LaTex Template for an EUCOMES 2010 Paper [311313]

Numerical Simulations and Experimental Human Gait Analysis Using Wearable Sensors

D. Tarnita1, I. Geonea2, A. Petcu3 and D.N. Tarnita4

1[anonimizat], e-mail: [anonimizat],

2[anonimizat], e-mail: [anonimizat]

3[anonimizat], e-mail: [anonimizat] ,

4[anonimizat], Romania, e-mail: [anonimizat]

Abstract. The paper presents a [anonimizat] a virtual mannequin computed in ADAMS virtual environment. Using Biometrics system which is a [anonimizat]. [anonimizat]. The obtained experimental data series were be introduced as input data in the joints of the virtual mannequin and a walking simulation was performed in ADAMS environment software. The variation of ground forces during walking are obtained by experimental data and by virtual simulation.

Key words: [anonimizat], [anonimizat], [anonimizat], ADAMS simulation

1 [anonimizat]-[anonimizat], offices, [anonimizat] [1-5]. In papers [4,5] the authors present a human-oriented approach to the study of the biped gait for a humanoid robot. [anonimizat], they pointed the main features of the human legs for a correct walking motion useful in the mechanical design of a humanoid robot. [anonimizat]. The more humanoid robots are expected to perform several application tasks in an actual human living environment.

Objective gait analysis techniques are based on the use of different devices to capture and measure information related to the various gait parameters [6].

The main four clinical applications of wearable sensors and acquisition data systems are: (a) identifying movement disorders, (b) assessing surgical outcomes, (c) [anonimizat] (4) reducing joint loading [7]. [anonimizat], and more robust than previous versions [7]. Wearable sensing systems make it possible to analyse data outside the laboratory and capture information about the human gait during the person’s everyday activities. [anonimizat], knees, thighs or waist. Different types of sensors are used to capture the various signals that characterize the human gait [8,9].

[anonimizat], hips and other human joints [10-15]. The main advantages of the wearable sensing systems are presented in [6]: Transparent analysis and monitoring of gait during daily activities and on the long term; Allows the possibility of deployment in any place, not needing controlled environments; Wireless systems enhance usability; In clinical gait analysis, promotes autonomy and active role of patients.

The force platform is now an indispensable tool in a state-of-the-art motion analysis laboratory which is used together with kinematic and EMG measurements in order to obtain maximum clinical implementation and interpretation of force plate measurements [16]. Force platforms are used in gait analysis to measure the gait by pressure or force sensors and moment transducers when the subject walks on them. Ground reaction force sensors (GRF) which measure the force exerted by the subject’s feet on the floor when he walks are determined using force platforms. GRF measurement as a part of many professional and sophisticated computer systems, was used earlier in the various gait analysis studies, i.e. identification of muscle forces in human lower limbs during sagittal plane movements in the monitoring of the patients physiotherapy process [17,18].

The first objective of this study is to measure the variation of flexion-extension angles of the human legs joints while the subject performs normal walking on a set of force platforms. The obtained experimental data series will be introduced as input data in the joints of a virtual mannequin and a virtual walking simulation will be performed in ADAMS environment software. The variation of ground forces will be obtained by experimental data and by virtual simulation and will be compared. The second objective supposes the determination, by numerical simulation, of the reactions developed in the lower limbs joints in order to perform an analysis with the finite element method and to obtain the stresses and deformations maps for the normal joints, for the osteoarthritic joints and for prosthetic joints.

2 Experimental study

The experimental method which allows obtaining the kinematic parameters diagrams for the human knee joint uses a Biometrics Ltd data acquisition system based on electrogonimeters and on force platforms. Biometrics Ltd is a world leader in the design, manufacture and distribution of technologically advanced sensors, instruments and software for the demanding needs in biomedical and engineering research [19]. DataLog MWX8 is the latest in data acquisition technology for portable data collection and ambulatory monitoring on 8 channels simultaneously in human gait, human performance, medical research, robotics [19].

A lot of research studies used for human gait analysis various wearable sensors such as, for example, electogoniometers, and the Biometrics data acquisition system [20-23]. The ankle movements of healthy individuals walking on a treadmill were studied in [20], while a set of Biometrics electrogoniometers were used to measure [22, 23]. In paper [21] the participants were asked to schedule a work day to wear Biometrics electrogoniometers and in goniometers were applied on hip, knee, and ankle joints of the paretic and nonparetic leg of stroke patients.

Using Biometrics system and the appropriate sensors (electrogoniometers and force platforms), the experimental measurements was performed for three trials of six consecutive walking cycles of a male subject. The subject was pain-free and had no evidence or known history of motor and skeletal disorders or record of surgery to the lower limbs. The experimental data were acquired for the right and left ankle joints, right and left knee joints and right and left hip joints, for overground walking cycles on six force platforms. In Table 1 the values ​​for anthropometric data of male subject are presented.

Table 1. Values of anthropometric data

In the performed experiments six force platforms, 3 for right leg and 3 for left leg, type FP4, from Bometrics system were used. These were arranged on a flat surface so the distance between them is equal to the steps of the subject. Taking into account the anthropometric data of the subject, the distance between the first and second force platform representing a step has been selected for a value equal to 580 mm and the distance between the left and right force platforms was chosen 280 mm. For data acquisition a 8-channel DataLOG at a frequency of 50 Hz (MWX 8 Biometrics Ltd) was used, each platform being assigned to a channel numbered from 1 to 6 in order of settlement on the floor surface. DataLOG was connected with each platform using a H2000 cable, and transfer data from the computer DataLOG was conducted in real time via Bluetooth.

For measuring angles variation during the gait, for the six joints of both lower limbs, six flexible electrogoniometers were used. For the ankle joint for both left and right legs were used two twin axis goniometers SG100. Attachment to the ankle joint was performed in two steps. First the distal endblock of the SG110 was attached using double-sided tape on the back of the heel. After flexion of ankle, the proximal endblock was attached to the posterior of the leg so that the axis of the leg and endblock coincide. The connection with the second DataLOG is performed using a cable J1500 using the channel 1 for the right leg and channel 4 for left leg from the DataLOG. For knee joints of both legs two twin axis goniometers SG 150 were used. Starting with the subject standing in the neutral position with the foot on a flat surface, the distal endblock was mounted using double-sided tape laterally on the leg so the axes of the leg and endblock coincide, when viewed in the sagittal plane , with the leg fully extended in the position of reference extend the goniometer to maximum length and attach the proximal endblock to the thigh so the axes of the thigh and endblock coincide. The connection with DataLOG is performed using a cable J1000 using the channel 2 for the right leg and channel 5 for left leg from the DataLOG. For hip joints of both legs two twin axis goniometers SG 150 were used. The proximal endblock was attached using double-sided tape to the side of the trunk in the pelvic region. With the limb in the position of reference, the goniometer was extend to maximum length and the distal endblock was attached to the thigh so that axes of the thigh and endblock coincide. The connection with DataLOG is performed using a cable J500 using the channel 3 for the right leg and channel 6 for left leg from the DataLOG. All measurements were performed with a frequency of 50 Hz

Three experimental trials were approved by the University of Craiova human ethics research committee and they were performed in the biomechanics laboratory at INCESA Research Centre, University of Craiova. Before starting the experiments there were collected anthropometric data from the human subject. In Fig. 1 the data acquisition system mounted on the subject is shown.

Fig. 2 Subject with mounted electrogoniometers and Biometrics data acquisition system

In Fig.2 different Biometrics flexible electrogoniometers and the DataLog device are presented.

Fig. 1 Biometrics flexible electrogoniometers and DataLog data acquisition system [19]

3 Numerical Simulations

3.1. Multibody Model of a Mannequin Walking

Multibody methods are used for human gait analysis when the investigation is focused on mechanical aspects, such establish of joints reaction forces [24]. For that purpose is used the inverse dynamics approach, using as inputs data the human leg joints angular variations to establish the joints reaction forces and torques.

In this research a virtual mannequin in the assumption of ground walking is modeled as a multibody model. The virtual model of the mannequin, which follows the anthropometric data of human subject, taken from Table 1 is developed in Solid Works virtual environment. A detail of designed mannequin presenting 14 rigid parts is presented in fig. 3.

Mannequin walking is a special kind of the multibody system with the occurrence of impacts, friction and slipping. All these aspects observed during the foot-ground contact have to be introduced into the model to make it real. In this research, both the slips and impacts are taken into consideration. The model was developed in ADAMS environment [25].

Mannequin legs are modeled realistically consisting of 3 parts: thigh, shank and foot. Each revolute joint, representing the hip, knee and ankle is additionally constrained by a motion generator, as presented in fig. 3

Fig. 3 Virtual model of a mannequin in ADAMS and ground-foot contact parameters

Motion laws of the human leg joints, established experimentally, are introduced in ADAMS using a spline function and the Akima fitting method [26] [fig.4].

Fig. 4 Left knee angular motion obtained in ADAMS by using spline function.

The impact and friction effects are considered in the ground reaction forces modeling. It is used the impact model contact in ADAMS, and the friction is specified according Coulomb model. A number of test simulations are performed in order to choose proper values of estimated stiffness k of the foot in shoe and proper values of damping parameters d and C. Contact parameters according to this model are presented in fig. 3. For this simulation approach the ground reaction forces are computed in ADAMS (they do not play the role of input data). Comparison of the measured ground reaction forces with the calculated ones are used to validate the model. The calculated and the measured results are similar, however there are some differences. These differences are caused mostly by the fact that foot is modeled as rigid. To obtain better results the model of foot and the model of contact should be made more complex [27].

Dynamic simulation in ADAMS is performed using the WSTIFF solver and SI2 integrator [28]. The gait patterns of the markers attached to the ankle, knee and hip, obtained from ADAMS dynamic simulation, are shown in fig. 5. Also superimposed frames from mannequin walk animation are shown.

Fig. 5 Mannequin leg joints trajectory and superimposed walking frames, computed in ADAMS

4 Results

The angular amplitudes of the six human joints flexion-extension during the gait performed on the treadmill were obtained from the report generated by the acquisition system as data files. The results of measurements using Biometrics system were exported in .txt as datasets and then imported into SIMIMOTION software. The diagrams of Ground Reactions Forces corresponding to the trial 1, obtained for the six force platforms are shown in Fig. 6 and the corresponding diagrams of the ankle, knee and hip flexion-extension angles collected using the acquisition data system for both legs (right and left) are shown in Fig. 7. The graphs were divided into gait cycles starting with the point of beginning of the first force platform and ending with the end of the six force platform.

Fig. 6 The experimental Ground Reactions Forces variations on the six force platforms

After splitting the initial cycles, the normalized flexion-extension cycles for each of six human joint and for each of six force platforms were obtained.

Fig. 7 The experimental diagrams obtained for six joints (right and left)

To obtain the average cycles of each movement cycle joints, in app „calculated envelope" of SIMIMOTION were introduced the gait cycles for right leg joints corresponding to the force platforms 1, 3 and 5 and for joints of left leg gait cycles corresponding force platforms 2, 4 and 6. The mean of experimental Ground Reaction Forces obtained from the six platforms during three walking trials, for right leg and for left leg are presented in Fig.8. The mean flexion-extension experimental cycles for the six lower limbs joints obtained from the collected data of three walking trials are also presented in Fig.8 (c) – h)).

b)

c) d)

e) f)

g) h)

Fig. 8 Medium experimental cycles obtained from collected data computed in SIMIMOTION: a) Mean Ground Reaction Forces for right leg cycle; b) Mean Ground Reaction Forces for right leg and left leg cycles, corresponding to a complete right leg cycle ; Flexion- extension cycles during a walking cycle for lower limbs joints: c) right ankle; d) left ankle; e) right knee; f) left knee; g) right hip; h) left hip

From Fig.6 and Fig.8 a) and b), we can see that the maximum value experimentally recorded is found in the second extreme point and it is 810 N (P1), ie 1.22 BW (body weight) where BW = 660N. In the first extreme point is recorded the value 780N (P2), ie 1.18 BW, and the minimum is 780 N (P3), ie 0.92 BW. The results are similar to those obtained in the papers [29-31] where the maximum value is comprising in the interval [1,15 BW;1,25 BW] and the minimum values was comprised in the interval [0,82 BW; 095 BW]. The maximum value corresponds to an abscissa value equal to the percentage 50%, while for a complete gait cycle normalized for right foot are considered 100% percentages.

Comparing the results from the papers [29-31] we notice that the percentage values are similar because 50% of the full cycle is equivalent to ≈80% of the active cycle (the period of contact of the foot with the force platform), as it is presented in the cited papers.

From the Fig. 8 c, d, e, f, g, h, it results a good repeatability of cycle gait corresponding to the five trials, obvious fact by parallel curves: Mean, Mean + StdDev, Mean-StdDev and by the fact that the standard deviation remains in allowed limits is <10%, excepting the areas corresponding to the maximum normalized values of the cycles.

a) b) c)

Fig. 9 Experimental flexion-extension medium cycles for: a) right and left ankles; b) right and left knees; c) right and left hips

Referring to Fig. 9 a, b, c, we can conclude that the subject are executing larger movements with the right foot in comparison with the left foot. By comparison the maximum values of flexion angles are: for the right ankle the value is 13° vs left ankle which is 7°; for the right knee the maximum value is 68° while for the left knee is 52°; for the right hip the value is 21° vs the hip left value which is 17°. The results obtained in the present study are comparable to those obtained in the papers [2,29,30] both in terms of asymmetry and maximum values.

The second set of results are the ones obtained by numerical simulations of mannequin walking, performed in ADAMS environment.

Ground contact forces variations in time obtained by virtual simulation, for right and left leg are presented in fig. 10.

Fig. 10 Ground contact forces for right and left leg, computed in ADAMS simulation.

Comparing the values of the ground reaction forces obtained by experimental tests on the force platforms and by numerical simulations in ADAMS we observe that they are similar, the maximum values of about 900 N being occurred when the heel touches the floor and when the toes push off. The maximum values of the ground reaction force computed in ADAMS for the right leg is 900 N and for the left leg is about 840 N. The values obtained from ADAMS simulation for the ground reaction forces are comparable with the experimentally measured ground reaction forces, presented in fig. 6. The differences could appear because of the contact conditions during mannequin walking. The horizontal displacements (on z-axis which represents the movement direction) and vertical displacements (on y-axis) of right hip attached marker, resulted from ADAMS simulation are presented in fig. 11. Similar diagrams of horizontal (on z-axis) and vertical displacements were obtained for the other five joints. Right hip translational velocity, upon horizontal and vertical directions, during mannequin walking, is shown in fig. 12. Similar diagrams of horizontal and vertical velocities were obtained for the other five joints.

Fig.11 Right hip translational displacements, upon horizontal direction (on z-axis) and vertical direction (on y-axis) computed in ADAMS

Fig.12 Right hip translational velocity, upon horizontal direction (on z-axis) and vertical direction (on y-axis) computed in ADAMS

Another simulation results consist in legs joints reaction forces and joints torques. Vertical reaction forces developed in the virtual knee joint, for right and left leg, are presented in fig.13. Similar diagrams were obtained for the other joints.

Fig. 13 Right and left knee vertical reaction forces (on y-axis), computed in ADAMS simulation

The joints torques are computed in ADAMS simulation. In fig. 14 are shown the variations in time of torques magnitude for both (right and left) mannequin knee joints. Similar diagrams of torques magnitude were obtained for the other joints.

Fig.14 Right and left knee joints torques magnitude computed in ADAMS

The obtained values of the joint torques from ADAMS simulation are comparable with those results obtained by other researchers [27].

5 Conclusions

A study based on the tools of gait analysis of human knee movement is presented. The purpose of this study was to investigate the human flexion-extension movements of the human lower limb joints and the ground reaction forces developed during walking on the platform forces. The study presents a comparison based on the experimental data collected during walking cycles for both legs joints of a healthy subject and on the numerical simulations in ADAMS environment. Biometrics acquisition system was considered to collect the data files from the six human joints of both legs and from the six force platforms. The kinematic data of the flexion extension angles for human ankle, knee and hip joints were obtained. Based on the average human anthropometric data and using SolidWorks software, the virtual model of a mannequin was developed and then transferred to ADAMS simulation environment in order to obtain the reaction forces in the six joints of both lower limbs. The results of ADAMS simulation obtained for the ground reaction forces are similar with the experimental measurements. The joint torque values and profiles of the simulated gait are similar with the results presented in other works. The reaction forces developed in every lower limb joint are very useful for understanding normal and pathological musculoskeletal motion and for the study of the stresses and displacements developed in normal joints, in affected and prosthetic joints, using the Finite Element Method. Considering the reaction forces obtained in joints by numerical simulation, the stress and deformation maps developed in the knee prosthetic joint could be obtained, as we show in fig.15 [32]. As future work we intend to take into consideration the lateral joints motion and the arm swing, as important factors which can influence the results during gait. As future work we intend to use the numerical simulation validated by experimental results to design and develop human-inspired robotic structures.

Fig.15. The prosthetic knee assembly joint; the stress maps and deformation maps

References

Tlalolini, D., Christine, C., Yannick, A.: Human-Like Walking: Optimal Motion of a Bipedal Robot With Toe-Rotation Motion. Transactions on Mechatronics, Institute of Electrical and Electronics Engineers 16 (2), 310-320 (2011)

Kanako, M., Mitsuharu, M., Fumio, K., Shuuji, K., Kenji, K., Kazuhito, Y.: Human-like Walking with Toe Supporting for Humanoids. International Conference on Intelligent Robots and Systems, San Francisco, CA, USA, (2011)

Jung-Yup, K., Ill-Woo, P., Jungho, L., Min-Su, K., Baek-kyu, C., Jun-Ho, Oh.: System Design and Dynamic Walking of Humanoid Robot KHR-2. International Conference on Robotics and Automation Barcelona, Spain, (2005)

Matthieu, L., et al.: The Poppy Humanoid Robot: Leg Design for Biped Locomotion, International Conference on Intelligent Robots and Systems, Tokyo, Japan, (2013)

Fumiya, L., Yohei, M., Jürgen, R., André, S.: Toward a human-like biped robot with compliant legs, Robotics and Autonomous Systems, 57, 139-144 (2009)

Muro-de-la-Herran, A., et al.: Gait Analysis Methods: An Overview of Wearable and Non-Wearable Systems. Highlighting Clinical Applications, Sensors 14, 3362-3394 (2014)

Shull, PB., Jirattigalachote, W., Hunt, MA., Cutkosky,MR., et al.: Quantified self and human movement: A review on the clinical impact of wearable sensing and feedback for gait analysis and intervention, Gait & Posture, 40, 11–19 (2014)

Tao, W., et al.: Gait analysis using wearable sensors. Sensors 12, 2255–2283 (2012)

Abdul, R. A. H., Zayegh, A., Begg, R. K., Wahab, Y.: Foot plantar pressure measurement system. A review, Sensors, 12, 9884–9912 (2012)

Tarnita, D., Marghitu, D.:Analysis of a hand arm system, Robotics and Computer-Integrated Manufacturing, 29(6), 493–501 (2013)

Tarnita D., Catana, M., Tarnita, D.N.: Nonlinear Analysis of Normal Human Gait for Different Activities with Application to Bipedal Locomotion. Ro. J. Tech. Sci. Appl. Mech., 58 (1-2), 177–192 (2013)

Mohamed, A. A., Baba, J., Beyea, J., Landry, J., Sexton, A., McGibbon, C. A. : Comparison of strain-gage and fiber-optic goniometry for measuring knee kinematics during activities of daily living and exercise, J Biomech Eng 134:084502 (2012)

Van der Linden, M. L., Rowe, P. J., Nutton, R. W. : Between-day repeatability of knee kinematics during functional tasks recorded using flexible electrogoniometry, Gait Posture 28, 292–6 (2008)

Indramohan, V. P., Valsan, G., Rowe, P. J.: Development and validation of a user friendly data logger (SUDALS) for use with flexible electrogoniometers to measure joint movement in clinical trials. JMed Eng Technol, 33, 650–5 (2009)

Huddleston, J., Alaiti, A., Goldvasser, D., Scarborough, D., Freiberg, A., Rubash, H., et al: Ambulatory measurement of knee motion and physical activity: preliminary evaluation of a smart activity monitor. J Neuroeng Rehabil, 3, 1–10 (2006)

Sutherland, D. H.: The evolution of clinical gait analysis part III – kinetics and energy assessment. Gait & Posture, 21, 447–46 (2005)

Leusmann, P., Mollering, C., Klack, L., et al.: Your Floor Knows Where You Are: Sensing and Acquisition of Movement Data. In Proceedings of 2011 12th IEEE International Conference on Mobile Data Management (MDM), Luleå, Sweden, 61–66 (2011)

Lincoln, L. S., et al.: An Elastomeric Insole for 3-Axis Ground Reaction Force Measurement. In Proceedings of 2012 4th IEEE RAS EMBS International Conference on Biomedical Robotics and Biomechatronics (BioRob), Rome, Italy, 1512–1517 (2012)

http://www.biometricsltd.com

Moriguchi, C.S., Sato, T.O. , Gil Coury, H.J.C.: Ankle movements during normal gait evaluated by flexible electrogoniometer. Rev. bras. Fisioter, 11(3) (2007)

Teran-Yengle, P., Yack, H. J.: Knee kinematics during daily standing and walking in healthy women with knee hyperextension. XXIV Congress of the international society of biomechanics (2013)

Yano, H., Tanaka, N., Kamibayashi, K., et al.: Development of a Portable Gait Rehabilitation System for Home-Visit Rehabilitation. The Scientific World Journal Volume, (2015)

Hirakawa, Y., Hara, M., Fujiwara, A., Hanada, H., Morioka, S.: The relationship among psychological factors, neglect-like symptoms and postoperative pain after total knee arthroplasty, Pain Res Manag, 19(5), 251-256 (2014)

Wojtyra, M.: Dynamical Analysis of Human Walking. 15th European ADAMS Users, Conference Technical Papers, Rome, Italy (2000)

Lu, T.W., Chang, C.F.: Biomechanics of human movement and its clinical applications. Kaohsiung Journal of Medical Sciences, 28, 13-25 (2012)

Kecskemethy, A.: Integrating efficient kinematics in biomechanics of human motions. Procedia IUTAM, 2, 86–92 (2011)

Chowdhury, S., and Kumar, N.: Estimation of Forces and Moments of Lower Limb Joints from Kinematics Data and Inertial Properties of the Body by Using Inverse Dynamics Technique. Journal of Rehabilitation Robotics, 1, 93-98 (2013)

MSC.ADAMS 2013 User Manual

Stansfield, B. W., Hillman, S. J., et al.: Regression analysis of gait parameters with speed in normal children walking at self-selected speeds. Gait & Posture, 23, 288–294 (2006)

Fregly, B. J., Reinbolt, J. A., et al.: Design of Patient-Specific Gait Modifications for Knee Osteoarthritis Rehabilitation. IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, 54 (9) (2007)

Akashi, P. M. H., et al.: The effect of diabetic neuropathy and previous foot ulceration in EMG and ground reaction forces during gait. Clinical Biomechanics, 23, 584–592 (2008)

Calafeteanu, D., Tarnita, D., et al.: Influences of Varus Tilt on the Stresses in Human Prosthetic Knee Joint. Applied Mechanics and Materials, 823, 143-148, (2016)