Parallel Algorithms and Architectures in Robotics

这是休斯研究实验室(HRL，James G.Nash)、南加州大学(USC，S.Y.Kung)和宾夕法尼亚大学(UPenn，Richard P.Paul)三个机构提议的联合研究工作。这项工作是为了开发一种用于机器人环境的特殊用途的脉动阵列计算机。在这项拟议的研究方面，三方每个人都有特殊的才能：HRL关注基于矩阵的线性代数计算机的体系结构和算法；南加州大学的研究旨在将问题映射到阵列；宾夕法尼亚大学专门研究机器人机械手控制问题。这项工作的目的是将高性能的脉动阵列处理器连接到机器人机械手，重新评估和改进处理器的设计，为这类机器开发新的并行算法，开发和评估各种适当的运动控制算法，以及开发将问题投影到这类机器上的算法。在项目负责人看来，这是三名一流的研究人员，他们提出了对可能对机器人控制领域产生重大影响的实验系统的研究。强烈建议使用修订后的预算和所附文件中增加的条件提供资金。

This is a proposed joint research effort among three institutions: Hughes Research Laboratory (HRL, James G. Nash), University of Southern California (USC, S.Y. Kung), and the University of Pennsylvania (UPenn, Richard P. Paul). This effort is to develop a special-purpose systolic-array computer for use in a robotics environment. Each of the three parties has special talents with regard to this proposed research: HRL is concerned with the architecture and algorithms for matrix-based linear algebraic computers; the research at USC is aimed at mapping problems into arrays; and UPenn specializes in robot manipulator control problems. The purposes of this effort are to interface a high-performance systolic-array processor to a robot manipulator, to re-evaluate and improve on the design of the processor, to develop new parallel algorithms for such machines, to develop and evaluate various appropriate kinematic control algorithms, and to develop algorithms for projecting problems onto machines of this type. In the opinion of the program director, these are three first-class researchers who propose research on an experimental system that may significantly impact the robotics control area. Funding with the revised budgets and added conditions in the attached documents is strongly recommended.