Development of Multi-Legged Robot Platform and Its Application to Make Evolution of Intelligent Machine Research

本研究项目的目标是开发一种多腿机器人平台，能够普遍用于智能机器人的各种研究。我们特别关注1995年至1997年在研究项目资助下开发的机器人平台，称为“智能机器人”，我们已经发展了协作工作，将平台转变为先进版本，以实现我们的目标。(1)多足机器人平台jrob2的开发：我们将jrob1的所有组件（视觉系统、电路板、计算机和软件，这些都是在研究资助项目上开发的）集成和标准化，并将它们组装成一个先进的多足机器人平台jrob2。视觉系统有两个摄像头，它的平移、倾斜和拥堵是可控的。它能够实时主动处理三维图像。配备陀螺仪和其他传感器来测量机器人的价值。作为操作系统，我们安装了Real-Time Linux，机器人的所有基本行为都是用C语言编写的。(2) jrob2在日本机械工程学会年会上的教程：开发jrob2的团队于1999年举办了jrob2的软硬件和基本行为库的教程，目的是使机器人研究人员广泛使用该机器人。(3)在基于平台的智能机器进化方面，协作研究者以自己的方式处理了广泛的研究课题。具体来说，他们研究了机器人在运动中的姿态与机动性之间的定量关系，通过叠加反应行为、运动规划、协同运输以及视觉和行为的整合来产生运动。这些研究结果发表在一份300多页的报告中。

The objective of this research project is to develop a multi-legged robot platform that is capable of being commonly utilized on performing various researches of intelligent robots. We have especially focused on the robot platform that has been developed from 1995-1997 on the research project grant-in-aid referred to as "intelligent robot," and we have evolved collaborative work to transform the platform into its advanced version for achieving our aim.(1) Development of a multi-legged robot platform, Jrob-2 : We have integrated and standardized all the components of Jrob-1 (which are a vision system, circuit boards, computer, and software and which have been developed on the research project grant-in-aid), and we have assembled them into an advanced multi-legged robot platform, Jrob-2. The vision system has two cameras and its pan, tilt, and congestion are controllable. It is capable of actively processing 3-dimensional images in real time. Gyroscopes and other sensors are equipped to take the valance of the robot. As the operating system, we have installed Real-Time Linux and all the fundamental behaviors of the robot are coded in C language.(2) Tutorial of Jrob-2 in the Annual Conference of the Japan Society of Mechanical Engineering : The team for developing Jrob-2 held a tutorial of its hard/soft ware and its fundamental behavior library in 1999 for the purpose of making the robot widely utilized by robotics researchers.(3) As for making evolutions of intelligent machine based on the platform, collaborative researchers have dealt with widely ranged research topics by taking their own approaches. Specifically, they have studied the quantitative relationship between the robot posture in locomotion and its mobility, generating locomotion by superimposing reactive behaviors, motion planning, cooperative transportation, and integration of vision and behavior. These research results are published in an over 300 pages report.