Simultaneous localization, and mapping) method to complete the self-driving car competition.

並列關鍵字

Auto Driving ； Artificial Intelligence ； Turtlebot3 Robot ； Road Sign Recognition ； SLAM Algorithm

主題瀏覽