• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>2018 IEEE 16th World Symposium on Applied Machine Intelligence and Informatics >Extrapolated state estimation in fixed point transformation-based adaptive control using fractional order feedback
【24h】

Extrapolated state estimation in fixed point transformation-based adaptive control using fractional order feedback

机译：基于分数阶反馈的基于定点变换的自适应控制中的外推状态估计

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

In this paper the possible advantages of model-based state extrapolation is investigated in control tasks in which the signals of observation “travel” from the location of sensing to that of the controller, and then the calculated control commands “propagate” to the place of actuation. These effects have crucial significance in the context of the “Fixed Point Transformation-based Adaptive Control” (FPTBAC) that adaptively deforms the input of an available imperfect model on the basis of fresh observations without improving this model by persistent modifications. For equal signal travel times (TD) the actuation at instant t is based on the controller's calculation at instant t - Td that is based on the observations made at instant t - 2Tq, that is in a 2Td long time-interval neither making observations, nor intervention by actuators are possible. This “dead period” has double effects in the degradation of the controller's operation: a) the adaptive deformation is calculated on the basis of partly “obsolete” information, and b) any extrapolation within this interval can be done on the basis of an imperfect model. Since in the FPTBAC design of the required tracking error relaxation can be designed independently of the details of the dynamic model, it naturally allows the application of “Fractional Order” (FO) error feedback terms that has “long memory effects” in contrast to its integer order-based counterparts. For this purpose a novel application of the Riemann-Liouville fractional order derivatives is suggested. It is concluded that this new FO feedback-based design can be successfully combined with the extrapolation based on the approximate model, and the FPTBAC-based adaptivity.
机译：在本文中，研究了基于模型的状态外推法在控制任务中的可能优势，在这种控制任务中，观察信号从传感位置“移动”到控制器的位置，然后将计算出的控制命令“传播”到控制位置。致动。这些影响在“基于定点变换的自适应控制”（FPTBAC）的背景下具有至关重要的意义，该功能基于新的观测值自适应地使可用的不完美模型的输入变形，而无需通过持续修改来改进该模型。对于相等的信号传播时间（TD），在时刻t的驱动是基于控制器在时刻t-Td的计算，该计算基于在时刻t-2Tq进行的观测，即在2Td的长时间间隔内均未进行观测，执行器也无法干预。该“死区”对控制器的运行性能具有双重影响：a）根据部分“过时”的信息来计算自适应变形，并且b）在此区间内的任何外推都可以基于不完美的结果进行模型。由于在FPTBAC设计中可以独立于动态模型的细节来设计所需的跟踪误差松弛，因此自然而然地允许使用“分阶”（FO）错误反馈项，与之相比，它具有“长记忆效应”基于整数的对应项。为此，提出了黎曼-利维尔分数阶导数的新应用。结论是，这种新的基于FO反馈的设计可以成功地与基于近似模型的外推法和基于FPTBAC的适应性相结合。

著录项

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