An experimental comparison of path planning techniques applied to micro-sized magnetic agents
An experimental comparison of path planning techniques applied to micro-sized magnetic agents
复制标题
应用于微型磁性体的路径规划技术的实验比较
DOI：
--
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
S. Misra
中科院分区：
文献类型：
--
作者：
S. Scheggi;S. Misra
微尺寸试剂可用于微组装、微操作和微创手术等应用。可以控制诸如顺磁性微粒的磁性试剂以将药剂递送到人体内难以接近的区域。为了自主地将这些微粒朝向目标/目标区域移动，必须使用路径规划算法来计算无障碍路径。在文献中已经开发了几种路径规划算法，然而，据我们所知，只有少数已被用于实验方案。在本文中，我们进行了实验比较的六个路径规划算法时，适用于顺磁性微粒的运动控制。在确定性和概率性路径规划器中，我们选择了我们认为最基本的路径规划器，例如：四叉树A*，均匀网格A*，D* Lite，人工势场，概率路线图和快速探索随机树。我们考虑一个由动态和静态障碍物构成的2D环境。四个场景进行了评估。三个指标，如计算时间，由微粒执行的轨迹的长度，以及到达目标的时间被用来比较的规划者。实验结果表明，几乎所有考虑的规划者之间的等价性的轨迹长度和完成时间。在计算时间方面，A* 与四叉树和人工势场取得了最好的性能。
Micro-sized agents can be used in applications such as microassembly, micromanipulation, and minimally invasive surgeries. Magnetic agents such as paramagnetic microparticles can be controlled to deliver pharmaceutical agents to difficult-to-access regions within the human body. In order to autonomously move these microparticles toward a target/goal area, an obstacle-free path must be computed using path planning algorithms. Several path planning algorithms have been developed in the literature, however, to the best of our knowledge, only few have been employed in an experimental scenario. In this paper we perform an experimental comparison of six path planning algorithms when applied to the motion control of paramagnetic microparticles. Among the families of deterministic and probabilistic path planners we select the ones that we consider the most fundamental, such as: A* with quadtrees, A* with uniform grids, D* Lite, Artificial Potential Field, Probabilistic Roadmap and Rapidly-exploring Random Tree. We consider a 2D environment made by both dynamic and static obstacles. Four scenarios are evaluated. Three metrics such as computation time, length of the trajectory performed by the microparticle, and time to reach the goal are used to compare the planners. Experimental results reveal equivalence between almost all the considered planners in terms of trajectory length and completion time. Concerning the computation time, A* with quadtrees and Artificial Potential Field achieve the best performances.