Collaborative Research: CISE-MSI: DP: CNS: Multi-Modal User-Centric Mobility Scooter Driving Safety Assessment System

电动滑板车是一种经济实惠且受欢迎的辅助移动技术，对于促进行动不便/残疾人的社会参与至关重要，从而提高他们的生活质量。然而，代步车驾驶安全的挑战也是显而易见的，代步车事故可能会造成严重的伤害，甚至可能导致死亡。为了应对这一挑战，本项目建立了一个基于多源数据的代步车驾驶员驾驶安全评估系统。该系统可在家中或临床环境中使用，以协助医生和康复专家。这项工作生成了一个移动滑板车安全评估指标框架和系统，提供了新的可用，准确和可解释的技术，以推进目标人群的康复和远程医疗过程，并通过基于深度学习的数据智能进行授权。跨学科的项目团队形成计算机科学和运动机能学和健康促进之间的协同作用，从两个西班牙裔服务机构，即，加州州立理工大学波莫纳分校和德克萨斯大学圣安东尼奥分校与卡萨科利纳医院和医疗保健中心以及加州州立大学北岭成就中心合作。这项工作向参与机构中代表性不足的少数民族学生介绍了现实世界的研究问题，研究能力扩展的示范可以传播到其他少数民族服务机构。该项目包括以下研究重点：1）建立一个纵向移动踏板车驾驶数据集，该数据集收集自具有各种条件的移动残疾/挑战人群（例如，神经系统疾病，年龄相关疾病和功能障碍）; 2）开发移动滑板车驾驶安全指标框架; 3）设计鲁棒的多模态深度学习算法，用于准确的安全评估，采用姿势估计和多模态数据的时间序列分析; 4）使用可定制的硬件和软件在用户的移动踏板车上实施所提出的驾驶安全评估系统，该奖项反映了NSF的法定使命，并被认为是值得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的支持。

Mobility scooters are an affordable and popular assistive mobility technology that is critical in facilitating the social participation of people with mobility challenges/disabilities, thereby improving their quality of life. However, the challenges of mobility scooter driving safety are evident, and mobility scooter accidents can cause serious injuries, with the potential consequence even being death.To tackle this challenge, this project builds a driving safety assessment system for mobility scooter drivers based on multi-sourced data. This system can be used at home or in clinical settings to assist doctors and rehabilitation experts. This work generates a mobility scooter safety assessment metrics framework and system that provides new usable, accurate and interpretable techniques to advance rehabilitation and tele-medicine processes for the target group of population, empowered by deep-learning-based data intelligence. The interdisciplinary project team forms synergy between Computer Science and Kinesiology and Health Promotion from two Hispanic Serving Institutions, i.e., California State Polytechnic University, Pomona and University of Texas San Antonio, in collaboration with Casa Colina Hospital and Centers for Healthcare and California State University Northridge Center of Achievement. This work introduces real-world research problems to underrepresented minority students in participating institutions, and the demonstration of research capacity expansion can be disseminated to other Minority Serving Institutions. This project consists of the following research thrusts: 1) Building a longitudinal mobility scooter driving dataset collected from the mobility-disabled/challenged population with various conditions (e.g., neurologic disorders, age-related diseases, and dysfunctions); 2) Developing a mobility scooter driving safety metrics framework; 3) Designing robust, multi-modal deep learning algorithms for accurate safety assessment, employing pose-estimation and time series analysis of multi-modal data; 4) Implementing the proposed driving safety assessment system on users’ mobility scooters using customizable hardware and software, and conducting extensive testing and evaluation of the proposed system.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.