文章詳目資料

車輛工程學刊

  • 加入收藏
  • 下載文章
篇名 基於實數編碼之遺傳演算法於無人自走車的路徑規劃
卷期 9
並列篇名 Path Planning of an Automatic Guidance Vehicle Using Real-Coded Genetic Algorithm Based Path Exploration
作者 林志哲劉政鑫林啟瑞游勝凱
頁次 067-082
關鍵字 基因演算法無人自走車路徑規劃避障任務Labviewreal-coded geneticalgorithm,path planningAGV
出刊日期 201206

中文摘要

近年來由於自動化技術與機器學習等相關技術快速發展,無人自走車(Automatic Guidance Vehicle,AGV)的技術逐漸成熟,因此AGV在居家生活中與自動化工業上亦逐漸重要。在居家移動機器人的研究中,路僅規劃一直是個重要的議題,尤其是在有障礙物的工作環境中,如何將路徑做最佳化的規劃與選擇,常用最佳化路徑的方法為最短路徑法、花費時間最少或是最安全路徑等。本文將研究實數編碼之基因演算法(Real-coded Genetic Algorithm,RGA),應用RGA於移動型機器人的路徑最佳化,以提供無人自動化載具在具有障礙之空間下,能有效的避開障礙物並進行任務。本文提出基於四種原則:(1)與任何障礙物不發生碰撞,(2)路徑盡可能短,執行時間盡可能少,(3)應與障礙物保持一定的安全距離,(4)路徑必為連續點,做為基因演算法所的基本準則。並應用Labview軟體將此基因演算法將以實現,求出最佳路徑的結果後,可將此程式提供於Robotino無人自走車之實現。

英文摘要

In the researches of automatic guidance vehicles (AGV), path planning with obstacle avoidance plays a very important role and has been a very challenging research topic. For path planning, it should produce continuous path from the starting point to the destination without colliding obstacles. Therefore, the real-coded genetic algorithm is developed in this work to search the path with the shortest path in Labview environment. The difficulty of the genetic algorithms applied to the mobile robot is how to reduce the complexity of the genetic operations, and how to avoid the region optimal solution with adaptation to environmental change. Many researchers studied genetic algorithms to identify the optimal solution for path planning, but the past literatures used the binary code for the gene encoding. If the environment exists the paths with many obstacles, using the binary code will be very difficult due to the length of gene. This situation results in large computing time and it makes the real-time control impossible. Therefore, a novel method using real-encoded genetic algorithm is proposed to solve the above problems as well as reducing the computing time.

相關文獻