Emergence of Sociality in Robots based on Embodied Evolution

基于具身进化的机器人社会性的出现

基本信息

批准号：
16500125
负责人：
ARITA Takaya
金额：
$ 2.24万
依托单位：
Nagoya University
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research (C)
财政年份：
2004
资助国家：
日本
起止时间：
2004 至 2005
项目状态：
已结题

来源：
https://kaken.nii.ac.jp/en/grant/KAKENHI-PROJECT-16500125/
关键词：
Embodied Evolution Theory of Mind Multilevel Selection Evolutionary Robotics Genetic Algorithm Swarm Robotics Prisoner's Dilemma Fitness Landscape 進化心理学誤認人工生命

项目摘要

Evolutionary robotics is the new and active field of applying computational models of evolution to the design of autonomous robots. However, it suffers from the problems that are derived from the gap between evolutionary simulation and real robot environments. This study established the concepts of embodied evolution and evolutionary computation methods for robots, and constructed a framework for experimental environments for them. Using the framework, we developed our understanding of human minds and society from an evolutionary perspective as follows. 1) We focused on the techniques of evolutionary computation for generating players performing tasks cooperatively and conducted comprehensive evaluation for 18 candidate methods using the ultimately simplified soccer game. 2) We constructed an agent-based model for multilevel selection, in which group-level selection and individual-level selection in population are supposed to work simultaneously, with dynamic group-restructuring, and conducted evolutionary experiments using it. 3) We defined a minimal model of emotion in robots based on a behavioristic theory on human emotions, and conducted evolutionary simulations based on the definition of emotions to verify the evolutionary adaptivity with the scenario of the origin and the evolution of human emotions in mind. 4) We conducted evolutionary simulations using a model describing physically-situated agents moving around avoiding collisions in order to investigate the dynamics inherent in the mechanism of recursion in the theory of mind. 5) We constructed an evolutionary model of quantitative traits by using an extended version of Kauffman's NK fitness landscape, in which whether each phenotype is plastic or not is genetically defined and plastic phenotypes are adjusted by learning, in order to give a new insight into the benefit and cost of learning by focusing on the quantitative evolution of phenotypic plasticity under the assumption of epistatic interactions.

进化机器人技术是将进化计算模型应用于自主机器人设计的新的活跃领域。但是，它遭受了从进化模拟与真实机器人环境之间的差距中得出的问题。这项研究确定了机器人的体现进化和进化计算方法的概念，并为其为实验环境构建了一个框架。使用该框架，我们从进化的角度开始对人类思想和社会的理解，如下所示。 1）我们专注于进化计算的技术，用于生成合作执行任务的玩家，并使用最终简化的足球游戏对18种候选方法进行了全面评估。 2）我们构建了一个基于代理的模型，用于多级选择，其中应该同时进行小组级别的选择和个体级别的选择，并具有动态的群体重塑，并使用它进行了进化实验。 3）我们根据人类情感的行为理论定义了机器人中情绪的最低模型，并根据情感的定义进行了进化模拟，以验证其起源场景的进化适应性和人类情感的进化。 4）我们使用描述物理固定的药物绕着避免碰撞的模型进行了进化模拟，以研究心理理论中递归机理固有的动力学。 5）我们通过使用Kauffman的NK NK健身景观的扩展版本构建了定量性状的进化模型，其中每种表型是否是塑料的，都是遗传定义的，并且通过学习来调整塑性表型，以通过学习的良好洞察力和成本来通过学习的益处和成本来进行ePSISTIPT塑料的定量进化，以使学习的益处和成本。