The difficulty of estimating joint kinematics remains a critical barrier toward widespread use of inertial measurement units in biomechanics. Traditional sensor-fusion filters are largely reliant on magnetometer readings, which may be disturbed in uncontrolled environments. Careful sensor-to-segment alignment and calibration strategies are also necessary, which may burden users and lead to further error in uncontrolled settings. We introduce a new framework that combines deep learning and top-down optimization to accurately predict lower extremity joint angles directly from inertial data, without relying on magnetometer readings. We trained deep neural networks on a large set of synthetic inertial data derived from a clinical marker-based motion-tracking database of hundreds of subjects. We used data augmentation techniques and an automated calibration approach to reduce error due to variability in sensor placement and limb alignment. On left-out subjects, lower extremity kinematics could be predicted with a mean (± STD) root mean squared error of less than 1.27° (± 0.38°) in flexion/extension, less than 2.52° (± 0.98°) in ad/abduction, and less than 3.34° (± 1.02°) internal/external rotation, across walking and running trials. Errors decreased exponentially with the amount of training data, confirming the need for large datasets when training deep neural networks. While this framework remains to be validated with true inertial measurement unit (IMU) data, the results presented here are a promising advance toward convenient estimation of gait kinematics in natural environments. Progress in this direction could enable large-scale studies and offer an unprecedented view into disease progression, patient recovery, and sports biomechanics.

估计关节运动学的难度仍然是惯性测量单元在生物力学中广泛使用的关键障碍。传统的传感器融合滤波器很大程度上依赖于磁力计读数，在不受控制的环境中，磁力计读数可能会受到干扰。传感器到段的对准和校准策略也必须谨慎，这可能会给用户带来负担，并导致不受控制的设置中出现进一步的错误。我们引入了一个新的框架，该框架结合了深度学习和自上而下的优化功能，可直接从惯性数据中准确预测下肢关节角度，而无需依赖磁力计读数。我们在大量的合成惯性数据集上训练了深度神经网络，这些惯性数据集来自基于临床标志物的数百个受试者的运动跟踪数据库。我们使用数据增强技术和自动校准方法来减少由于传感器放置和肢体对齐的可变性而引起的错误。对于遗留受试者，可以预测下肢运动学，屈曲/伸展度的平均（±STD）均方根误差小于1.27°（±0.38°），ad /小于2.52°（±0.98°）在步行和跑步试验中，被绑架，并且内部/外部旋转角度小于3.34°（±1.02°）。误差随着训练数据量的增加呈指数下降，这证实了在训练深度神经网络时需要大型数据集。尽管此框架仍有待使用真实的惯性测量单位（IMU）数据进行验证，但此处介绍的结果是在自然环境中方便地估计步态运动学的有希望的进展。