Wearable sensing for human motor control research and mobility assistive devices: Capturing environmental features

出于近300万加拿大行动不便者面临的挑战，该提案寻求资金，以推进可穿戴式传感方法，以自动获取物理环境的关键环境信息，并调查集成移动辅助设备(MAD)情景感知控制的效用。虽然新出现的MAD(如外骨骼、机器人步行器、先进假肢)促进行动能力的潜在好处已受到极大关注，但它们的吸收和实用仍然有限。MAD有效性的一个常见障碍是在克服现实世界环境障碍方面的局限性，并将跌倒风险降至最低。拟议研究的基本假设是，结合环境特征存在(和接近)的数据将提高目标MAD和可穿戴传感应用在促进安全移动方面的有效性。最初的研究目标是开发与适应性机动性相关的环境特征分类，并评估候选传感成像方式以满足技术和可行性要求。在多模式数据收集和注释活动之后，将生成机器视觉算法，以自动检测分类法定义的移动性特征。为了检验实用性，这些新功能将在样例MAD和可穿戴式传感应用中进行集成和测试。实现拟议的研究目标将产生描述移动性环境特征的基础知识，开发新的可穿戴传感功能，并检查腿部和轮式MAD的上下文感知自动控制的效用。虽然这项研究的主要动机是促进MAD，但拟议的研究结果有望通过提供新的工具来从生态学角度研究自由生活现象，从而影响人类运动控制研究。这一基础研究计划的成果，如新的传感器套件、算法和数据集，为检查和测量人与环境的流动性互动提供了新的视角。通过提供增强的背景感知分析，这些基本发展预计将吸引工业和卫生研究合作伙伴进行应用研究。这项研究将通过商业化机会、工具开发和高素质人员培训(HQP)，使设计辅助技术的加拿大企业(如Bionik实验室、Able Innovation、Trexo Robotics)和其他与移动相关的市场(如移动机器人、增强现实、康复服务)受益。通过投资培训14名HQP(10名本科生、2名硕士和2名博士生)，该研究计划将直接支持加拿大在该领域的专业知识的发展。

Motivated by the challenges of  nearly 3 million Canadians living with a mobility disability, this proposal seeks funding to advance wearable sensing methods to automatically acquire key environmental information of the physical environment, and to investigate the utility of integrating context-aware control for mobility assistive devices (mAD). While the potential benefits of emerging mAD (e.g., exoskeletons, robotic walkers, advanced prosthetics) to facilitate mobility have received significant attention, their uptake and utility remain limited. A common barrier to effectiveness of mAD are limitations in overcoming real-world environmental obstacles with minimal risk for falls. The underlying hypothesis of the proposed research theorizes that incorporating data on presence (and proximity) of environmental features will advance the effectiveness of target mAD and wearable sensing applications in facilitating safe mobility. Initial research objectives are aimed at developing a taxonomy of environmental features related to adaptive mobility, and evaluating candidate sensing imaging modalities to meet technical and feasibility requirements. Following multimodal data collection and annotation activities, machine vision algorithms will be generated to automatically detect mobility features, as defined by the taxonomy. To examine utility, these new capabilities will be integrated and tested in exemplar mAD and wearable sensing applications. Achieving the proposed research objectives will generate foundational knowledge in characterizing environmental features for mobility, develop new wearable sensing capabilities, and examine utility of context-aware automated control of legged and wheeled mAD. While a major motivation of the research is to advance mAD, the outcomes of the proposed research are expected to impact human motor control research by providing new tools to investigate free-living phenomena from an ecological perspective. The deliverables of this basic research program, such as new sensor kits, algorithms, and datasets, afford a new lens to examine and measure human-environment mobility interactions. By providing enhanced context-aware analyses, these fundamental developments are expected to attract industry and health research partners for applied research. This research will benefit Canadian businesses designing assistive technologies (e.g., Bionik Laboratories, Able Innovations, Trexo Robotics) and other markets related to mobility (e.g., mobile robotics, augmented reality, rehabilitation services) through commercialization opportunities, tool development, and training of highly qualified personnel (HQP). By investing in the training of 14 HQP (10 undergraduate, 2 Master's, and 2 Doctoral students), this research program will directly support the development of Canadian expertise in this field.