NRI: FND: Rapid Operator Awareness via Mobile Robotics (ROAMR), Customizable Human Safety using Mobile and Wearable Co-Robots

在建筑工地、灾区和其他快速变化的环境中，人类操作员面临着与移动机械和下落部件碰撞而受伤的风险。碰撞的风险会降低人类的表现，因为它会迫使人类缓慢地工作，同时跟踪多种情况。该项目旨在探索协作机器人团队如何通过检测人类可能没有意识到的潜在碰撞，提醒他们注意危险，并帮助他们远离危险，从而提高人类的安全性。该项目将开发通过人类穿戴的外骨骼向人类操作员传达潜在威胁的方法，该外骨骼还可以帮助人们尽快转移到安全地带。该方法将增强人类在非结构化、户外和变化环境中的意识、性能和安全性，在这些环境中，动态物体会造成迫在眉睫的、未被注意的碰撞威胁。具体来说，该项目将开发一种新的安全架构，其中移动系统可以快速检测碰撞并规划安全响应，可穿戴外骨骼将情况传达给人类，外骨骼有助于实现安全响应。该项目将侧重于1)为可穿戴协作机器人开发和描述一种物理“语言”，以提供基于视觉显示、音频引导、振动触觉和外骨骼驱动器物理信号的态势感知；2)研究可穿戴传感如何利用机器学习推断人类运动，通过考虑自动避障计划、人类运动策略、关节运动学、地面接触力和肌肉激活的测量来预测人类行为；3)研究可穿戴外骨骼如何利用有针对性的外骨骼扭矩来增强推进力，通过减小肢体的表观惯性来提高响应速度，从而实时加速安全响应。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Human operators in construction sites, disaster zones, and other rapidly changing environments risk injury from collisions with moving machinery and falling components. The risk of collisions can reduce human performance by forcing them to work very slowly and keep track of multiple situations simultaneously. This project aims to explore how teams of co-robots can improve human safety by detecting potential collisions that the humans may not be aware of, alerting them to the danger, and helping them move away from the danger. The project will develop methods for communicating potential threats to human operators via a human-worn exoskeleton, which can also assist the people in moving to safety as rapidly as possible.The approach will augment human awareness, performance, and safety in unstructured, outdoor, and varying environments, where dynamic objects pose threats of imminent, unnoticed collision. Specifically, the project will develop a new safety architecture where mobile systems rapidly detect collisions and plan a safe response, a wearable exoskeleton communicates the situation to the human, and the exoskeleton helps to achieve a safe response. The project will focus on 1) developing and characterizing a physical "language" for wearable co-robots to provide situational awareness based on visual displays, audio guidance, vibro-tactile sensations, and physical signals from the exoskeleton actuators; 2) investigating how wearable sensing can infer human motions using machine learning to predict human behavior by accounting for automated obstacle avoidance plans, human motion strategies, and measurements of joint kinematics, ground contact forces, and muscle activation; and 3) researching how a wearable exoskeleton can accelerate safe responses in real-time by using targeted exoskeleton torques to enhance propulsive forces and increase speed of response by reducing the apparent inertia of the limbs.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.