NRI: Toward Safe and Reliable Robotic Eye Examinations

该项目建议启用一个完整的系统，用于机器人验光I眼科检查 独立式患者，包括自动角膜、晶状体、视网膜成像和眼镜 处方测量。机器人辅助体检有可能改善获得 预防性医疗，改善健康状况，降低护理成本，减少疾病 传播风险，但它们构成了安全问题，因为人类和 机械设备。此外，患者交互非常多样，因此在 异常情况的“长尾”需要稳健和适应性的感知和决策。 为了解决这些安全和可靠性问题，这个为期三年的项目将把基础研究整合到 实时人体运动感知和机器人运动规划及其在眼科和医学中的应用研究 生物医学成像。这项研究将沿着三个目标进行。AIM 1将通过增强提出者在机器人对准光学方面的现有工作来构建机器人安装的传感仪器 使用其他模式的相干断层扫描成像。目标2将通过利用 传感器、实时人体跟踪、人体运动预测和机器人方面的最新技术进展 实施主动安全系统的运动规划，以消除机器人引发的碰撞并将 人类发起的。目标3将通过对头部/眼睛模体的大规模研究和对人类受试者的较小规模的研究来评估和改进安全性和可靠性。

This project proposes to enable a complete system for robotic optometric I ophthalmic examination of freestanding patients, including automated cornea , lens, and retinal imaging as well as eyeglass prescription measurement. Robot-aided medical exams have the potential to improve access to preventative medical care, improve health outcomes, reduce costs of care, and reduce disease transmission risk, but they pose safety concerns due to the extremely close proximity of humans and machinery. Moreover, patient interactions are extremely varied , so automating data collection in the "long tail" of anomalous situations requires robust and adaptive perception and decision-making. To address these safety and reliability concerns , this three-year project will integrate basic research in real-time human motion sensing and robot motion planning with applied research in ophthamology and biomedical imaging. This research will be conducted along three aims. Aim 1 will build the robot-mounted sensing instrument by augmenting the proposers' existing work in robotically-aligned optical coherence tomography imaging with additional modalities. Aim 2 will address safety by leveraging recent technical advances in sensors, real-time human tracking , human motion prediction , and robot motion planning to implement active safety systems that eliminate robot-initiated collisions and minimize human-initiated ones. Aim 3 will address evaluation and refinement of safety and reliability using large-scale studies on head / eye phantoms and smaller-scale studies on human subjects.