Reaction Force Measurements Research Articles

Correlation Between Maximal Eccentric and Isometric Multi-Joint Lower-Extremity Strength and Vertical Jumping Performance in Young Adults

Maximal eccentric (MES) and isometric (MIS) muscle strength may enhance vertical jump performance by facilitating preloading and reducing energy loss during the eccentric (ECC) phase of the stretch-shortening cycle (SSC). However, the contributions of ECC and isometric (ISO) strength to the countermovement (CMJ) and depth jump (DJ) remain unclear due to variability in assessment methods (e.g., dynamometry, isometric mid-thigh pull) and the limited range of metrics examined in prior research. The aim of this study was to assess correlations between multi-joint lower extremity MES and MIS, obtained using a seated multi-joint isokinetic dynamometer, and 13 vertical ground reaction force (GRF) measures derived from the performance of three maximal effort DJs and CMJs. Twenty-five healthy young adults participated in this study (age = 21.9 ± 2.9 years). Pearson r correlation coefficients were used to assess the statistical significance (α = 0.05) of the relationships between absolute (N) and body mass normalized (BN) maximal strength measures and vertical jumping metrics. Moderate-to-strong positive correlations were identified between MES and MIS with broad performance metrics in CMJ and DJ, including reactive strength index (r = 0.45–0.53, p < 0.05), modified reactive strength index (r = 0.41–0.62, p < 0.05), and jump height (r = 0.59–0.75, p < 0.05). Moderate-to-strong positive correlations were also observed between MES and MIS with CON work (r = 0.58–0.71, p < 0.05) and CON peak power (r = 0.44–0.71, p < 0.05) for both the CMJ and DJ. In contrast, moderate-to-strong negative correlations were observed between MES and MIS with ECC work (r = 0.42–0.62, p < 0.005) and ECC peak power (r = 0.45–0.60, p < 0.05). These findings suggest that enhanced neuromuscular efficiency and joint stiffness in stronger musculature reduce energy absorption during the eccentric phase, minimizing mechanical deformation and preserving elastic energy for concentric propulsion. Combined, MES and MIS optimize force application, energy utilization, and control, which are crucial for maximizing jump height. These findings underscore the role of MES and MIS in influencing jumping performance across both the ECC and CON phases of the SSC. This insight is valuable for practitioners designing training programs aimed at improving vertical jumping ability.

Comparing the Ground Reaction Forces, Toe Clearances, and Stride Lengths of Young and Older Adults Using a Novel Shoe Sensor System.

In this study, we developed a lightweight shoe sensor system equipped with four high-capacity, compact triaxial force sensors and an inertial measurement unit. Remarkably, this system enabled measurements of localized three-directional ground reaction forces (GRFs) at each sensor position (heel, first and fifth metatarsal heads, and toe) and estimations of stride length and toe clearance during walking. Compared to conventional optical motion analysis systems, the developed sensor system provided relatively accurate results for stride length and minimum toe clearance. To test the performance of the system, 15 older and 8 young adults were instructed to walk along a straight line while wearing the system. The results reveal that compared to the young adults, older adults exhibited lower localized GRF contributions from the heel and greater localized GRF contribution from the toe and fifth metatarsal locations. Furthermore, the older adults exhibited greater variability in their stride length and smaller toe clearance with greater variability compared to the young adults. These results underscore the effectiveness of the proposed gait analysis system in distinguishing the gait characteristics of young and older adults, potentially replacing traditional motion capture systems and force plates in gait analysis.

The Variable Stiffness Treadmill 2: Development and Validation of a Unique Tool to Investigate Locomotion on Compliant Terrains

Abstract Understanding legged locomotion in various environments is valuable for many fields, including robotics, biomechanics, rehabilitation, and motor control. Specifically, investigating legged locomotion in compliant terrains has recently been gaining interest for the robust control of legged robots over natural environments. At the same time, the importance of ground compliance has also been highlighted in poststroke gait rehabilitation. Currently, there are not many ways to investigate walking surfaces of varying stiffness. This article introduces the variable stiffness treadmill (VST) 2, an improvement of the first version of the VST, which was the first treadmill able to vary belt stiffness. In contrast to the VST 1, the device presented in this paper (VST 2) can reduce the stiffness of both belts independently, by generating vertical deflection instead of angular, while increasing the walking surface area from 0.20m2 to 0.74m2. In addition, both treadmill belts are now driven independently, while high-spatial-resolution force sensors under each belt allow for measurement of ground reaction forces and center of pressure. Through validation experiments, the VST 2 displays high accuracy and precision. The VST 2 has a stiffness range of 13kN/m to 1.5MN/m, error of less than 1%, and standard deviations of less than 2.2kN/m, demonstrating its ability to simulate low-stiffness environments reliably. The VST 2 constitutes a drastic improvement of the VST platform, a one-of-its-kind system that can improve our understanding of human and robotic gait while creating new avenues of research on biped locomotion, athletic training, and rehabilitation of gait after injury or disease.

Bilateral hip stability variation in the functional ambulation and kinetic parameters after total hip arthroplasty during leveled walking.

This study determines gender variation, comparing the significance level between men and women related to functional ambulation characteristics after hip arthroplasty. The study focuses on the broader female pelvis and how it affects the rehabilitation regimen following total hip arthroplasty. In this cross-sectional study, 20 cases of right hip arthroplasty were divided into 10 male and 10 female cases, aged 40-65years. The functional ambulation parameters (walking cadence, gait speed, stride length, and gait cycle time) were acquired from the GAITRite device, as well as kinematic values for hip frontal plane displacement and kinetic parameters for ground response force in the medial-lateral direction. An independent t-test showed a significant difference in the kinematic parameter variables for the anterior superior iliac spine, more significant trochanter displacement, and hip abduction angle between the operated and non-operated limbs for each group separately. Regarding the functional ambulation parameters, there was a significant difference in the walking cadence between the operated and non-operated limbs of both male and female groups. Moreover, the output variables of ground reaction force measures revealed significant differences between their operated and non-operated limbs. The linear regression model used was consistent with the current results, demonstrating a weak negative correlation between the abduction angle of the operated hip and gait speed for both male and female groups. Based on the findings, we draw the conclusion that improving a rehabilitated physical therapy program for the abductors of both male and female patients' operated and non-operated limbs is essential for normalizing the ground reaction force value, avoiding focus on the operated hip, and reducing the amount of time that the operated hip's abductors must perform. This involves exposing the surgically repaired limb to the risk of post-operative displacement or dislocation, particularly in female patients.

Development of an Ankle Sensor for Ground Reaction Force Measurement in Intelligent Prosthesis

In this study, a new low-cost, three-degree-of-freedom force sensor is developed to measure Ground Reaction Forces (GRFs) and to be used in the safe control of active transfemoral prosthesis. Initially, the proposed sensor was designed with the Finite Element Method (FEM). Then, the sensor's control board was developed to include an electronic circuit with its microcontroller module, four load cell amplifiers, and an orientation sensor. A test platform was also developed to conduct the sensor tests. To test the accuracy of the results obtained from the developed test platform, the same tests were also carried out with a commercial sensor and similar results were obtained, thus proving that the sensor is suitable for use in prosthetics.

Effects of the intermittent theta burst stimulation on gait, balance and lower limbs motor function in stroke: study protocol for a double-blind randomised controlled trial with multimodal neuroimaging assessments

IntroductionApproximately, 50% of stroke survivors experience impaired walking ability 6 months after conventional rehabilitation and standard care. However, compared with upper limb motor function, research on lower limbs rehabilitation through...

Benchmark and validation of state-of-the-art muscle recruitment strategies in shoulder modelling

AbstractShoulder muscle forces estimated via modelling are typically indirectly validated against measurements of glenohumeral joint reaction forces (GHJ-RF). This validation study benchmarks the outcomes of several muscle recruitment strategies against public GHJ-RF measurements. Public kinematics, electromyography, and GHJ-RF data from a selected male participant executing a 2.4 kg weight shoulder abduction task up to 92° GHJ elevation were obtained. The Delft Shoulder and Elbow Model was scaled to the participant. Muscle recruitment was solved by 1) minimising muscle activations squared (SO), 2) accounting for dynamic muscle properties (CMC) and 3) constraining muscle excitations to corresponding surface electromyography measurements (CEINMS). Moreover, the spectrum of admissible GHJ-RF in the model was determined via Markov-chain Monte Carlo stochastic sampling. The experimental GHJ-RF was compared to the resultant GHJ-RF of the different muscle recruitment strategies as well as the admissible stochastic range. From 21 to 40 degrees of humeral elevation, the experimental measurement of the GHJ-RF was outside the admissible range of the model (21 to 659% of body weight (%BW)). Joint force RMSE was between 21 (SO) and 24%BW (CEINMS). At high elevation angles, CMC (11%BW) and CEINMS (14%BW) performed better than SO (25%BW). A guide has been proposed to best select muscle recruitment strategies. At high elevation angles, CMC and CEINMS were the two most accurate methods in terms of predicted GHJ-RF. SO performed best at low elevation angles. In addition, stochastic muscle sampling highlighted the lack of consistency between the model and experimental data at low elevation angles.

Running Biomechanics and Clinical Features Among Adolescent Athletes With Lower Leg Chronic Exertional Compartment Syndrome.

To compare clinical measures between patients with chronic exertional compartment syndrome (CECS) and healthy controls and evaluate running biomechanics, physical measurements, and exertional intracompartmental (ICP) changes in adolescent athletes with lower leg CECS. Cross-sectional case-control study. Large tertiary care hospital and affiliated injury prevention center. Forty-nine adolescents with CECS (39 F, 10 M; age: 16.9 ± 0.8 years; body mass index (BMI): 23.1 ± 2.9 kg/m 2 ; symptom duration: 8 ± 12 months) were compared with 49 healthy controls (39 F, 10 M; age: 6.9 ± 0.8 years; BMI: 20.4 ± 3.7 kg/m 2 ). All participants underwent gait analyses on a force plate treadmill and clinical lower extremity strength and range of motion testing. Patients with chronic exertional compartment syndrome underwent Stryker monitor ICP testing. Symptoms, menstrual history, and ICP pressures of the patients with CECS using descriptive statistics. Mann-Whitney U and χ 2 analyses were used to compare CECS with healthy patients for demographics, clinical measures, and gait biomechanics continuous and categorical outcomes, respectively. For patients with CECS, multiple linear regressions analyses were used to assess associations between gait biomechanics, lower extremity strength and range of motion, and with ICP measures. The CECS group demonstrated higher mass-normalized peak ground reaction force measures (xBW) compared with controls (0.21 ± 0.05 xBW ( P < 0.001) and were more likely to have impact peak at initial contact ( P = 0.04). Menstrual dysfunction was independently associated with higher postexertion ICP (ß = 14.6; P = 0.02). The CECS group demonstrated increased total force magnitude and vertical impact transient peaks. In women with CECS, menstrual dysfunction was independently associated with increased postexertion ICP. These biomechanical and physiological attributes may play a role in the development of CECS.

MEMS highly sensitive and large-area force plate for total ground reaction force measurement of running ant

Terrestrial insects exhibit agile manoeuvrability while running. However, the ground reaction force (GRF) in small insects remains poorly understood owing to its size and acting force. Here, we present a force plate for the measurement of total GRF during the running motion of ants with a force on the order of several tens of μN. The proposed force plate consists of a sufficiently large plate for several step cycles and four supporting cantilevers with highly sensitive piezoresistive elements. The plate and sensor chips were fabricated separately and combined during postprocessing. Two sizes of force plates were designed and fabricated for a large ant (Camponotus japonicas) and small ant (Messor aciculatus) with force resolution less than 1 μN. The developed force plates were calibrated by applying vertical forces at 32 points on the plate. Using the fabricated force plate, we measured the total GRF over several step cycles as the ant ran along the plate. Consequently, it was suggested that the ant ran with small vibrations in the direction of gravity.

Acute effects of caffeine supplementation on kinematics and kinetics of sprinting.

We investigated the acute effects of caffeine supplementation (6 mg･kg-1 ) on 60-m sprint performance and underlying components with a step-to-step ground reaction force measurement in 13 male sprinters. After the first round sprint as a control, caffeine supplementation-induced improvement in 60-m sprint times (7.811 s at the first versus 7.648 s at the second round, 2.05%) were greater compared with the placebo condition (7.769 s at the first versus 7.768 s at the second round, 0.02%). Using average values for every four steps, in the caffeine condition, higher running speed (all six step groups), higher step frequency (5th-16th and 21st-24th step groups), shorter support time (all the step groups except for 13th-16th step) and shorter braking time (9th-24th step groups) were found. Regarding ground reaction forces variables, greater braking mean force (13th-19th step group), propulsive mean force (1st-12th and 17th-20th step groups), and effective vertical mean force (9th-12th step group) were found in the caffeine condition. For the block clearance phase at the sprint start, push-off and reaction times did not change, while higher total anteroposterior mean force, average horizontal external power, and ratio of force were found in the caffeine condition. These results indicate that, compared with placebo, acute caffeine supplementation improved sprint performance regardless of sprint sections during the entire acceleration phase from the start through increases in step frequency with decreases in support time. Moreover, acute caffeine supplementation promoted increases in the propulsive mean force, resulting in the improvement of sprint performance.

Device for Measuring Contact Reaction Forces during Animal Adhesion Landing/Takeoff from Leaf-like Compliant Substrates.

A precise measurement of animal behavior and reaction forces from their surroundings can help elucidate the fundamental principle of animal locomotion, such as landing and takeoff. Compared with stiff substrates, compliant substrates, like leaves, readily yield to loads, presenting grand challenges in measuring the reaction forces on the substrates involving compliance. To gain insight into the kinematic mechanisms and structural-functional evolution associated with arboreal animal locomotion, this study introduces an innovative device that facilitates the quantification of the reaction forces on compliant substrates, like leaves. By utilizing the stiffness-damping characteristics of servomotors and the adjustable length of a cantilever structure, the substrate compliance of the device can be accurately controlled. The substrate was further connected to a force sensor and an acceleration sensor. With the cooperation of these sensors, the measured interaction force between the animal and the compliant substrate prevented the effects of inertial force coupling. The device was calibrated under preset conditions, and its force measurement accuracy was validated, with the error between the actual measured and theoretical values being no greater than 10%. Force curves were measured, and frictional adhesion coefficients were calculated from comparative experiments on the landing/takeoff of adherent animals (tree frogs and geckos) on this device. Analysis revealed that the adhesion force limits were significantly lower than previously reported values (0.2~0.4 times those estimated in previous research). This apparatus provides mechanical evidence for elucidating structural-functional relationships exhibited by animals during locomotion and can serve as an experimental platform for optimizing the locomotion of bioinspired robots on compliant substrates.

Absorption function loss due to the history of previous ankle sprain explored by unsupervised machine learning

BackgroundAnkle sprains are common and cause persistent ankle function reduction. To biomechanically evaluate the ankle function after ankle sprains, the ground reaction force (GRF) measurement during the single-legged landing had been used. However, previous studies focused on discrete features of vertical GRF (vGRF), which largely ignored vGRF waveform features that could better identify the ankle function. PurposeTo identify how the history of ankle sprain affect the vGRF waveform during the single-legged landing with unsupervised machine learning considering the time-series information of vGRF. MethodsEighty-seven currently healthy basketball athletes (12 athletes without ankle sprain, 49 athletes with bilateral, and 26 athletes with unilateral ankle sprain more than 6 months before the test day) performed single-legged landings from a 20 centimeters (cm) high box onto the force platform. Totally 518 trials vGRF data were collected from 87 athletes of 174 ankles, including 259 ankle sprain trials (from previous sprain ankles) and 259 non-ankle sprain trials (from without sprain ankles). The first 100 milliseconds (ms) vGRF waveforms after landing were extracted. Principal component analysis (PCA) was applied to the vGRF data, selecting 8 principal components (PCs) representing 96% of the information. Based on these 8 PCs, k-means method (k = 3) clustered the 518 trials into three clusters. Chi-square test assessed significant differences (p < 0.01) in the distribution of ankle sprain and non-ankle sprain trials among clusters. FindingsThe ankle sprain trials accounted for a significantly larger percentage (63.9%) in Cluster 3, which exhibited rapidly increased impulse vGRF waveforms with larger peaks in a short time. SignificancePCA and k-means method for vGRF waveforms during single-legged landing identified that the history of previous ankle sprains caused a loss of ankle absorption ability lasting at least 6 months from an ankle sprain.

Estimating vertical ground reaction forces from plantar pressure using interpretable high-dimensional approximation

Several studies have reported methods for capturing ground reaction forces in the field. However, these forces are usually measured indirectly with non-interpretable “black box” models. Here we report an interpretable model with a function approximation algorithm. The model uses time series of plantar pressure from instrumented insoles to estimate vertical ground reaction forces. The study included data from 16 persons moving at different speeds on a two-belt treadmill equipped with force plates. The introduced regression model, based on a high-dimensional approximation, using only low-dimensional variable interactions, demonstrated its ability to estimate vertical ground reaction forces. The normalized mean square error for velocities between 3 and 9kmh-1\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$9\\,\\hbox {km}\\,\\hbox {h}^{-1}$$\\end{document} ranged from 10.6 to 24.4%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$24.4\\,\\%$$\\end{document}. The accuracy of the presented approach, which can be used to analyze and interpret the learned model, was comparable to that reported in the literature. Furthermore, the evaluation of the learned model is particularly suitable for embedded and portable systems and, after a one-time calibration measurement, allows permanent and laboratory-independent measurement of vertical ground reaction forces.

Lower Limb Joint Torque Prediction Using Long Short-Term Memory Network and Gaussian Process Regression

The accurate prediction of joint torque is required in various applications. Some traditional methods, such as the inverse dynamics model and the electromyography (EMG)-driven neuromusculoskeletal (NMS) model, depend on ground reaction force (GRF) measurements and involve complex optimization solution processes, respectively. Recently, machine learning methods have been popularly used to predict joint torque with surface electromyography (sEMG) signals and kinematic information as inputs. This study aims to predict lower limb joint torque in the sagittal plane during walking, using a long short-term memory (LSTM) model and Gaussian process regression (GPR) model, respectively, with seven characteristics extracted from the sEMG signals of five muscles and three joint angles as inputs. The majority of the normalized root mean squared error (NRMSE) values in both models are below 15%, most Pearson correlation coefficient (R) values exceed 0.85, and most decisive factor (R2) values surpass 0.75. These results indicate that the joint prediction of torque is feasible using machine learning methods with sEMG signals and joint angles as inputs.

Force plate coverings significantly affect measurement of ground reaction forces.

The purpose of this study was to carry out a material test to investigate the effect of different force plate coverings on vertical and horizontal ground reaction force and derived parameters. Four surface conditions were analysed; bare plate, vinyl, sportflex, and astroturf on a Kistler force plate. Vertical data were collected by dropping a 2 kg rigid, textured medicine ball from a low (61 cm) and a high (139 cm) height. Horizontal data were collected using a custom-built, rigid, metal pendulum device. A one-way ANOVA revealed a significant main effect of surface on peak force and rate of force development for high height, low height, and horizontal force conditions (all p<0.001), with effect sizes in the post-hoc analysis being mostly large to very large. Interestingly, sportflex yielded the highest vertical but the lowest horizontal ground reaction forces. This study showed the use of current force platform coverings had a significant effect on peak force and rate of force development measurements during a standardised testing procedure. Future research should try to obtain rate of force development values that more closely replicate aspects of human performance during standardised testing procedures. Also further investigate the effect of the different surfaces on ground reaction forces during human movement.

Robust and novel attention guided MultiResUnet model for 3D ground reaction force and moment prediction from foot kinematics

Ground reaction force and moment (GRF&M) measurements are vital for biomechanical analysis and significantly impact the clinical domain for early abnormality detection for different neurodegenerative diseases. Force platforms have become the de facto standard for measuring GRF&M signals in recent years. Although the signal quality achieved from these devices is unparalleled, they are expensive and require laboratory setup, making them unsuitable for many clinical applications. For these reasons, predicting GRF&M from cheaper and more feasible alternatives has become a topic of interest. Several works have been done on predicting GRF&M from kinematic data captured from the subject’s body with the help of motion capture cameras. The problem with these solutions is that they rely on markers placed on the whole body to capture the movements, which can be very infeasible in many practical scenarios. This paper proposes a novel deep learning-based approach to predict 3D GRF&M from only 5 markers placed on the shoe. The proposed network “Attention Guided MultiResUNet” can predict the force and moment signals accurately and reliably compared to the techniques relying on full-body markers. The proposed deep learning model is tested on two publicly available datasets containing data from 66 healthy subjects to validate the approach. The framework has achieved an average correlation coefficient of 0.96 for 3D ground reaction force prediction and 0.86 for 3D ground reaction momentum prediction in cross-dataset validation. The framework can provide a cheaper and more feasible alternative for predicting GRF&M in many practical applications.

Imaging of the Microstructural Failure Mechanism in the Human Hip.

Imaging the bone microstructure under progressively increasing loads allows for observing the microstructural failure behavior of bone. Here, we describe a protocol for obtaining a sequence of three-dimensional microstructural images of the entire proximal femur under progressively increasing deformation, causing clinically relevant fractures of the femoral neck. The protocol is demonstrated using four femora from female donors aged 66-80 years at the lower end of bone mineral density in the population (T-score range = -2.09 to -4.75). A radio-transparent compressive stage was designed for loading the specimens replicating a one-leg stance, while recording the applied load duringmicro-computed tomography (micro-CT) imaging. The field of view was 146 mm wide and 132 mm high, and the isotropic pixel size was 0.03 mm. The force increment was based on finite-element predictions of the fracture load. The compressive stage was used to apply the displacement to the specimen and enact the prescribed force increments. Sub-capital fractures due to opening and shear of the femoral neck occurred after four to five load increments. The micro-CTimages and the reaction force measurements were processed to study the bone strain and energy absorption capacity. Instability of the cortex appeared at the early loading steps. The subchondral bone in the femoral head displayed large deformations reaching 16% beforefracture, and a progressive increase in the support capacity up tofracture. The deformation energy linearly increased with the displacement up tofracture, while the stiffness decreased to near-zero values immediately beforefracture. Three-fourths of the fracture energy was taken by the specimen during the final 25% force increment. In conclusion, the protocol developed revealed a remarkable energy absorption capacity, or damage tolerance, and a synergic interaction between the cortical and trabecular bone at an advanced donor age.

Scaled dynamic loading tests on seismic isolation bearing excluding the contamination of friction and inertia forces

Conventional seismic isolation-bearing testing facilities require moving horizontal platens under high compression. However, these dynamic moving platens inevitably lead to the friction and inertial forces generated by large vertical loads and inertial forces due to the dynamic mass action of the moving platen, which eventually affects the accuracy of the measured force. To solve this problem, the authors have proposed a direct reaction force measurement system (horizontal-vertical separated type) that can exclude the contamination of friction and inertial forces. Measuring pure reaction forces without the contamination of friction forces and inertia is crucial for the reliability of test results. In this paper, prior to the construction of large-scale facility, reduced-size–mockup dynamic tests were conducted using the proposed measurement concept, and its validity and accuracy were discussed. Furthermore, this system was applied to hybrid simulation tests, and the significant improvement of the proposed measurement system was observed.

Controlling jumps through latches in small jumping robots

Small jumping robots can use springs to maximize jump performance, but they are typically not able to control the height of each jump owing to design constraints. This study explores the use of the jumper’s latch, the component that mediates the release of energy stored in the spring, as a tool for controlling jumps. A reduced-order model that considers the dynamics of the actuator pulling the latch and the effect of spring force on the latch is presented. This model is then validated using high speed video and ground reaction force measurements from a 4g jumper. Both the model and experimental results demonstrate that jump performance in small insect-inspired resource-constrained robots can be tuned to a range of outputs using latch mediation, despite starting with a fixed spring potential energy. For a fixed set of input voltages to the latch actuator, the results also show that a jumper with a larger latch radius has greater tunability. However, this greater tunability comes with a trade-off in maximum performance. Finally, we define a new metric, ‘Tunability Range,’ to capture the range of controllable jump behaviors that a jumper with a fixed spring compression can attain given a set of control inputs (i.e. latch actuation voltage) to choose from.

Characterizing Bodyweight-Supported Treadmill Walking on Land and Underwater Using Foot-Worn Inertial Measurement Units and Machine Learning for Gait Event Detection.

Gait rehabilitation commonly relies on bodyweight unloading mechanisms, such as overhead mechanical support and underwater buoyancy. Lightweight and wireless inertial measurement unit (IMU) sensors provide a cost-effective tool for quantifying body segment motions without the need for video recordings or ground reaction force measures. Identifying the instant when the foot contacts and leaves the ground from IMU data can be challenging, often requiring scrupulous parameter selection and researcher supervision. We aimed to assess the use of machine learning methods for gait event detection based on features from foot segment rotational velocity using foot-worn IMU sensors during bodyweight-supported treadmill walking on land and underwater. Twelve healthy subjects completed on-land treadmill walking with overhead mechanical bodyweight support, and three subjects completed underwater treadmill walking. We placed IMU sensors on the foot and recorded motion capture and ground reaction force data on land and recorded IMU sensor data from wireless foot pressure insoles underwater. To detect gait events based on IMU data features, we used random forest machine learning classification. We achieved high gait event detection accuracy (95-96%) during on-land bodyweight-supported treadmill walking across a range of gait speeds and bodyweight support levels. Due to biomechanical changes during underwater treadmill walking compared to on land, accurate underwater gait event detection required specific underwater training data. Using single-axis IMU data and machine learning classification, we were able to effectively identify gait events during bodyweight-supported treadmill walking on land and underwater. Robust and automated gait event detection methods can enable advances in gait rehabilitation.