Implementing dual quaternions as a basis for sensor fusion

实现双四元数作为传感器融合的基础

• 批准号: 463677-2014
• 负责人: Blohm, Gunnar
• 金额: $ 1.82万
• 依托单位: Queen's University
• 依托单位国家: 加拿大
• 项目类别: Engage Grants Program
• 财政年份: 2014
• 资助国家: 加拿大
• 起止时间: 2014-01-01 至 2015-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=558538
• 关键词: Implementing; dual; quaternions; basis; sensor

Sensor fusion is the process by which (potentially redundant) information from different sensory sources is combined. In pose estimation, data can arise from optical imaging or inertial sensors, among others. Since the data is measured in different reference frames, it is not trivial to combine (fuse) them within a rigid body skeletal model to obtain instantaneous pose. Here, we propose a novel sensor fusion approach using Bayesian principles, Kalman filters and dual quaternions to obtain optimal estimates of pose from different sensors. This approach will be developed in collaboration with HAS-Motion and implemented in their Visual3D software. We will also record walking/running data to test our new algorithm against existing procedures. This will provide critically needed sensor fusion technology for biomechanical assessment and rehabilitation.

传感器融合是将来自不同感官来源的（可能冗余的）信息组合在一起的过程。在姿态估计中，数据可以来自光学成像或惯性传感器等。由于数据是在不同的参考系中测量的，因此将它们合并（融合）在刚体骨骼模型中以获得瞬时姿态是很重要的。在这里，我们提出了一种新的传感器融合方法，利用贝叶斯原理、卡尔曼滤波和对偶四元数来获得不同传感器的姿态最优估计。这种方法将与HAS-Motion合作开发，并在他们的Visual3D软件中实现。我们还将记录步行/跑步数据，以对照现有程序测试我们的新算法。这将为生物力学评估和康复提供急需的传感器融合技术。