Inseung Kang

This record contains the code and data supporting the preprint, “Fall risk-aware adaptation explains suboptimal locomotor performance.” The included files are: locomotor_learning_model-CODE.zip, which contains the MATLAB and Python implementations of the locomotor adaptation model, simulation code, analysis routines, figure-generation scripts, and validation utilities. locomotor_learning_model-DATA.zip, which contains the experimental data used in the study. Together, these materials support reproduction of the computational analyses and manuscript figures described in the preprint. Software requirements, environment setup, and usage instructions are provided in the accompanying README documentation. This record is being shared for editorial and peer-review purposes and will be made publicly available upon publication.

This record contains the code and data supporting the preprint, “Fall risk-aware adaptation explains suboptimal locomotor performance.” The included files are: locomotor_learning_model-CODE.zip, which contains the MATLAB and Python implementations of the locomotor adaptation model, simulation code, analysis routines, figure-generation scripts, and validation utilities. locomotor_learning_model-DATA.zip, which contains the experimental data used in the study. Together, these materials support reproduction of the computational analyses and manuscript figures described in the preprint. Software requirements, environment setup, and usage instructions are provided in the accompanying README documentation. This record is being shared for editorial and peer-review purposes and will be made publicly available upon publication.

Designing generalizable control policies for lower-limb exoskeletons remains fundamentally constrained by exhaustive data collection or iterative optimization procedures, which limit accessibility to clinical populations. To address this challenge, we introduce a device-agnostic framework that combines physiologically plausible musculoskeletal simulation with reinforcement learning to enable scalable personalized exoskeleton assistance for both able-bodied and clinical populations. Our control policies not only generate physiologically plausible locomotion dynamics but also capture clinically observed compensatory strategies under targeted muscular deficits, providing a unified computational model of both healthy and pathological gait. Without task-specific tuning, the resulting exoskeleton control policies produce assistive torque profiles at the hip and ankle that align with state-of-the-art profiles validated in human experiments, while consistently reducing metabolic cost across walking speeds. For simulated impaired-gait models, the learned control policies yield asymmetric, deficit-specific exoskeleton assistance that improves both energetic efficiency and bilateral kinematic symmetry without explicit prescription of the target gait pattern. These results demonstrate that physiologically plausible musculoskeletal simulation via reinforcement learning can serve as a scalable foundation for personalized exoskeleton control across both able-bodied and clinical populations, eliminating the need for extensive physical trials.

Designing generalizable control policies for lower-limb exoskeletons remains fundamentally constrained by exhaustive data collection or iterative optimization procedures, which limit accessibility to clinical populations. To address this challenge, we introduce a device-agnostic framework that combines physiologically plausible musculoskeletal simulation with reinforcement learning to enable scalable personalized exoskeleton assistance for both able-bodied and clinical populations. Our control policies not only generate physiologically plausible locomotion dynamics but also capture clinically observed compensatory strategies under targeted muscular deficits, providing a unified computational model of both healthy and pathological gait. Without task-specific tuning, the resulting exoskeleton control policies produce assistive torque profiles at the hip and ankle that align with state-of-the-art profiles validated in human experiments, while consistently reducing metabolic cost across walking speeds. For simulated impaired-gait models, the learned control policies yield asymmetric, deficit-specific exoskeleton assistance that improves both energetic efficiency and bilateral kinematic symmetry without explicit prescription of the target gait pattern. These results demonstrate that physiologically plausible musculoskeletal simulation via reinforcement learning can serve as a scalable foundation for personalized exoskeleton control across both able-bodied and clinical populations, eliminating the need for extensive physical trials.

A bstract Human locomotion requires balancing multiple biological objectives, such as metabolic energy efficiency, stability, and symmetry. While models based on optimization successfully predict how humans walk in familiar settings, they fail to explain why individuals adopt inefficient movement patterns in novel environments, even after extensive practice. Here, we show that such suboptimality in a novel environment arises from a fundamental prioritization of safety. We find that individuals do not simply fail to reach an optimal solution; instead, they navigate an environment-dependent risk landscape by mitigating the statistical probability of falling. We find that this risk-averse strategy is explained by adjusting internal learning parameters: specifically, the learning rate and the tradeoff between metabolic cost and symmetry, in a manner that lowers fall risk. To quantify this process, we developed an ‘inverse adaptation’ modeling framework; this approach works backwards from locomotor performance data to mathematically infer the underlying internal learning parameters and how they vary with fall risk. Our analysis reveals that the observed motor performance is explained by a global probabilistic fall risk rather than a local step-based measure of instability. Ultimately, these findings reveal that fall risk-aware adaptation explains suboptimal locomotor behavior, providing a new data-driven framework to understand the drivers of motor performance.

Developing exoskeleton controllers that generalize across diverse locomotor conditions typically requires extensive motion-capture data and biomechanical labeling, limiting scalability beyond instrumented laboratory settings. Here, we present a physics-based neuromusculoskeletal learning framework that trains a hip-exoskeleton control policy entirely in simulation, without motion-capture demonstrations, and deploys it on hardware via policy distillation. A reinforcement learning teacher policy is trained using a muscle-synergy action prior over a wide range of walking speeds and slopes through a two-stage curriculum, enabling direct comparison between assisted and no-exoskeleton conditions. In simulation, exoskeleton assistance reduces mean muscle activation by up to 3.4% and mean positive joint power by up to 7.0% on level ground and ramp ascent, with benefits increasing systematically with walking speed. On hardware, the assistance profiles learned in simulation are preserved across matched speed-slope conditions (r: 0.82, RMSE: 0.03 Nm/kg), providing quantitative evidence of sim-to-real transfer without additional hardware tuning. These results demonstrate that physics-based neuromusculoskeletal simulation can serve as a practical and scalable foundation for exoskeleton controller development, substantially reducing experimental burden during the design phase.

Portable active back support devices (BSDs) offer tunable assistance but are often bulky and heavy, limiting their usability. In contrast, passive BSDs are lightweight and compact but lack the ability to adapt their assistance to different back movements. We present a soft, lightweight, and compact BSD that combines a variable-stiffness passive element and an active element (an artificial muscle) in parallel. The device provides tunable assistance through discrete changes in stiffness values and active force levels. We validate the device's tuning capabilities through bench testing and on-body characterization. Further, we use the device's tuning capabilities to provide weight-adaptive object lifting and lowering assistance. We detect the weight handled by the user based on forearm force myography and upper-back inertial measurement unit data. Furthermore, electromyography analyses in five participants performing symmetric object lifting and lowering tasks showed reductions in back extensor activity. Preliminary results in one participant also indicated reduced muscle activity during asymmetric lifting.

Portable active back support devices (BSDs) offer tunable assistance but are often bulky and heavy, limiting their usability. In contrast, passive BSDs are lightweight and compact but lack the ability to adapt their assistance to different back movements. We present a soft, lightweight, and compact BSD that combines a variable-stiffness passive element and an active element (an artificial muscle) in parallel. The device provides tunable assistance through discrete changes in stiffness values and active force levels. We validate the device's tuning capabilities through bench testing and on-body characterization. Further, we use the device's tuning capabilities to provide weight-adaptive object lifting and lowering assistance. We detect the weight handled by the user based on forearm force myography and upper-back inertial measurement unit data. Furthermore, electromyography analyses in five participants performing symmetric object lifting and lowering tasks showed reductions in back extensor activity. Preliminary results in one participant also indicated reduced muscle activity during asymmetric lifting.

Developing exoskeleton controllers that generalize across diverse locomotor conditions typically requires extensive motion-capture data and biomechanical labeling, limiting scalability beyond instrumented laboratory settings. Here, we present a physics-based neuromusculoskeletal learning framework that trains a hip-exoskeleton control policy entirely in simulation, without motion-capture demonstrations, and deploys it on hardware via policy distillation. A reinforcement learning teacher policy is trained using a muscle-synergy action prior over a wide range of walking speeds and slopes through a two-stage curriculum, enabling direct comparison between assisted and no-exoskeleton conditions. In simulation, exoskeleton assistance reduces mean muscle activation by up to 3.4% and mean positive joint power by up to 7.0% on level ground and ramp ascent, with benefits increasing systematically with walking speed. On hardware, the assistance profiles learned in simulation are preserved across matched speed-slope conditions (r: 0.82, RMSE: 0.03 Nm/kg), providing quantitative evidence of sim-to-real transfer without additional hardware tuning. These results demonstrate that physics-based neuromusculoskeletal simulation can serve as a practical and scalable foundation for exoskeleton controller development, substantially reducing experimental burden during the design phase.

Falls are the leading cause of injury related hospitalization and mortality among older adults. Consequently, mitigating age-related declines in gait stability and reducing fall risk during walking is a critical goal for assistive devices. Lower-limb exoskeletons have the potential to support users in maintaining stability during walking. However, most exoskeleton controllers are optimized to reduce the energetic cost of walking rather than to improve stability. While some studies report stability benefits with assistance, the effects of specific parameters, such as assistance magnitude and duration, remain unexplored. To address this gap, we systematically modulated the magnitude and duration of torque provided by a bilateral hip exoskeleton during slip perturbations in eight healthy adults, quantifying stability using whole-body angular momentum (WBAM). WBAM responses were governed by a significant interaction between assistance magnitude and duration, with duration determining whether exoskeleton assistance was stabilizing or destabilizing relative to not wearing the exoskeleton device. Compared to an existing energy-optimized controller, experimentally identified stability-optimal parameters reduced WBAM range by 25.7% on average. Notably, substantial inter-subject variability was observed in the parameter combinations that minimized WBAM during perturbations. We found that optimizing exoskeleton assistance for energetic outcomes alone is insufficient for improving reactive stability during gait perturbations. Stability-focused exoskeleton control should prioritize temporal assistance parameters and include user-specific personalization. This study represents an important step toward personalized, stability-focused exoskeleton control, with direct implications for improving stability and reducing fall risk in older adults.

Falls are the leading cause of injury related hospitalization and mortality among older adults. Consequently, mitigating age-related declines in gait stability and reducing fall risk during walking is a critical goal for assistive devices. Lower-limb exoskeletons have the potential to support users in maintaining stability during walking. However, most exoskeleton controllers are optimized to reduce the energetic cost of walking rather than to improve stability. While some studies report stability benefits with assistance, the effects of specific parameters, such as assistance magnitude and duration, remain unexplored. To address this gap, we systematically modulated the magnitude and duration of torque provided by a bilateral hip exoskeleton during slip perturbations in eight healthy adults, quantifying stability using whole-body angular momentum (WBAM). WBAM responses were governed by a significant interaction between assistance magnitude and duration, with duration determining whether exoskeleton assistance was stabilizing or destabilizing relative to not wearing the exoskeleton device. Compared to an existing energy-optimized controller, experimentally identified stability-optimal parameters reduced WBAM range by 25.7% on average. Notably, substantial inter-subject variability was observed in the parameter combinations that minimized WBAM during perturbations. We found that optimizing exoskeleton assistance for energetic outcomes alone is insufficient for improving reactive stability during gait perturbations. Stability-focused exoskeleton control should prioritize temporal assistance parameters and include user-specific personalization. This study represents an important step toward personalized, stability-focused exoskeleton control, with direct implications for improving stability and reducing fall risk in older adults.

Mobility impairments from aging, injury, or medical conditions limit independence and social participation. Conventional assistive devices lack adaptability in complex environments. Recent wearable technologies integrating neural sensing, electronics, and co-design offer personalized, responsive mobility support. This perspective focuses on advances in wearable sensing and multimodal fusion for intent recognition, environmental interaction, and adaptive control in exoskeletons, prosthetics, smart wheelchairs, and navigation systems. Emphasizing human-in-the-loop and cognitive-sensorimotor integration, it outlines emerging trends and challenges, promoting intelligent, user-centered solutions to restore function and enhance autonomy, accessibility, and inclusion for individuals with mobility impairments.

Virtual models of human gait, or digital twins, offer a promising solution for studying mobility without the need for labor-intensive data collection. However, challenges such as the sim-to-real gap and limited adaptability to diverse walking conditions persist. To address these, we developed and validated a framework to create a skeletal humanoid agent capable of adapting to varying walking speeds while maintaining biomechanically realistic motions. The framework combines a synthetic data generator, which produces biomechanically plausible gait kinematics from open-source biomechanics data, and a training system that uses adversarial imitation learning to train the agent's walking policy. We conducted comprehensive analyses comparing the agent's kinematics, synthetic data, and the original biomechanics dataset. The agent achieved a root mean square error of $5.24 \pm 0.09$ degrees at varying speeds compared to ground-truth kinematics data, demonstrating its adaptability. This work represents a significant step toward developing a digital twin of human locomotion, with potential applications in biomechanics research, exoskeleton design, and rehabilitation.

Accurate estimation of a user's biological joint moment from wearable sensor data is vital for improving exoskeleton control during real-world locomotor tasks. However, most state-of-the-art methods rely on deep learning techniques that necessitate extensive in-lab data collection, posing challenges in acquiring sufficient data to develop robust models. To address this challenge, we introduce a locomotor task set optimization strategy designed to identify a minimal, yet representative, set of tasks that preserves model performance while significantly reducing the data collection burden. In this optimization, we performed a cluster analysis on the dimensionally reduced biomechanical features of various cyclic and non-cyclic tasks. We identified the minimal viable clusters (i.e., tasks) to train a neural network for estimating hip joint moments and evaluated its performance. Our cross-validation analysis across subjects showed that the optimized task set-based model achieved a root mean squared error of $0.29 \pm 0.06 \text{Nm} / \text{kg}$. This performance was significantly better than using only cyclic tasks ($\mathbf{p}<0.05$) and was comparable to using the full set of tasks. Our results demonstrate the ability to maintain model accuracy while significantly reducing the cost associated with data collection and model training. This highlights the potential for future exoskeleton designers to leverage this strategy to minimize the data requirements for deep learning-based models in wearable robot control.

Falls during daily ambulation activities are a leading cause of injury in older adults due to delayed physiological responses to disturbances of balance. Lower-limb exoskeletons have the potential to mitigate fall incidents by detecting and reacting to perturbations before the user. Although commonly used, the standard metric for perturbation detection, whole-body angular momentum, is poorly suited for exoskeleton applications due to computational delays and additional tunings. To address this, we developed a novel ground perturbation detector using lower-limb kinematic states during locomotion. To identify perturbations, we tracked deviations in the kinematic states from their nominal steady-state trajectories. Using a data-driven approach, we optimized our detector with an open-source ground perturbation biomechanics dataset. A nine-subject cross-validation demonstrated that our model distinguished perturbed from unperturbed gait cycles with 95.5% accuracy and only a delay of 33.1% within the gait cycle, outperforming the benchmark by 49.4% in detection accuracy. The results of our study offer exciting promise for our detector and its potential utility to enhance the controllability of robotic assistive exoskeletons.

Accurate lower-limb joint kinematics estimation is critical for patient monitoring, rehabilitation, and exoskeleton control. While previous studies have employed wearable sensor-based deep learning (DL) models for estimating joint kinematics, these methods often require extensive new datasets to adapt to unseen gait patterns. Meanwhile, researchers in computer vision have advanced human pose estimation models, which are easy to deploy and capable of real-time inference. However, such models are infeasible in scenarios where cameras cannot be used. To address these limitations, we propose a computer vision-based DL adaptation framework for realtime joint kinematics estimation. This framework requires only a small dataset (i.e., 1-2 gait cycles) and does not rely on professional motion capture setups. Using transfer learning, we adapted our temporal convolutional network (TCN) to stiff knee gait data, allowing the model to reduce root mean square error by 9.7 % and 19.9 % compared to a TCN trained on only able-bodied and stiff knee dataset, respectively. Our framework demonstrated a potential for smartphone camera-trained DL model to estimate real-time joint kinematics across novel users in clinical populations with applications in wearable robots.

Fall incidents due to slips are some of the most common causes of injuries for industry workers and older adults, motivating research to assist balance recovery following slips. To assist balance recovery during a slip, a detection algorithm that can work with an assistive device, such as an exoskeleton, needs to be able to detect slips rapidly after onset, which remains a critical gap in the field. Here, we compared the ability of linear discriminant analysis (LDA), extreme gradient boosting (XGBoost), and convolutional neural networks (CNN) to detect slip using only native sensors on a hip exoskeleton. We trained and evaluated user-independent models on early-stance (ES) and late-stance (LS) slips of various magnitudes collected through treadmill-based slips. All models, except LDA with LS slips, detected slips with >90% accuracy. Overall, he best model was XGBoost, with its fastest results achieving average detection times and median accuracies of 155.06 ms at 96.25% for ES slips and 228.88 ms at 93.75% for LS slips, while also achieving 100% sensitivity at 195.64 ms (ES) and 266.24 ms (LS). Our results indicate a promising direction for further research into designing a generalizable model for balance recovery during slip perturbations using robotic hip exoskeletons.

This study introduces a deep learning framework for estimating lower-limb joint kinematics using inertial measurement units (IMUs). While deep learning methods avoid sensor drift, extensive calibration, and complex setup procedures, they require substantial data. To meet this demand, we leveraged an open-source dataset to develop and evaluate three training approaches. The first involved training a model exclusively on data from a single user, resulting in high accuracy for that individual only. The second approach trained a model on data from multiple users to generalize across individuals; however, demonstrated lower accuracy due to variations in gait patterns across users. The third approach added transfer learning to the second, improving estimation accuracy for new users through fine-tuning with a small portion of their data. This model overcame the limitations of the previous methods' dependency on extensive data collection, and achieved comparable performance to inverse kinematics, making it an effective solution for diverse populations. Additionally, our analysis on IMU combinations suggests that IMUs placed on the femur and calcaneus are the best for most cases. This framework not only reduces the need for extensive data collection but also enhances personalized gait analysis, enabling more efficient and accessible applications in both clinical assessments and real-world environments for broader use.

OBJECTIVE: This study applies human-in-the-loop optimization to identify optimal hip exoskeleton assistance patterns for stair climbing. METHODS: Ten participants underwent optimization to individualize hip flexion and extension assistance, followed by a validation comparing optimized assistance (OPT) to biological hip moment-based assistance (BIO), no assistance (No-Assist), and no exoskeleton (No-Exo) conditions. RESULTS: OPT reduced metabolic cost by 4.5% compared to No-Exo, 11.44% compared to No-Assist, and 5.02% compared to BIO, demonstrating the effectiveness of the optimization approach. Statistical analysis revealed distinct characteristics in optimal assistance timing and magnitude that deviated systematically from biological hip moment patterns. Compared to BIO, OPT exhibited later peak flexion timing (76.4 $\pm$ 3.7% vs 65.0%), shorter flexion duration (29.2 $\pm$ 3.6% vs 40.0%), later peak extension timing (26.7 $\pm$ 3.8% vs 20.0% of gait cycle), and higher peak flexion magnitude (11.1 $\pm$ 1.5 Nm vs 10.0 Nm). While individual optimal assistance profiles varied across participants, comparison between individually optimized parameters and the best subject-independent parameters identified through post-hoc analysis showed consistency. On average, metabolic rate convergence was achieved after 18 iterations, while most exoskeleton control parameters did not reach our convergence criteria within 20 iterations. CONCLUSION: These findings demonstrate that human-in-the-loop optimization can effectively identify task-specific assistance patterns for stair climbing, while the consistency between individual and subject-independent parameters suggests the potential for developing generalized assistance strategies. The systematic differences between optimized and biological moment-based assistance underscore the fundamental distinctions between biological torque-based control and optimal control for exoskeleton assistance.

Robotic exoskeletons can transform mobility for individuals with lower-limb disabilities. However, their widespread adoption is limited by controller degradation caused by varying gait dynamics across different users and environments. Here, we propose an online adaptation framework that leverages real-time data streams to continuously update the user state estimator model. This approach allows the exoskeleton to learn the user-specific gait patterns, effectively customizing the model for each new user. Additionally, we demonstrate a sensor signal transformation technique that enables model transfer across different exoskeleton hardware (from a research-grade exoskeleton to a commercial device). With less than one minute of adaptation, our framework improved gait phase estimation, which directly affects assistance timing, by 40.9% for able-bodied subjects and 65.9% for stroke survivors (p<0.05), and reduced torque profile error by 32.7% compared to the baseline model (p<0.05). Furthermore, in a pilot test, we applied our adaptation framework with human-in-the-loop optimization for control tuning. In a single stroke survivor, this approach led to a 21.8% increase in walking speed and a 6.5% reduction in metabolic cost compared to walking without exoskeleton. While preliminary, these results suggest the potential for personalized exoskeleton assistance in clinical populations.

Accurate estimation of a user's biological joint moment from wearable sensor data is vital for improving exoskeleton control during real-world locomotor tasks. However, most state-of-the-art methods rely on deep learning techniques that necessitate extensive in-lab data collection, posing challenges in acquiring sufficient data to develop robust models. To address this challenge, we introduce a locomotor task set optimization strategy designed to identify a minimal, yet representative, set of tasks that preserves model performance while significantly reducing the data collection burden. In this optimization, we performed a cluster analysis on the dimensionally reduced biomechanical features of various cyclic and non-cyclic tasks. We identified the minimal viable clusters (i.e., tasks) to train a neural network for estimating hip joint moments and evaluated its performance. Our cross-validation analysis across subjects showed that the optimized task set-based model achieved a root mean squared error of 0.30$\pm$0.05 Nm/kg. This performance was significantly better than using only cyclic tasks (p&lt;0.05) and was comparable to using the full set of tasks. Our results demonstrate the ability to maintain model accuracy while significantly reducing the cost associated with data collection and model training. This highlights the potential for future exoskeleton designers to leverage this strategy to minimize the data requirements for deep learning-based models in wearable robot control.

Falls during daily ambulation activities are a leading cause of injury in older adults due to delayed physiological responses to disturbances of balance. Lower-limb exoskeletons have the potential to mitigate fall incidents by detecting and reacting to perturbations before the user. Although commonly used, the standard metric for perturbation detection, whole-body angular momentum, is poorly suited for exoskeleton applications due to computational delays and additional tunings. To address this, we developed a novel ground perturbation detector using lower-limb kinematic states during locomotion. To identify perturbations, we tracked deviations in the kinematic states from their nominal steady-state trajectories. Using a data-driven approach, we further optimized our detector with an open-source ground perturbation biomechanics dataset. A pilot experimental validation with five able-bodied subjects demonstrated that our model distinguished perturbed from unperturbed gait cycles with 98.8% accuracy and only a delay of 23.1% within the gait cycle, outperforming the benchmark by 47.7% in detection accuracy. The results of our study offer exciting promise for our detector and its potential utility to enhance the controllability of robotic assistive exoskeletons.

Virtual models of human gait, or digital twins, offer a promising solution for studying mobility without the need for labor-intensive data collection. However, challenges such as the sim-to-real gap and limited adaptability to diverse walking conditions persist. To address these, we developed and validated a framework to create a skeletal humanoid agent capable of adapting to varying walking speeds while maintaining biomechanically realistic motions. The framework combines a synthetic data generator, which produces biomechanically plausible gait kinematics from open-source biomechanics data, and a training system that uses adversarial imitation learning to train the agent's walking policy. We conducted comprehensive analyses comparing the agent's kinematics, synthetic data, and the original biomechanics dataset. The agent achieved a root mean square error of 5.24 +- 0.09 degrees at varying speeds compared to ground-truth kinematics data, demonstrating its adaptability. This work represents a significant step toward developing a digital twin of human locomotion, with potential applications in biomechanics research, exoskeleton design, and rehabilitation.

Accurate lower-limb joint kinematic estimation is critical for applications such as patient monitoring, rehabilitation, and exoskeleton control. While previous studies have employed wearable sensor-based deep learning (DL) models for estimating joint kinematics, these methods often require extensive new datasets to adapt to unseen gait patterns. Meanwhile, researchers in computer vision have advanced human pose estimation models, which are easy to deploy and capable of real-time inference. However, such models are infeasible in scenarios where cameras cannot be used. To address these limitations, we propose a computer vision-based DL adaptation framework for real-time joint kinematic estimation. This framework requires only a small dataset (i.e., 1-2 gait cycles) and does not depend on professional motion capture setups. Using transfer learning, we adapted our temporal convolutional network (TCN) to stiff knee gait data, allowing the model to further reduce root mean square error by 9.7% and 19.9% compared to a TCN trained on only able-bodied and stiff knee datasets, respectively. Our framework demonstrates a potential for smartphone camera-trained DL models to estimate real-time joint kinematics across novel users in clinical populations with applications in wearable robots.

Robotic lower-limb exoskeletons can augment human mobility, but current systems require extensive, context-specific considerations, limiting their real-world viability. Here, we present a unified exoskeleton control framework that autonomously adapts assistance on the basis of instantaneous user joint moment estimates from a temporal convolutional network (TCN). When deployed on our hip exoskeleton, the TCN achieved an average root mean square error of 0.142 newton-meters per kilogram across 35 ambulatory conditions without any user-specific calibration. Further, the unified controller significantly reduced user metabolic cost and lower-limb positive work during level-ground and incline walking compared with walking without wearing the exoskeleton. This advancement bridges the gap between in-lab exoskeleton technology and real-world human ambulation, making exoskeleton control technology viable for a broad community.

Proportional myoelectric controller (PMC) has been one of the most common assistance strategies for robotic exoskeletons due to its ability to modulate assistance level directly based on the user's muscle activation. However, existing PMC strategies (static or user-adaptive) scale torque linearly with muscle activation level and fail to address complex and non-linear mapping between muscle activation and joint torque. Furthermore, previously presented adaptive PMC strategies do not allow for environmental changes (such as changes in ground slopes) and modulate the system's assistance level over many steps. In this work, we designed a novel user- and environment-adaptive PMC for a knee exoskeleton that modulates the peak assistance level based on the slope level during locomotion. We recruited nine able-bodied adults to test and compare the effects of three different PMC strategies (static, user-adaptive, and user- and environment-adaptive) on the user's metabolic cost and the knee extensor muscle activation level during load-carriage walking (6.8 kg) in three inclination settings (0°, 4.5°, and 8.5°). The results showed that only the user- and environment-adaptive PMC was effective in significantly reducing user's metabolic cost (5.8% reduction) and the knee extensor muscle activation (19% reduction) during 8.5° incline walking compared to the unpowered condition while other PMCs did not have as large of an effect. This control framework highlights the viability of implementing an assistance paradigm that can dynamically adjust to the user's biological demand, allowing for a more personalized assistance paradigm.