Programming Entertainment Robots

题目：CISE实验计算机科学博士后助理：娱乐机器人编程为了使人形机器人成为小规模娱乐应用的可行替代方案，需要更简单的编程方法。它们必须具有自然和娱乐的动作风格，并且通常需要大量的动作数据库来确保足够多样的行为。这个问题的一个潜在解决方案是使用人体运动数据作为关节轨迹的来源。用于人体运动数据使用的自动技术将提供引人注目的机器人性能所需的各种运动。该博士后助理将开发技术，将允许人类运动数据用于控制各种人形和娱乐机器人和模拟。该项目将涉及优化技术的开发以及物理机器人的构建和控制方面的重要实验。该研究将结合博士后在动力学、规划和运动滤波方面的专业知识，以及首席研究员在模拟、人体运动控制系统和处理运动捕捉数据方面的经验。

EIA - 0203912 Hodgins, Jessica Carnegie Mellon UniversityTITLE: CISE Postdoctoral Associates in Experimental Computer Science: Programming Entertainment RobotsTo make humanoid robots a viable alternative for small scale entertainment applications, easier ways of programming them are needed. They must have a natural and entertaining style of motion and often require substantial motion databases to ensure a sufficient variety of behaviors. One potential solution to this problem is to use human motion data as the source of joint trajectories. Automatic techniques for human motion data use would provide the variety of motion that is required for compelling robotic performances. The postdoctoral associate will develop techniques that will allow human motion data to be used to control a wide variety of humanoid and entertainment robots and simulations. This project will involve significant experimentation both in the development of optimization techniques and in the construction and control of physical robots. The research will combine the expertise of the postdoctoral associate in dynamics, planning and motion filtering with the principal investigator's experience in simulation, control systems for human motion, and processing motion capture data.