利用合作式基因最佳化法之機器人路徑規劃

Abstract

機器人路徑規劃功能需考慮：最短路徑、避免與障礙物碰撞、平滑的路徑以及快速的運算處理，基本上是一最佳化的問題。故本論文提出一最佳路徑規劃的作法，利用基因演算法(Genetic Algorithm)所具備之最佳化搜尋能力，配合菁英化策略與路徑優化處理，有效提高收斂效率以及導航成功率。實驗結果證明，本論文所提出的路徑規劃法有效改進了傳統基因演算法，實現更安全、更高效率的機器人路徑規劃。以及考慮到近年來智慧型手機應用程式的普及化，本文將改良後的基因演算法實現在Android系統並應用於室內行動機器人導航，發展出讓人類與機器人的互動更為貼近的室內行動機器人導航App。

Path planning for mobile robots needs to consider several issues including the shortest path, obstacle avoidance, and computation efficiency, which can be regarded as an optimization problem. Taking advantage of the genetic algorithms to solve various optimization problems, this paper first proposes a Cooperative Genetic Optimization (CGO) Algorithm, including the establishment of an elite policy and larger selection region to minimize the occurrence of local optima so as to increase the speed of convergence. Based on the proposed CGO, a global path planning approach for robots is then presented. As a result, the proposed method leads to a better performance to reach the goal in terms of a safer and shorter path in comparison with the traditional genetic algorithm. Considering of App development is very popular recently. In this paper, App based development of robots path planning using Cooperative Genetic Optimization is a great contribution and innovation for development of indoor mobile robot navigation.