RI: Medium: Active Sensing, Localization, and Mapping in Dynamic Deformable Environments for Image-Guided Interventions

这项研究旨在解决实现基于图像引导的针和导管手术的智能机器人系统的两个重要研究问题。第一个研究重点是利用医学成像对外科介入手术中的柔性器械和解剖目标进行定位和跟踪。在嘈杂、不完整的传感器测量下，稳健跟踪器械、目标组织和其他相关解剖结构(包括需要避免的障碍物和解剖结构)的算法对于机器人系统成功执行手术至关重要。第二个研究重点是开发动态控制过程中的医学成像系统的算法。目标是最大限度地减少传感器测量的数量，从而降低与传感相关的成本，同时仍然获得足够的传感信息。该项目将通过开发将改进经皮干预的技术，对医疗保健产生直接的广泛社会影响，经皮干预被广泛用于诊断和治疗应用。本研究的目的是开发主动控制术中医学成像系统的方法，以便在经皮介入治疗期间定位和跟踪器械和解剖结构。具体地说，研究将集中在动态变形环境下的主动定位和映射算法的开发，以及术中成像系统动态控制的主动传感算法的开发。所开发的技术将在硬件中实现，并在使用实时磁共振成像技术的临床介入性磁共振成像系统的台式实验研究中得到验证。该项目的学术价值来自于对机器人和医学成像领域的科学贡献，以及这两项技术的协同整合。

This study aims to address two important research problems in realizing intelligent robotic systems for image-guided needle and catheter based surgical interventions. The first research thrust focuses on localization and tracking of flexible instruments and anatomical targets during the surgical intervention using medical imaging. Algorithms for robustly tracking the instrument, target tissue, and other relevant anatomical structures, including obstacles and anatomical structures that need to be avoided, under noisy, incomplete sensor measurements are essential for the robotic system for successfully performing the procedure. The second research thrust aims to develop algorithms for dynamically controlling the medical imaging system during the procedure. The goal is to minimize the number of sensor measurements, and hence reduce costs associated with sensing, while still obtaining sufficient sensing information. This project will have direct broad societal impacts in healthcare, by developing technologies that will improve percutaneous interventions, which are widely used for both diagnostic and therapeutic applications. The purpose of the present study is the development of methods for active control of intra-operative medical imaging systems for localization and tracking of instruments and anatomical structures during percutaneous interventions. Specifically, the research will focus on development of algorithms for active localization and mapping in dynamic deformable environments, and active sensing algorithms for dynamically controlling intra-operative imaging system. The developed technologies will be realized in hardware and validated in benchtop experimental studies on a clinical interventional magnetic resonance imaging system using real-time magnetic resonance imaging techniques. The intellectual merit of the project comes from the scientific contributions to the fields of robotics and medical imaging as well as a synergistic integration of the two technologies.