18 December 2023 in IEEE Transactions on Instrumentation and Measurement
Motivated by the need for underwater exoskeleton control, swimming phase monitoring is critical to adjust robotic assistance to human motions. However, harsh underwater environments bring in challenges of poor robustness and serious disturbances to the few available sensing methods leading to coarse phase segmentation. Given skeletal muscle activation for limb joint motions, this paper proposes a muscle-joint synergy method to combine muscle mechanics captured by a novel capacitive sensor with joint kinematics by traditional inertial measurements. Development of the wearable sensing system is provided with details on design, fabrication and assembly. The muscle-joint synergy model is rigorously formulated and experimentally validated with breaststroke data for the first time. Immediate applications are illustrated with discrete mode identification and continuous phase monitoring, justifying the robustness of the proposed method to various artificial intelligent algorithms and different tested subjects. This work is anticipated to be valuable to unveil the mechanism of human natural motions for human-centered robotics, sport science and rehabilitation engineering in the future.
5 March 2025 in IEEE Transactions on Instrumentation and Measurement
Swimming pattern recognition plays an important role in underwater wearable robot control and competitive sports training. However, serious disturbances of harsh underwater environments result in few available sensing methods, and existing single–mode sensing inherently limits the efficiency of swimming monitoring. This paper combines muscle deformations measured by flexible capacitive sensors with limb joint rotations by inertial measurement units. Heatmaps and polar coordinate systems are used to exhibit differences of cyclic features in four strokes. The t–SNE method is used to convert high–dimensional fusion features to a 2D plane for visualization. Using 1D–CNN classifier for swimming pattern recognition at three kicking frequencies, effectiveness of the method was experimentally validated by nine subjects with the average accuracy of 97.9%, while that in single sensing modality is 82.6% with muscle deformations and 85.9% with limb rotations. To the best of our knowledge, this paper is the first to use multimodal fusion for swimming motion monitoring, and is expected to enhance underwater wearable sensing devices for human-centered robotics and sports science.
30 May 2024 in IEEE Sensors Journal
Accurate estimation of gait phases is crucial in controlling wearable robots designed to restore or enhance human walking capabilities. Asynchronous human-machine motions can lead to reduced efficiency, increased interaction loads, and potential harm to the human body. Variations in walking speeds and terrain ramps pose additional challenges to continuous phase estimation. This paper presents a novel synergy method based on muscle deformations measured by flexible wearable sensors to dynamically monitor gait phases while adapting to changes in speed and terrain ramp. Polar coordinate diagrams are used to elucidate the cyclic features of joint angles and muscle deformations during walking. The consistency of muscle deformation synergy is studied to build a phase estimation model, with parameters being determined by neural network algorithms. The average root-mean-square error for the estimated phase is less than 4.4% within a speed range of 3 to 6 km/h and less than 8.8% in a ramp range of 0 to 10°. Muscle deformation presents a viable alternative in scenarios involving changes in speed, outperforming joint kinematics in gait phase estimation amidst variations in ramp inclinations. The proposed unified muscle synergy model, adaptable across a wide spectrum of walking speeds and ramp inclinations, holds promise for enhancing human-machine coordination for robotic assistance and rehabilitation.
17 June 2025 in IEEE/ASME Transactions on Mechatronics
Tri-axial reaction forces (TRF), which support body mass and provide dynamic propulsion, are critical to improve motion stability, gait transitions, and energy efficiency in legged motions of both humans and robots. The accurate and robust sensing of TRF presents significant challenges due to the cyclic and high-impact loads. Existing commercial sensors with rigid encapsulation have a heavy weight and poor wearability, while flexible sensors with conformal capability are susceptible to damage under cyclic impacts due to complex structure and stress concentration between flexible and hard interfaces. This article proposes a flexible magnetoelastic sensing method that couples three-dimensional elastomer deformations with magnetic flux intensities to capture tri-axial load distributions. The locally-stiffened design mitigates stress concentration at the interfaces among soft elastomer, flexible substrate, and rigid chips. The plantar load sensing array is developed to measure loads up to 116, 113, and 433 kPa in the x-, y-, and z-axes, respectively, with durability validated by the 150 000-cycle compression test and 9-month clinical gait monitoring. The proposed method is illustrated with an immediate application to gait pattern identification, and it is anticipated to contribute to tri-axial force sensing in various scenarios, including but not limited to rehabilitation gait monitoring, legged robot navigation, and human–robot interaction.
4 March 2025 in IEEE Transactions on Instrumentation and Measurement
Gait pattern transition is important for wearable robot control to assist human walking on various terrains. The mismatch between human and machine motions during gait transitions poses challenges to human-robot interaction, leading to undesirable robotic assistance and high impacts or even damages to the human body. This paper proposes a gait transition monitoring method based on wearable sensing of muscle deformation synergy during multi-pattern natural motions. The cyclic features of different gaits, as depicted in the polar coordinate diagrams, facilitate the use of neural networks for pattern identification and phase estimation during gait transition. The proposed method has demonstrated adaptability to multiple natural gaits, as validated by the consistent transition phase (standard deviation ± 3.8%) and duration (standard deviation ± 3.5%) of nine subjects observed by four transition modes across normal walking and stair ascent / decent motions. This research is expected to provide valuable guidelines to develop robotic assistance control and improve human-machine coordination during gait transition.
18 June 2025 in IEEE Transactions on Instrumentation and Measurement
Accurate recognition of human motion modes and continuous phases is crucial to exoskeleton control to provide proper assistance. However, harsh underwater environments severely restrict the study on swimming motion monitoring, where existing researches either focus on a single swimming mode or discrete phases, limiting underwater exoskeleton control. To address this limitation, this paper develops a two-stage network consisting of one mode classifier (1st stage) and four phase regressors (2nd stage), where muscle deformation features are used instead of traditional joint kinematics. Swimming tests are conducted on nine subjects with four modes at three frequencies. Effectiveness of the proposed method is justified by mode identification accuracy of 99.72% and phase estimation error of 3.92%, where the error is 52.89% smaller than that in the traditional time-based estimation method. This paper is the first to simultaneously recognize the swimming mode and the continuous phase, which is valuable to adapt the smooth exoskeleton assistance to harsh underwater environment and multimodal motion scenarios.