Optimal Sensor Placement for Estimation of Center of Plantar Pressure Based on the Improved Genetic Algorithms
Optimal Sensor Placement for Estimation of Center of Plantar Pressure Based on the Improved Genetic Algorithms
复制标题
DOI：
10.1109/jsen.2021.3125021
复制
发表时间：
2021-12-15
影响因子：
4.3
通讯作者：
Xie, Longhan
中科院分区：
文献类型：
--
作者：
Xian, Xiaoming;Zhou, Zikang;Xie, Longhan
足底压力分析可用于临床诊断、运动指导和日常监测。在实际使用中，COP轨迹是动态分析的一个重要参数，在户外和日常监控中通常是通过鞋内系统获得的。因此，在低成本传感鞋垫中设计传感器布局以获得准确的COP估计是一个关键问题。提出了一种基于足底压力分布数据的传感器布局方法--改进遗传算法，以减小轨迹估计误差和增加信息量为目标，将传感器布局问题抽象为多目标组合优化问题。通过优化迭代，确定了一组优化的传感器配置，并应用于实际应用。6名受试者穿上最优放置鞋垫，与测量平台提供的COP轨迹比较，平均绝对误差分别为3.81 mm(内侧向)和8.61 mm(前后向)。与以往的结果相比，本文提出的方法提供了更准确的COP估计，提高了9.7%。这项研究为选择足底压力传感器的位置提供了新的指导方针，并将智能优化算法融入到新的方法中，以提高可穿戴设备分析的准确性。
Plantar pressure analysis can be used for clinical diagnosis, exercise guidance and daily monitoring. In actual use, the CoP trajectory is an important parameter for dynamic analysis, which is generally obtained with an in-shoe system in outdoor and daily monitoring. Therefore, it is a critical issue to design the sensor placement to obtain accurate CoP estimation in a low-cost sensing insole. In this paper, a new sensor placement method, an improved genetic algorithm, was proposed, driven by a large amount of plantar pressure distribution data, with the objectives of reducing the trajectory estimation error and increasing the amount of information, and abstract the placement problem as a combinatorial optimization problem under multiple objectives. Through optimization iterations, a set of optimized sensor placements are determined and applied to practical use. Six subjects wore the optimal placement insoles and the mean absolute error was 3.81 mm (medial-lateral direction) and 8.61 mm (anterior-posterior direction) for comparison with the CoP trajectory provided by the measurement platform. Compared with previous results, the method proposed in this paper provides a more accurate CoP estimation with a 9.7% improvement. This study provides new guidelines for the selection of plantar pressure sensor placements and incorporates intelligent optimization algorithms into new ways to improve the accuracy of wearable device analysis.