K2MUSE: A human lower limb multimodal dataset under diverse conditions for facilitating rehabilitation robotics

TL;DR

K2MUSE provides the first open, large-scale multimodal lower-limb dataset combining kinematics, kinetics, amplitude-mode ultrasound, and sEMG across diverse walking conditions and nonideal scenarios. The dataset supports data-driven control and intention recognition for rehabilitation robotics, backed by rigorous synchronization, processing, and validation against public benchmarks. With detailed data organization and open-source tooling, K2MUSE enables robust biomechanical analysis and development of end-to-end control strategies for exoskeletons and assistive devices. Future work aims to broaden tasks, include clinical populations, and test in real-world environments to advance embodied intelligent robotics.

Abstract

The natural interaction and control performance of lower limb rehabilitation robots are closely linked to biomechanical information from various human locomotion activities. Multidimensional human motion data significantly deepen the understanding of the complex mechanisms governing neuromuscular alterations, thereby facilitating the development and application of rehabilitation robots in multifaceted real-world environments. However, currently available lower limb datasets are inadequate for supplying the essential multimodal data and large-scale gait samples necessary for effective data-driven approaches, and they neglect the significant effects of acquisition interference in real applications.To fill this gap, we present the K2MUSE dataset, which includes a comprehensive collection of multimodal data, comprising kinematic, kinetic, amplitude-mode ultrasound (AUS), and surface electromyography (sEMG) measurements. The proposed dataset includes lower limb multimodal data from 30 able-bodied participants walking under different inclines (0$^\circ$, $\pm$5$^\circ$, and $\pm$10$^\circ$), various speeds (0.5 m/s, 1.0 m/s, and 1.5 m/s), and different nonideal acquisition conditions (muscle fatigue, electrode shifts, and inter-day differences). The kinematic and ground reaction force data were collected via a Vicon motion capture system and an instrumented treadmill with embedded force plates, whereas the sEMG and AUS data were synchronously recorded for thirteen muscles on the bilateral lower limbs. This dataset offers a new resource for designing control frameworks for rehabilitation robots and conducting biomechanical analyses of lower limb locomotion. The dataset is available at https://k2muse.github.io/.

Figures (14)

• Figure 1: Experiments were conducted in the biomechanics laboratory. (a) The experimental scene shows a participant equipped with all the devices: a motion capture system, a treadmill with embedded force plates, an sEMG system, and an AUS device. (b) Participants performed experiments on a treadmill under diverse conditions, including different ascending and descending ramps and walking speeds. In the fatigue-induced experiment, the participants alternated between squatting and walking continuously.
• Figure 2: Summary of publicly available biomechanics datasets for lower limb locomotion.
• Figure 3: The modified marker set for motion capture. The markers were attached to the lower limbs in a generally symmetrical arrangement, with the markers on the left side shown. Markers marked with '*' were defined according to the Plug-in Gait lower body model, which implements the Conventional Gait Model. Detailed marker placement instructions for the Plug-in Gait lower body model can be found in the Plug-in Gait Reference Guide.
• Figure 4: The sEMG sensors and AUS transducers were attached to the participants' skin. The channels of different instrumentation are highlighted in different colors for easy distinction. The symbol '#' corresponds to the channel numbers of different devices.
• Figure 5: Experimental setups for muscle fatigue and electrode shifts. (a) Fatigue-induced lower limb exercises, including dorsiflexion/plantar flexion and squats, with a barbell held in the hands. (b) Experimental setup simulating electrode shifts, where different electrode pairs and transducers correspond to initial positions and four shift directions.