细菌趋向性启发的多机器人图案生成策略

Bacterial chemotaxis-inspired strategy for multirobot pattern formation

针对多机器人图案生成问题,本文提出了一种细菌趋向性启发的多机器人分阶式控制策略。该策略将图案生成过程分解为2个阶段:聚集阶段和图案生成阶段。聚集阶段,设计基于平均距离的适应度函数,其值作为机器人的感知输入。机器人根据当前时刻与上一时刻适应度数值变化执行类似细菌趋向的直行或翻转来完成聚集。引入搜索因子促进多子群融合,提升多机器人聚集的成功率。图案生成阶段,针对机器人翻转角度生成问题,引入决策因子来评估邻居作用程度,提升多机器人图案生成的成功率。仿真实验结果表明:针对多机器人六边形和三角形图案生成,本文所提出策略平均迭代次数分别为25.36和93.83,成功率分别为83.33%和96.67%,优于相关对比算法。

We present a multirobot hierarchical control strategy inspired by bacterial chemotaxis for multirobot pat-tern generation.This strategy divides the pattern generation process into two stages:the aggregation stage and the pattern generation stage.In the aggregation stage,the fitness function based on the average distance is developed,and its value is employed as the perception input of the robot.Robots aggregate by moving forward or rolling over,similar to bacterial chemotaxis,based on the fitness value change between the current time and the last time to ag-gregate.The search factor is introduced to promote multiple subgroup fusion and enhance the success rate of multi-robot aggregation.In the pattern generation stage,aiming at the problem of rolling-over angle generation of the ro-bot,the decision factor is introduced to assess the degree of the neighbor effect to enhance the success rate of multi-robot pattern generation.The simulation results show that for the hexagon and triangle pattern generation of the mul-tirobot,the average iteration times of the proposed strategy are 25.36 and 93.83,respectively,and the success rates are 83.33%and 96.67%,respectively.The new algorithm is superior to the compared algorithms.