Recent Machine Learning Progress in Lower Limb Running Biomechanics With Wearable Technology: A Systematic Review

Rethinking running biomechanics: a critical review of ground reaction forces, tibial bone loading, and the role of wearable sensors

This study presents a comprehensive review of the correlation between tibial acceleration (TA), ground reaction forces (GRF), and tibial bone loading, emphasizing the critical role of wearable sensor technology in accurately measuring these biomechanical forces in the context of running. This systematic review and meta-analysis searched various electronic databases (PubMed, SPORTDiscus, Scopus, IEEE Xplore, and ScienceDirect) to identify relevant studies. It critically evaluates existing research on GRF and tibial acceleration (TA) as indicators of running-related injuries, revealing mixed findings. Intriguingly, recent empirical data indicate only a marginal link between GRF, TA, and tibial bone stress, thus challenging the conventional understanding in this field. The study also highlights the limitations of current biomechanical models and methodologies, proposing a paradigm shift towards more holistic and integrated approaches. The study underscores wearable sensors' potential, enhanced by machine learning, in transforming the monitoring, prevention, and rehabilitation of running-related injuries.

Integrating an LSTM framework for predicting ankle joint biomechanics during gait using inertial sensors

The ankle joint plays a crucial role in gait, facilitating the articulation of the lower limb, maintaining foot-ground contact, balancing the body, and transmitting the center of gravity. This study aimed to implement long short-term memory (LSTM) networks for predicting ankle joint angles, torques, and contact forces using inertial measurement unit (IMU) sensors. Twenty-five healthy participants were recruited. Two IMU sensors were attached to the foot dorsum and the vertical axis of the distal anteromedial tibia in the right lower limb to record acceleration and angular velocity during running. We proposed a LSTM-MLP (multilayer perceptron) model for training time-series data from IMU sensors and predicting ankle joint biomechanics. The model underwent validation and testing using a custom nested k-fold cross-validation process. The average values of the coefficient of determination (R2), mean absolute error (MAE), and mean squared error (MSE) for ankle dorsiflexion joint and moment, subtalar inversion joint and moment, and ankle joint contact forces were 0.89 ± 0.04, 0.75 ± 1.04, and 2.96 ± 4.96 for walking, and 0.87 ± 0.07, 0.88 ± 1.26, and 4.1 ± 7.17 for running, respectively. This study demonstrates that IMU sensors, combined with LSTM neural networks, are invaluable tools for evaluating ankle joint biomechanics in lower limb pathological diagnosis and rehabilitation, offering a cost-effective and versatile alternative to traditional experimental settings.

Predicting Tissue Loads in Running from Inertial Measurement Units

BACKGROUND

Runners have high incidence of repetitive load injuries, and habitual runners often use smartwatches with embedded IMU sensors to track their performance and training. If accelerometer information from such IMUs can provide information about individual tissue loads, then running watches may be used to prevent injuries.

METHODS

We investigate a combined physics-based simulation and data-based method. A total of 285 running trials from 76 real runners are subjected to physics-based simulation to recover forces in the Achilles tendon and patella ligament, and the collected data are used to train and test a data-based model using elastic net and gradient boosting methods.

RESULTS

Correlations of up to 0.95 and 0.71 for the patella ligament and Achilles tendon forces, respectively, are obtained, but no single best predictive algorithm can be identified.

CONCLUSIONS

Prediction of tissues loads based on body-mounted IMUs appears promising but requires further investigation before deployment as a general option for users of running watches to reduce running-related injuries.

Location Matters-Can a Smart Golf Club Detect Where the Club Face Hits the Ball?

In golf, the location of the impact, where the clubhead hits the ball, is of imperative nature for a successful ballflight. Direct feedback to the athlete where he/she hits the ball could improve a practice session. Currently, this information can be measured via, e.g., dual laser technology; however, this is a stationary and external method. A mobile measurement method would give athletes the freedom to gain the information of the impact location without the limitation to be stationary. Therefore, the aim of this study was to investigate whether it is possible to detect the impact location via a motion sensor mounted on the shaft of the golf club. To answer the question, an experiment was carried out. Within the experiment data were gathered from one athlete performing 282 golf swings with an 7 iron. The impact location was recorded and labeled during each swing with a Trackman providing the classes for a neural network. Simultaneously, the motion of the golf club was gathered with an IMU from the Noraxon Ultium Motion Series. In the next step, a neural network was designed and trained to estimate the impact location class based on the motion data. Based on the motion data, a classification accuracy of 93.8% could be achieved with a ResNet architecture.

A Systematic Review of Gait Analysis in the Context of Multimodal Sensing Fusion and AI

BACKGROUND

Neurological diseases are a leading cause of disability and mortality. Gait, or human walking, is a significant predictor of quality of life, morbidity, and mortality. Gait patterns and other kinematic, kinetic, and balance gait features are accurate and powerful diagnostic and prognostic tools.

OBJECTIVE

This review article focuses on the applicability of gait analysis using fusion techniques and artificial intelligence (AI) models. The aim is to examine the significance of mixing several types of wearable and non-wearable sensor data and the impact of this combination on the performance of AI models.

METHOD

In this systematic review, 66 studies using more than two modalities to record and analyze gait were identified. 40 studies incorporated multiple gait analysis modalities without the use of artificial intelligence to extract gait features such as kinematic, kinetic, margin of stability, temporal, and spatial gait parameters, as well as cerebral activity. Similarly, 26 studies analyzed gait data using multimodal fusion sensors and AI algorithms.

RESULTS

The research summarized here demonstrates that the quality of gait analysis and the effectiveness of AI models can both benefit from the integration of data from many sensors. Meanwhile, the utilization of EMG signals in fusion data is especially advantageous.

CONCLUSION

The findings of this review suggest that a smart, portable, wearable-based gait and balance assessment system can be developed using multimodal sensing of the most cutting-edge, clinically relevant tools and technology available. The information presented in this article may serve as a vital springboard for such development.

A Data-Driven Approach for Fatigue Detection during Running Using Pedobarographic Measurements

Background

Detecting fatigue at the early stages of a run could aid training programs in making adjustments, thereby reducing the heightened risk of injuries from overuse. The study aimed to investigate the effects of running fatigue on plantar force distribution in the dominant and nondominant feet of amateur runners.

Methods

Thirty amateur runners were recruited for this study. Bilateral time-series plantar forces were employed to facilitate automatic fatigue gait recognition using convolutional neural network (CNN) and CNN-based long short-term memory network (ConvLSTM) models. Plantar force data collection was conducted both before and after a running-induced fatigue protocol using a FootScan force plate. The Keras library in Python 3.8.8 was used to train and tune deep learning models.

Results

The results demonstrated that more mid-forefoot and heel force occurs during bilateral plantar and less midfoot fore force occurs in the dominant limb after fatigue (p < 0.001). The time of peak forces was significantly shortened at the midfoot and sum region of the nondominant foot, while it was delayed at the hallux region of the dominant foot (p < 0.001). In addition, the ConvLSTM model showed higher performance (Accuracy = 0.867, Sensitivity = 0.874, and Specificity = 0.859) in detecting fatigue gait than CNN (Accuracy = 0.800, Sensitivity = 0.874, and Specificity = 0.718).

Conclusions

The findings of this study could offer empirical data for evaluating risk factors linked to overuse injuries in a single limb, as well as facilitate early detection of fatigued gait.

The Application of Wearable Sensors and Machine Learning Algorithms in Rehabilitation Training: A Systematic Review

The integration of wearable sensor technology and machine learning algorithms has significantly transformed the field of intelligent medical rehabilitation. These innovative technologies enable the collection of valuable movement, muscle, or nerve data during the rehabilitation process, empowering medical professionals to evaluate patient recovery and predict disease development more efficiently. This systematic review aims to study the application of wearable sensor technology and machine learning algorithms in different disease rehabilitation training programs, obtain the best sensors and algorithms that meet different disease rehabilitation conditions, and provide ideas for future research and development. A total of 1490 studies were retrieved from two databases, the Web of Science and IEEE Xplore, and finally 32 articles were selected. In this review, the selected papers employ different wearable sensors and machine learning algorithms to address different disease rehabilitation problems. Our analysis focuses on the types of wearable sensors employed, the application of machine learning algorithms, and the approach to rehabilitation training for different medical conditions. It summarizes the usage of different sensors and compares different machine learning algorithms. It can be observed that the combination of these two technologies can optimize the disease rehabilitation process and provide more possibilities for future home rehabilitation scenarios. Finally, the present limitations and suggestions for future developments are presented in the study.

A Wearable Biofeedback Device for Monitoring Tibial Load During Partial Weight-Bearing Walking

Patients with tibial fractures are usually advised to follow a partial weight-bearing gait rehabilitation program after surgery to promote bone healing and lower limb functional recovery. Currently, the biofeedback devices used for gait rehabilitation training in fracture patients use ground reaction force (GRF) as the indicator of tibial load. However, an increasing body of research has shown that monitoring GRF alone cannot objectively reflect the load on the lower limb bones during human movement. In this study, a novel biofeedback system was developed utilizing inertial measurement units and custom instrumented insoles. Based on the data collected from experiments, a hybrid approach combining a physics-based model and neural network architectures was used to predict tibial force. Compared to the traditional physics-based algorithm, the physical guided neural networks method showed better predictive performance. The study also found that regardless of the type of weight-bearing walking, the peak tibial force was significantly higher than the peak tibial GRF, and the time at which the peak tibial compression force occurs may not be consistent with the time at which the peak vertical GRF occurs. This further supports the idea that during gait rehabilitation training for patients with tibial fractures, monitoring and providing feedback on the actual tibial force rather than just the GRF is necessary. The developed device is a non-invasive and reliable portable device that can provide audio feedback, providing a viable solution for gait rehabilitation training outside laboratory and helping to optimize patients' rehabilitation treatment strategies.

Performance of multiple neural networks in predicting lower limb joint moments using wearable sensors

Joint moment measurements represent an objective biomechemical parameter in joint health assessment. Inverse dynamics based on 3D motion capture data is the current 'gold standard' to estimate joint moments. Recently, machine learning combined with data measured by wearable technologies such electromyography (EMG), inertial measurement units (IMU), and electrogoniometers (GON) has been used to enable fast, easy, and low-cost measurements of joint moments. This study investigates the ability of various deep neural networks to predict lower limb joint moments merely from IMU sensors. The performance of five different deep neural networks (InceptionTimePlus, eXplainable convolutional neural network (XCM), XCMplus, Recurrent neural network (RNNplus), and Time Series Transformer (TSTPlus)) were tested to predict hip, knee, ankle, and subtalar moments using acceleration and gyroscope measurements of four IMU sensors at the trunk, thigh, shank, and foot. Multiple locomotion modes were considered including level-ground walking, treadmill walking, stair ascent, stair descent, ramp ascent, and ramp descent. We show that XCM can accurately predict lower limb joint moments using data of only four IMUs with RMSE of 0.046 ± 0.013 Nm/kg compared to 0.064 ± 0.003 Nm/kg on average for the other architectures. We found that hip, knee, and ankle joint moments predictions had a comparable RMSE with an average of 0.069 Nm/kg, while subtalar joint moments had the lowest RMSE of 0.033 Nm/kg. The real-time feedback that can be derived from the proposed method can be highly valuable for sports scientists and physiotherapists to gain insights into biomechanics, technique, and form to develop personalized training and rehabilitation programs.

Foot Pronation Prediction with Inertial Sensors during Running: A Preliminary Application of Data-Driven Approaches

Abnormal foot postures may affect foot movement and joint loading during locomotion. Investigating foot posture alternation during running could contribute to injury prevention and foot mechanism study. This study aimed to develop feature-based and deep learning algorithms to predict foot pronation during prolonged running. Thirty-two recreational runners have been recruited for this study. Nine-axial inertial sensors were attached to the right dorsum of the foot and the vertical axis of the distal anteromedial tibia. This study employed feature-based machine learning algorithms, including support vector machine (SVM), extreme gradient boosting (XGBoost), random forest, and deep learning, i.e., one-dimensional convolutional neural networks (CNN1D), to predict foot pronation. A custom nested k-fold cross-validation was designed for hyper-parameter tuning and validating the model's performance. The XGBoot classifier achieved the best accuracy using acceleration and angular velocity data from the foot dorsum as input. Accuracy and the area under curve (AUC) were 74.7 ± 5.2% and 0.82 ± 0.07 for the subject-independent model and 98 ± 0.4% and 0.99 ± 0 for the record-wise method. The test accuracy of the CNN1D model with sensor data at the foot dorsum was 74 ± 3.8% for the subject-wise approach with an AUC of 0.8 ± 0.05. This study found that these algorithms, specifically for the CNN1D and XGBoost model with inertial sensor data collected from the foot dorsum, could be implemented into wearable devices, such as a smartwatch, for monitoring a runner's foot pronation during long-distance running. It has the potential for running shoe matching and reducing or preventing foot posture-induced injuries.

Estimation of horizontal running power using foot-worn inertial measurement units

Feedback of power during running is a promising tool for training and determining pacing strategies. However, current power estimation methods show low validity and are not customized for running on different slopes. To address this issue, we developed three machine-learning models to estimate peak horizontal power for level, uphill, and downhill running using gait spatiotemporal parameters, accelerometer, and gyroscope signals extracted from foot-worn IMUs. The prediction was compared to reference horizontal power obtained during running on a treadmill with an embedded force plate. For each model, we trained an elastic net and a neural network and validated it with a dataset of 34 active adults across a range of speeds and slopes. For the uphill and level running, the concentric phase of the gait cycle was considered, and the neural network model led to the lowest error (median ± interquartile range) of 1.7% ± 12.5% and 3.2% ± 13.4%, respectively. The eccentric phase was considered relevant for downhill running, wherein the elastic net model provided the lowest error of 1.8% ± 14.1%. Results showed a similar performance across a range of different speed/slope running conditions. The findings highlighted the potential of using interpretable biomechanical features in machine learning models for the estimating horizontal power. The simplicity of the models makes them suitable for implementation on embedded systems with limited processing and energy storage capacity. The proposed method meets the requirements for applications needing accurate near real-time feedback and complements existing gait analysis algorithms based on foot-worn IMUs.

Machine Learning for Optical Motion Capture-Driven Musculoskeletal Modelling from Inertial Motion Capture Data

Marker-based Optical Motion Capture (OMC) systems and associated musculoskeletal (MSK) modelling predictions offer non-invasively obtainable insights into muscle and joint loading at an in vivo level, aiding clinical decision-making. However, an OMC system is lab-based, expensive, and requires a line of sight. Inertial Motion Capture (IMC) techniques are widely-used alternatives, which are portable, user-friendly, and relatively low-cost, although with lesser accuracy. Irrespective of the choice of motion capture technique, one typically uses an MSK model to obtain the kinematic and kinetic outputs, which is a computationally expensive tool increasingly well approximated by machine learning (ML) methods. Here, an ML approach is presented that maps experimentally recorded IMC input data to the human upper-extremity MSK model outputs computed from ('gold standard') OMC input data. Essentially, this proof-of-concept study aims to predict higher-quality MSK outputs from the much easier-to-obtain IMC data. We use OMC and IMC data simultaneously collected for the same subjects to train different ML architectures that predict OMC-driven MSK outputs from IMC measurements. In particular, we employed various neural network (NN) architectures, such as Feed-Forward Neural Networks (FFNNs) and Recurrent Neural Networks (RNNs) (vanilla, Long Short-Term Memory, and Gated Recurrent Unit) and a comprehensive search for the best-fit model in the hyperparameters space in both subject-exposed (SE) as well as subject-naive (SN) settings. We observed a comparable performance for both FFNN and RNN models, which have a high degree of agreement (ravg,SE,FFNN=0.90±0.19, ravg,SE,RNN=0.89±0.17, ravg,SN,FFNN=0.84±0.23, and ravg,SN,RNN=0.78±0.23) with the desired OMC-driven MSK estimates for held-out test data. The findings demonstrate that mapping IMC inputs to OMC-driven MSK outputs using ML models could be instrumental in transitioning MSK modelling from 'lab to field'.

A Wearable-Sensor System with AI Technology for Real-Time Biomechanical Feedback Training in Hammer Throw

Developing real-time biomechanical feedback systems for in-field applications will transfer human motor skills' learning/training from subjective (experience-based) to objective (science-based). The translation will greatly improve the efficiency of human motor skills' learning and training. Such a translation is especially indispensable for the hammer-throw training which still relies on coaches' experience/observation and has not seen a new world record since 1986. Therefore, we developed a wearable wireless sensor system combining with artificial intelligence for real-time biomechanical feedback training in hammer throw. A framework was devised for developing such practical wearable systems. A printed circuit board was designed to miniaturize the size of the wearable device, where an Arduino microcontroller, an XBee wireless communication module, an embedded load cell and two micro inertial measurement units (IMUs) could be inserted/connected onto the board. The load cell was for measuring the wire tension, while the two IMUs were for determining the vertical displacements of the wrists and the hip. After calibration, the device returned a mean relative error of 0.87% for the load cell and the accuracy of 6% for the IMUs. Further, two deep neural network models were built to estimate selected joint angles of upper and lower limbs related to limb coordination based on the IMUs' measurements. The estimation errors for both models were within an acceptable range, i.e., approximately ±12° and ±4°, respectively, demonstrating strong correlation existed between the limb coordination and the IMUs' measurements. The results of the current study suggest a remarkable novelty: the difficulty-to-measure human motor skills, especially in those sports involving high speed and complex motor skills, can be tracked by wearable sensors with neglect movement constraints to the athletes. Therefore, the application of artificial intelligence in a wearable system has shown great potential of establishing real-time biomechanical feedback training in various sports. To our best knowledge, this is the first practical research of combing wearables and machine learning to provide biomechanical feedback in hammer throw. Hopefully, more wearable biomechanical feedback systems integrating artificial intelligence would be developed in the future.

Predicting Coordination Variability of Selected Lower Extremity Couplings during a Cutting Movement: An Investigation of Deep Neural Networks with the LSTM Structure

There are still few portable methods for monitoring lower limb joint coordination during the cutting movements (CM). This study aims to obtain the relevant motion biomechanical parameters of the lower limb joints at 90°, 135°, and 180° CM by collecting IMU data of the human lower limbs, and utilizing the Long Short-Term Memory (LSTM) deep neural-network framework to predict the coordination variability of selected lower extremity couplings at the three CM directions. There was a significant (p < 0.001) difference between the three couplings during the swing, especially at 90° vs the other directions. At 135° and 180°, t13-he coordination variability of couplings was significantly greater than at 90° (p < 0.001). It is important to note that the coordination variability of Hip rotation/Knee flexion-extension was significantly higher at 90° than at 180° (p < 0.001). By the LSTM, the CM coordination variability for 90° (CMC = 0.99063, RMSE = 0.02358), 135° (CMC = 0.99018, RMSE = 0.02465) and 180° (CMC = 0.99485, RMSE = 0.01771) were accurately predicted. The predictive model could be used as a reliable tool for predicting the coordination variability of different CM directions in patients or athletes and real-world open scenarios using inertial sensors.