本論文基於蜂群演算法 (artificial bee colony algorithm, ABC algorithm) 提出改良式蜂群演算法 (Modified artificial bee colony algorithm, MABC algorithm) 解決最佳化問題,並透過適當選取目標函數應用於多移動機器人之即時路徑規劃與運動控制。蜂群演算法為基於蜜蜂覓食行為提出之仿生搜索法,其包括工蜂、觀察蜂、偵查蜂之行為模式。本文提出之改良式蜂群演算法能改善傳統蜂群演算法之性能,其中包含了選取精英個體以維持好的演化、解的分享提供當搜索的方向、即時更新策略立即更新最佳個體資訊、協同機制產生額外的個體、族群管理動態調整族群大小。我們將之應用在多移動機器人之路徑規劃問題上,首先為即時路徑規劃,根據與目標之距離、障礙物避障、避免機器人彼此碰撞,選取目標函數之方式規劃下一步之路徑; 路徑選擇後為機器人移動控制,利用PID類神經網路產生適當的控制力使機器人移動到理想位置目標附近,模擬結果與分析可說明提出之改良式蜂群演算法之性能與優勢。
Based on the artificial bee colony (ABC) algorithm, we propose the modified artificial bee colony (MABC) algorithm for solving optimization problem and apply it to the on-line path planning of multi-robots system by choosing proper objective function. The ABC motivated by the foraging behavior of bee colony .The colony consists of three kinds of bees, employed bees, onlooker bees, and scout bees. The MABC algorithm enhances the performance of ABC algorithm by using elite individuals for perserving good evolution, the solution sharing provides a proper direction for searching, the instant update strategy provides the newest inforamtion of solution, the cooperative strategy generates extra individuals for searching the solution. Finally, the population manager is proposed to adjust the population size. We use the MABC algorithm for solving on-line path planning of multi-robot system. At first, the on-line path planning based on objective values for Goal, obstacles avoidance, and robots avoidance. And then, the motion control to the planned neighborhood target is done by the PID-type neural network (PIDNN) controller. Simulation results are introduced to show the effectiveness and performance of the proposed approach.