Development of Active Antennase based on The Nesting Behavior of Wasps

黄蜂可以构造有规则排列的六边形蜂窝。本研究观察了天线在筑巢过程中的运动，并探讨了其构造方法。通过录像记录了基于天线运动的筑巢行为和巢的构建过程，以供分析和讨论。在筑巢过程中，天线的运动是主动的，在每个筑巢过程中，天线的运动是有序的、有特点的。观察结果表明，天线的特征运动可能决定了巢的形状，而筑巢行为与所谓的“趋向性”有关。用计算机模拟的方法讨论了黄蜂的构造方法。通过模拟来确定构造的形状和简单的移动规则（如“出租车”）之间的关系。仿真结果表明，简单的移动规则，如“的士”，导致直线，曲线和六边形的构造，从基地的形状为导向基于黄蜂的筑巢行为，研制了一种建筑机器人。这个建筑机器人有两个类似黄蜂触角的传感器。进行了施工试验。基于传感器运动和简单运动规则的构造算法，以基本形状为导向标记，构造出相似形状和不同形状我们试图获得一个最优的规则，可以追踪建筑的基本形状。用神经网络表示机器人的构造规则。通过计算机模拟，确定了一种可以跟踪基本形状的规则。采用遗传算法确定了最优跟踪规则。分析了最优规则。分析表明，遗传操作得到的规则与“趋向性”有关。

1 wasps can construct combs of hexagonal cells with a regular arrangement. In this study, the movements of antennae in nesting were observed and construction methods were discussed. Nesting behavior based on the movements of antennae and the process of comb construction were recorded on video tape for analysis and discussion. Antennae are moved actively in nesting and the movements are orderly and characteristic in each process of nesting. Observation results show that characteristic movements of antennae are probably used in determining how to shape the nest and that nesting behavior is connected with so called "taxis."2 Construction methods of wasps were disussed using computer simulation. Simulation was conducted to ascertain relationships between constructed shapes and simple moving rules such as "taxis." The simulation results show that simple moving rules such as "taxis" lead to the construction of straight lines, curved lines and hexagons, starting from the shapes of the base as a guide.3 A construction robot based on the nesting behavior of wasps was developed. The construction robot has two sensors similar to the antennae of wasps. Construction experiments were conducted. Using base shapes as a guide mark, construction algorithms based on the movements of sensors and simple moving rules lead to the construction of similar shapes and different shapes.4 We attempted to obtain an optimal rule that can trace basic shapes for construction. Construction rules of the robot were represented by means of neural networks. Computer simulation was conducted to ascertain a rule that can trace basic shapes. An optimal rule for tracing was determined by genetic algorithms. The optimal rule was analyzed. Analysis shows that the rule that was obtatined as the result of genetic operations is connected with "taxis."