• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文学位 >SINGLE STEP OPTIMIZATION OF COLLISION AVOIDANCE MANIPULATOR MANEUVERS (ROBOT, OBSTACLE).
【24h】

SINGLE STEP OPTIMIZATION OF COLLISION AVOIDANCE MANIPULATOR MANEUVERS (ROBOT, OBSTACLE).

机译：避免碰撞操纵器动作（机器人，障碍物）的单步优化。

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The dynamics of a robot manipulator represents a coupled nonlinear system subject to bounded control inputs. Current investigations have attacked the problem by linearizing and decoupling the original nonlinear system using external feedback transformation. It is called the "inverse dynamics" or, "computed torque method." The bounds of the new inputs created from this approach, are functions of the state, hence become a difficulty for any long term planning algorithm.; Instead of using a long term strategy design in two stages, in which one plans a complete path first then follows that path, a short term suboptimal strategy approach is proposed. In this short term suboptimal strategy one only plans one step ahead to immediately follow that path, with an N-step ahead precaution measure. Unlike the long term strategy, which requires heavy computation effort and most of the time can only be done off-line, the next step optimization strategy, since it only considers the situation in the next step, can easily handle the computation on-line. Thus the difficulty of the state dependent input bounds arisen from the inverse dynamics method becomes manageable. The other task demands, collision avoidance and stability, can also be incorporated in the proposed N-step ahead precaution measure. The proposed approach starts with the externally linearized robot dynamics. Then a discrete time model relating the present time step to the next time step is formulated, and the main task demand (i.e., maneuvering to the target positions) is expressed as a minimization problem. This problem minimizes the error between the next step and the target state. The state-dependent input bounds, the desire of avoiding obstacles and stability considerations are incorporated in the form of N-step (where N is a design parameter) ahead precaution constraints. Distortion of the computed torques or forces due to plant parameter variations is handled by variable structure control. The algorithms were tested first by simulations. Experiments using an IBM PC to drive a three-degree-of-freedom cylindrical arm were also performed. Simulations and experiments are discussed. The evaluations of the results are also presented.
机译：机器人操纵器的动力学表示受约束控制输入约束的耦合非线性系统。当前的研究通过使用外部反馈变换线性化和解耦原始非线性系统来解决该问题。它被称为“逆动力学”或“计算转矩方法”。用这种方法创建的新输入的界限是状态的函数，因此对于任何长期计划算法都是困难的。代替在两个阶段中使用长期策略设计（其中先计划一条完整路径，然后遵循该路径）的方法，提出了一种短期次优策略方法。在这种短期的次优策略中，仅计划提前一步，立即采取这条道路，并采取N步预防措施。与需要大量计算工作并且大部分时间只能离线进行的长期策略不同，下一步优化策略仅考虑下一步的情况，因此可以轻松地在线处理计算。因此，由逆动力学方法引起的与状态有关的输入范围的困难变得可控。其他任务要求（避免碰撞和稳定性）也可以纳入提议的N步提前预防措施中。所提出的方法从外部线性化机器人动力学开始。然后，建立将当前时间步骤与下一时间步骤相关联的离散时间模型，并将主要任务需求（即，操纵到目标位置）表示为最小化问题。此问题使下一步和目标状态之间的错误最小化。状态相关的输入范围，避免障碍的需求和稳定性考虑因素以预防措施约束之前的N步（其中N是设计参数）的形式包含在内。由于设备参数变化而导致的计算出的转矩或力的失真通过可变结构控制进行处理。首先通过仿真测试算法。还进行了使用IBM PC驱动三自由度圆柱臂的实验。讨论了仿真和实验。还介绍了结果评估。

著录项

  • 作者

    CHEN, NING.;

  • 作者单位

    Colorado State University.;

  • 授予单位 Colorado State University.;
  • 学科 Engineering Electronics and Electrical.
  • 学位 Ph.D.
  • 年度 1986
  • 页码 165 p.
  • 总页数 165
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类 无线电电子学、电信技术;
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号