Kinematics, Electromyography, and Human-Robot Interaction Dataset of Assisted Treadmill Walking with an Ankle Exoskeleton

Dataset Description This dataset contains kinematic and electromyographic data from three healthy participants (two males and one female) walking on a treadmill while wearing a right-ankle exoskeleton. The participants walked at three different speeds: 1.0 km/h, 1.3 km/h, and 1.6 km/h. The ankle exoskeleton used in this study is a 1-DoF module of the H2 exoskeleton (Technaid, Spain). It operates under two distinct control modes. Passive Mode: The device does not actively assist but instead follows the user's movements, reacting to minimize interactions and mimicking the natural foot trajectory. Active Mode (Position Control): The device imposes a predefined gait pattern on the ankle by controlling its position. Data Collection and Sensors The dataset includes data recorded from multiple sensors embedded in the exoskeleton, an electromyography (EMG) system, and a motion capture suit: Exoskeleton sensor data: Angular position of the exoskeleton; Reference (set to zero in passive mode; predefined trajectory in active mode); Interaction torque between the exoskeleton and the user; Motor-generated torque; Force-sensitive resistor (FSR) data from sensors embedded in the sole (heel and toe regions). EMG data was acquired using Delsys Trigno system. Surface EMG signals were recorded from two muscles in the right leg: Gastrocnemius Medialis and Tibialis Anterior. EMG data was recorded in RMS mode at 148 Hz and later resampled to 100 Hz. Maximum voluntary contraction (MVC) data for normalization is available in the MetaData.xlsx file. Kinematic Data was acquired with Xsens MTw Awinda Motion Capture Suit. It captured data at 60Hz, that was further interpolated to 100 Hz using MATLAB (linear interpolation with interp1). It includes foot contact data for both legs: RHeel, RToe, LHeel, LToe; and joint angles in the sagittal plane of all lower limb joints: Right and left ankles (RAnkle, LAnkle), right and left knees (RKnee, LKnee), and right and left hips (RHip, LHip). Additional Information Participant information is available in MetaData.xlsx. The dataset is organized into two main structures: passive and active assistance. Each condition is further divided by walking speed (V10, V13, V16). For each walking speed, there are three tables, one for each participant.数据集说明 本数据集包含3名健康受试者（2男1女）穿戴右踝外骨骼在跑步机上行走时的运动学与肌电（electromyography, EMG）数据。受试者分别以1.0 km/h、1.3 km/h及1.6 km/h三种速度行走。 本研究使用的踝部外骨骼为西班牙Technaid公司研发的H2外骨骼的单自由度（1-DoF）模块，具备两种截然不同的控制模式。 被动模式：设备不主动提供助力，而是跟随使用者的运动，通过响应以减少交互作用力，并模拟自然足部运动轨迹。 主动模式（位置控制）：设备通过控制踝关节位置，向使用者施加预设的步态模式。 数据采集与传感器 本数据集包含外骨骼内置传感器、肌电系统以及动作捕捉服采集的多类数据： 外骨骼传感器数据包括：外骨骼的角位置；参考位置（被动模式下设为0，主动模式下为预设轨迹）；外骨骼与使用者间的交互力矩；电机输出力矩；内置在鞋垫足跟与足尖区域的力敏电阻（force-sensitive resistor, FSR）数据。 肌电数据采用Delsys Trigno系统采集。受试者右侧下肢的两块肌肉：腓肠肌内侧头（Gastrocnemius Medialis）与胫骨前肌（Tibialis Anterior）的表面肌电信号被记录。肌电数据以148 Hz的频率按均方根（root mean square, RMS）模式采集，后被重采样至100 Hz。用于归一化的最大自主收缩（maximum voluntary contraction, MVC）数据可在MetaData.xlsx文件中获取。 运动学数据采用Xsens MTw Awinda动作捕捉服采集，采集频率为60 Hz，后通过MATLAB的interp1函数（线性插值）重采样至100 Hz。数据包含双侧下肢的足部接触数据：右足跟（RHeel）、右足尖（RToe）、左足跟（LHeel）、左足尖（LToe）；以及所有下肢关节在矢状面的关节角度：双侧踝关节（RAnkle、LAnkle）、双侧膝关节（RKnee、LKnee）及双侧髋关节（RHip、LHip）。 附加说明 受试者的详细信息可在MetaData.xlsx文件中查询。本数据集分为两大主要结构：被动助力模式与主动助力模式。每种模式又按行走速度分为V10、V13、V16三个组别。针对每种行走速度，均包含对应3名受试者的三份数据表。