Developmental System Based on the Attention Control of Humanoid Robot with Sensor Flesh Embedded with Multi-axes Tactile Sensors

本研究的目的是阐明多轴软触觉传感器仿人机器人硬件和软件的构建方法。作为本研究的成果，开发了嵌入多轴软触觉传感器的小型和成人类人机器人的软测量外部。同时，针对海量的分布式触觉传感器，提出了一种结构化数据的生成方法。此外，还开发了强化学习系统，该系统在机器人与对象之间的近距离交互行为期间调整适应函数。

The aim of this research is to clarify the method to construct both hardware and software for humanoid robots with multi-axes soft tactile sensors. As results of this research, soft sensor exteriors embedded with multi-axes soft tactile sensors for both small and adult-size humanoid robots are developed. Also, method to make structured data for the enormous amount of distributed tactile sensors is developed for them. Furthermore, reinforcement learning system which adjusts a function of adaptation during close interaction behavior between the robot and an object is developed.