一种基于足底压力的踝关节角度估计方法OA
Ankle angle estimation based on a distributed plantar pressure sensing system
针对目前行走状态下关节角度估计方法存在的问题,本文提出了一种基于足底压力的踝关节角度估计方法.首先,本文开发了一套分布式足底压力采集系统,该系统可以实时采集人体双足各8 个点位的分布式足底压力数据.随后,本文使用动作捕捉系统与足底压力采集系统进行了足底压力与关节角度同步采集实验,并构建了足底压力-关节角度数据集.基于该数据集,本文使用多项式特征方法对输入的足底压力进行了预处理,随后使用机器学习中的岭回归算法对系数矩阵进行求解,构建了足底压力数据与关节角度数据之间的映射模型,该模型在测试集上表现出的均方根误差为 4.04 °.该结果表明本文模型可以成为一种良好的踝关节角度估计模型.
To solve the existing problems in joint angle estimation methods during walking states,this paper proposes a method for estimating the ankle joint angle based on plantar pressure.Firstly,a distributed plantar pressure acqui-sition system is developed in this paper,which can collect real-time distributed plantar pressure data from eight points on both human feet.Subsequently,this paper conducts synchronized experiments for plantar pressure and joint angle collection using a motion capture system and the plantar pressure acquisition system,and constructs a plantar pressure-joint angle dataset.Based on this dataset,the paper preprocesses the input plantar pressure using the polynomial feature method,and then uses the ridge regression algorithm in machine learning for fitting to solve the coefficient matrix,constructing a mapping model between plantar pressure data and joint angle data.The model demonstrates a root mean square error of 4.04 °on the test set,indicating that this method can be a good ankle joint angle estimation model.
吕博文;蔡世波;都明宇;张科文;李刚
浙江工业大学机械工程学院 杭州 310023浙江工业大学机械工程学院 杭州 310023浙江工业大学机械工程学院 杭州 310023浙江工业大学机械工程学院 杭州 310023浙江工业大学机械工程学院 杭州 310023
踝关节角度足底压力多项式特征岭回归
ankle joint angleplantar pressurepolynomial featuresridge regression
《高技术通讯》 2026 (7)
759-766,8
国家重点研发计划(2022YFC3601702)资助项目.
评论