认知启发的稳态可以平衡机器人中相互冲突的需求
Cognition-inspired homeostasis can balance conflicting needs in robots
摘要:
稳态使动物存活;这是一个基本的过程,使动物能够迅速适应环境。人工稳态可以用来帮助机器人适应不断变化的环境。以前为机器人开发人工稳态的尝试是通过模仿驱动人类稳态的生化机制来驱动的。通过从认知的角度考虑稳态,我们开发了一个相对简单的机器人控制器名为CogSis(COGESHOMEostaSIS),并证明它可以提供稳态的机器人,即使有冲突的需求。我们目前的实验表明,运行CogSis的机器人能够从以前的经验中学习,并利用它们来影响未来的行为;可以保持其充电水平,同时参加另一项任务(在与充电站分开的区域中加热自己);并且能够保持其充电水平,同时避免冲突的需要(保持凉爽,当充电站放置在环境的炎热区域时)。结果在模拟和从一个真正的机器人平台。
英文摘要:
Homeostasis keeps animals alive; it is a fundamental process that allows animals to adapt quickly to their environment. Artificial homeostasis can be used to help robots adapt to changing environments. Previous attempts at developing artificial homeostasis for robots were driven by mimicry of the biochemical machinery that drives homeostasis in humans. By considering homeostasis from a cognitive perspective, we develop a comparatively simple robot controller named CogSis (COGnitive HomeostaSIS) and demonstrate that it can provide homeostasis to a robot, even when there are conflicting needs. We present experiments showing that a robot running CogSis is able to learn from previous experiences and use them to influence future behaviour; can maintain its charge level while attending to another task (warming itself in an area separate from the charging station); and is able to maintain its charge level while avoiding a conflicting need (keeping cool, when the charging station is placed in a hot region of the environment). Results are presented in simulation and from a real robot platform.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。