Ground reaction force (GRF) and ground reaction moment (GRM) are critical in gait analysis. While force plates provide accurate measurements, they are costly and spatially limiting.
Objective
This study aimed to evaluate the reliability and accuracy of GRF and GRM predictions using a feedforward neural network (FNN) integrating infrared camera-based positional data with accelerometer (ACC) data from wearable devices.
Methods
Eighty participants walked at their usual pace along a 10-meter walkway over force plates. Positional and ACC data of body segments were used to train the FNN to predict GRF and GRM. Prediction accuracy was assessed using multiple metrics, including root mean square error (RMSE) and normalized RMSE (NRMSE).
Results
Combining positional and ACC data improved GRF prediction in all directions (vertical, anterior-posterior, medial-lateral). The combined dataset achieved a correlation coefficient of 0.979 for medial-lateral GRF and an NRMSE of 6.07%. GRM predictions also benefited from ACC integration, especially in the sagittal plane, where R2 reached 0.939, outperforming other models. The vertical direction and transverse axis yielded the lowest RMSE and NRMSE.
Conclusions
These findings surpass many previously reported results, demonstrating the superior performance of the proposed model compared with current state-of-the-art methods. The approach offers a cost-effective, flexible alternative to traditional force plates for clinical and sports assessments.
BoulicRThalmannNMThalmannD. A global human walking model with real-time kinematic personification. Vis Comput1990; 6: 344–358.
2.
BastienGJGosseyeTPPentaM. A robust machine learning enabled decomposition of shear ground reaction forces during the double contact phase of walking. Gait Posture2019; 73: 221–227.
3.
DavisBLCavanaghPR. Decomposition of superimposed ground reaction forces into left and right force profiles. J Biomech1993; 26: 593–597.
4.
GoldbergSROunpuuSArnoldAS, et al.Kinematic and kinetic factors that correlate with improved knee flexion following treatment for stiff-knee gait. J Biomech2006; 39: 689–698.
5.
KutznerIHeinleinBGraichenF, et al.Loading of the knee joint during activities of daily living measured in vivo in five subjects. J Biomech2010; 43: 2164–2173.
6.
MullerAPontonnierCRobert-LachaineX, et al.Motion-based prediction of external forces and moments and back loading during manual material handling tasks. Appl Ergon2020; 82: 102935.
7.
RouhaniHFavreJCrevoisierX, et al.Ambulatory assessment of 3D ground reaction force using plantar pressure distribution. Gait Posture2010; 32: 311–316.
8.
HouckJKneissJBukataSV, et al.Analysis of vertical ground reaction force variables during a sit to stand task in participants recovering from a hip fracture. Clin Biomech2011; 26: 470–476.
9.
NikooyanAAZadpoorAA. Effects of muscle fatigue on the ground reaction force and soft-tissue vibrations during running: a model study. IEEE Trans Biomed Eng2011; 59: 797–804.
10.
NüeschCValderrabanoVHuberC, et al.Gait patterns of asymmetric ankle osteoarthritis patients. Clin Biomech2012; 27: 613–618.
11.
RenLJonesRKHowardD. Whole body inverse dynamics over a complete gait cycle based only on measured kinematics. J Biomech2008; 41: 2750–2759.
12.
TurcotKAissaouiRBoivinK, et al.New accelerometric method to discriminate between asymptomatic subjects and patients with medial knee osteoarthritis during 3-D gait. IEEE Trans Biomed Eng2008; 55: 1415–1422.
13.
ChoiARLeeJMMunJH. Ground reaction forces predicted by using artificial neural network during asymmetric movements. J Precis Eng Manuf2013; 14: 475–483.
14.
ShinKYRimYHKimYS, et al.A joint normalcy index to evaluate patients with gait pathologies in the functional aspects of joint mobility. J Mech Sci Tech2010; 24: 1901–1909.
15.
WuJZChiouSSPanCS. Analysis of musculoskeletal loadings in lower limbs during stilts walking in occupational activity. Ann Biomed Eng2009; 37: 1177–1189.
16.
KaratsidisABellusciGSchepersHM, et al.Estimation of ground reaction forces and moments during gait using only inertial motion capture. Sensors2016; 17: 75.
17.
MartínezMVillagraFCastelloteJM, et al.Kinematic and kinetic patterns related to free-walking in Parkinson’s disease. Sensors2018; 18: 4224.
18.
FaberHVan SoestAJKistemakerDA. Inverse dynamics of mechanical multibody systems: an improved algorithm that ensures consistency between kinematics and external forces. PLoS One2018; 13: e0204575.
19.
JohnsonWRMianADonnellyCJ, et al.Predicting athlete ground reaction forces and moments from motion capture. Med Biol Eng Comput2018; 56: 1781–1792.
20.
O’DonovanKJKamnikRO’KeeffeDT, et al.An inertial and magnetic sensor based technique for joint angle measurement. J Biomech2007; 40: 2604–2611.
21.
HorstFLapuschkinSSamekW, et al.Explaining the unique nature of individual gait patterns with deep learning. Sci Rep2019; 9: 2391.
22.
StetterBJRinghofSKrafftFC, et al.Estimation of knee joint forces in sport movements using wearable sensors and machine learning. Sensors2019; 19: 3690.
23.
HahnMEO’KeefeKB. A neural network model for estimation of net joint moments during normal gait. J Musculoskelet Res2008; 11: 117–126.
24.
LiuYShihSMTianSL, et al.Lower extremity joint torque predicted by using artificial neural network during vertical jump. J Biomech2009; 42: 906–911.
25.
OhSEChoiAMunJH. Prediction of ground reaction forces during gait based on kinematics and a neural network model. J Biomech2013; 46: 2372–2380.
26.
MundtMDavidSKoeppeA, et al.Intelligent prediction of kinetic parameters during cutting manoeuvres. Med Biol Eng Comput2019; 57: 1833–1841.
27.
FluitRAndersenMSKolkS, et al.Prediction of ground reaction forces and moments during various activities of daily living. J Biomech2014; 47: 2321–2329.
28.
ArdestaniMMZhangXWangL, et al.Human lower extremity joint moment prediction: a wavelet neural network approach. Expert Syst Appl2014; 41: 4422–4433.
29.
KomarisDSPérez-ValeroEJordanL, et al.Predicting three-dimensional ground reaction forces in running by using artificial neural networks and lower body kinematics. IEEE Access2019; 7: 156779–156786.
30.
JohnsonWRMianARobinsonMA, et al.Multidimensional ground reaction forces and moments from wearable sensor accelerations via deep learning. IEEE Trans Biomed Eng2020; 68: 289–297.
31.
AvdanGOnalSRekabdarB. Regression transfer learning for the prediction of three-dimensional ground reaction forces and joint moments during gait. Int J Biomed Eng Technol2023; 42: 317–338.
32.
HavashinezhadianSChiasson-PoirierLSylvestreJ, et al.Inertial sensor location for ground reaction force and gait event detection using reservoir computing in gait. Int J Environ Res Public Health2023; 20: 3120.
LimHKimBParkS. Prediction of lower limb kinetics and kinematics during walking by a single IMU on the lower back using machine learning. Sensors2019; 20: 130.
35.
CollinsTDGhoussayniSNEwinsDJ, et al.A six degrees-of-freedom marker set for gait analysis: repeatability and comparison with a modified Helen Hayes set. Gait Posture2009; 30: 173–180.
36.
CappozzoACappelloACroceUD, et al.Surface-marker cluster design criteria for 3-D bone movement reconstruction. IEEE Trans Biomed Eng1997; 44: 1165–1174.
37.
WuGSieglerSAllardP, et al.ISB Recommendation on definitions of joint coordinate system of various joints for the reporting of human joint motion—part I: ankle, hip, and spine. J Biomech2002; 35: 543–548.
38.
WelchNRichterCFranklyn-MillerA, et al.Principal component analysis of the biomechanical factors associated with performance during cutting. J Strength Cond Res2021; 35: 1715–1723.
39.
TaylorKLSheppardJMLeeH, et al.Negative effect of static stretching restored when combined with a sport specific warm-up component. J Sci Med Sport2009; 12: 657–661.
40.
McMillianDJMooreJHHatlerBS, et al.Dynamic vs. static-stretching warm up: the effect on power and agility performance. J Strength Cond Res2006; 20: 492–499.
41.
HaupenthalAFontanaHHaupenthalD, et al.Prediction of ground reaction forces while walking in water. Plos one2019; 14: e0219673.
42.
SimTYKwonHBOhSE, et al.Predicting complete ground reaction forces and moments during gait with insole plantar pressure information using a wavelet neural network. J Biomech Eng2015; 137: 091001.
43.
WeiFCrechioloAHautRC. Prediction of ground reaction forces in level and incline/decline walking from a multistage analysis of plantar pressure data. J Biomech2019; 84: 46–51.
44.
WalshGSHarrisonI. Gait and neuromuscular dynamics during level and uphill walking carrying military loads. Eur J Sport Sci2022; 22: 1364–1373.
45.
RoeslerHHaupenthalASchützGR, et al.Dynamometric analysis of the maximum force applied in aquatic human gait at 1.3 m of immersion. Gait Posture2006; 24: 412–417.
46.
KloecknerJVisscherRMTaylorWR, et al.Prediction of ground reaction forces and moments during walking in children with cerebral palsy. Front Hum Neurosci2023; 17: 1127613.
47.
SakamotoSHutabaratYOwakiD, et al.Ground reaction force and moment estimation through EMG sensing using long short-term memory network during posture coordination. Cyborg Bionic Syst2023; 4: 0016.
