首页> 外文会议>IEEE Annual Conference on Decision and Control >A fast method for solving guard set intersection in nonlinear hybrid reachability
【24h】

A fast method for solving guard set intersection in nonlinear hybrid reachability

机译:非线性混合可达性中保护集相交的快速求解方法

获取原文

摘要

Reachability computation formulates the problem of simulating the behavior of a continuous or hybrid dynamical system in a set-theoretical framework. Compared to the stochastic approach, it provides guaranteed results and has been shown highly valuable for verification or synthesis tasks. This issue is still quite challenging for uncertain nonlinear hybrid dynamical systems.Recently, [1] proposed a method for solving the flow/guard intersection problem that is at the core of hybrid reachability. It first derives an analytical expression for the boundaries of continuous flows using interval Taylor methods and techniques for controlling the wrapping effect. It then expresses the event detection and localization problem underlying flow/guard intersection as a constraint-satisfaction problem (CSP). One of the main issues in interval integration is to control, at each step, the overestimation of the reachable state set due to the wrapping effect. For this purpose, [1] only relies on the geometrical transformation induced by Lohner's QR-factorization method [4], which acts at the integration step. But when dealing with hybrid systems, another source of overestimation exists at the transition step. This paper describes an efficient method for solving flow/guard intersection: using the standard contractor HC4Revise at the transitions step, we will show how to minimize both the overestimation of the flow/guard intersection and the computational complexity, hence computation time. Interestingly, the geometrical transformation introduced by Lohner's QR-factorization method combined with our method, eventually minimizes the overestimation for the whole hybrid flow trajectory. The performance of the new method is illustrated on examples involving typical hybrid systems.
机译:可达性计算提出了在集合理论框架中模拟连续或混合动力系统行为的问题。与随机方法相比,它提供了有保证的结果,并且已被证明对验证或综合任务非常有价值。对于不确定的非线性混合动力系统来说,这个问题仍然具有很大的挑战性。最近,[1]提出了一种解决流动/保护相交问题的方法,它是混合可达性的核心。它首先使用区间泰勒方法和控制包裹效应的技术来导出连续流边界的解析表达式。然后将流/保护交叉点下面的事件检测和定位问题表示为约束满足问题(CSP)。间隔积分的主要问题之一是在每个步骤中控制由于包裹效应而导致的可到达状态集的高估。为此,[1]仅依靠由Lohner的QR分解方法[4]引起的几何变换,该变换在积分步骤中起作用。但是,在处理混合系统时,过渡阶段存在另一个高估来源。本文介绍了一种有效的解决流/护相交的方法:在过渡步骤使用标准承包商HC4Revise,我们将展示如何最小化对流/护相交的过高估计和计算复杂度,从而减少计算时间。有趣的是,Lohner的QR分解方法与我们的方法相结合引入的几何变换最终使整个混合流轨迹的高估最小化。在涉及典型混合系统的示例中说明了新方法的性能。

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