ARTICLE IN PRESS Journal of Biomechanics 43 (2010) 2017–2021
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Short communication
A novel gait platform to measure isolated plantar metatarsal forces during walking Wen-Ming Chen a, Peter Vee-Sin Lee b, Seung-Bum Park c, Sung-Jae Lee d, Victor Phyau Wui Shim e, Taeyong Lee a,n a
Division of Bioengineering, National University of Singapore, Block E3A #07-15, 7 Engineering Drive 1, Singapore 117574, Singapore Department of Mechanical Engineering, University of Melbourne, Australia c Footwear Biomechanics Team, Footwear Industrial Promotion Center, Busan Economic Promotion Agency, South Korea d Department of Biomedical Engineering, Inje University, South Korea e Department of Mechanical Engineering, National University of Singapore, Singapore b
a r t i c l e in fo
abstract
Article history: Accepted 19 March 2010
A new gait platform described in this report allows an isolated measurement of the vertical and shear forces under an individual metatarsal head during barefoot walking. The apparatus incorporated a customized tactile force sensor and a high-speed camera system, which enabled easy identification of a single anatomical landmark at the forefoot’s plantar surface that is in contact with the sensor throughout stance. After calibration, the measured peak forces under the 2nd MTH showed variability of 3.7%, 9.2%, and 8.9% in vertical, anterior–posterior, and medial–lateral directions, respectively. The device therefore provides information about the magnitude and timing of such local metatarsal forces, and has been shown to be of significant research and clinical interest. Its ability to achieve this with a high degree of accuracy ensures its potential as a valuable research tool. & 2010 Elsevier Ltd. All rights reserved.
Keywords: Gait platform Shear force Metatarsal head Forefoot’s plantar surface
1. Introduction The ball of the forefoot supports a significant portion of the body’s weight during human locomotion. The highest ground reaction force (GRF) has been estimated to be beneath the foot’s plantar metatarsal sites, including five metatarsal heads (MTHs) and the underlying protective fat pads, serving as pivot points during push-off (Hicks, 1955). These localized loadings produce high pressures on MTHs, which could cause problems to a pathological foot. Excessive pressure (i.e., arising from vertical force) and shear stresses (i.e., arising from shear force) at MTHs have been shown to be associated with metatarsalgia in rheumatoid arthritis (Roy, 1988), and have been implicated in the development of ulcers in the insensate foot (Cavanagh et al., 1993) and other forefoot structure abnormalities (Sanders et al., 1992). It is therefore important to be able to accurately determine the localized forces acting on MTHs, in order to prevent tissue damage in the pathological foot. There is still a lack of suitable systems capable of measuring the three-dimensional forces acting on localized anatomical sites such as the plantar MTHs. Force sensor technology is still far from miniaturization to the point where it can accurately relate force
n
Corresponding author. Tel.: + 65 6516 1471; fax: +65 6872 3069. E-mail address:
[email protected] (T. Lee).
0021-9290/$ - see front matter & 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbiomech.2010.03.036
distribution data to specific anatomical sites (e.g., the 2nd MTH). Davis et al. (1998) presented a method whereby strain-gauged force sensors were arranged in an array to measure the plantar force distribution. Recently, Mackey and Davis (2006) developed a similar optical based force sensor array. Gross plantar force patterns obtained, however, can only be ‘‘mapped’’ onto a portion of the foot (Yavuz et al., 2009). Subtle variations within individual MTH often cannot be distinguished. Today, many researchers have theoretically estimated the local GRF acting at foot areas of concern based on the local plantar pressure distribution and the global GRF (Abuzzahab Jr. et al., 1997; Uccioli et al., 2001; Giacomozzi et al., 2008). However, its accuracy can be significantly compromised due to the fact that the vertical and the shear force components may not have a simple linear relationship at the foot–ground interface (Yavuz et al., 2007). This study describes the construction of a gait platform-type apparatus, and a pilot study to obtain the vertical, anterior– posterior (AP), and medial–lateral (ML) GRF components acting at the 2nd MTH during walking.
2. A new gait platform system A mini tactile force sensor capable of detecting the forces at three orthogonal directions at the foot–ground interface was developed. The sensor was incorporated into a gait platform,
ARTICLE IN PRESS 2018
W.-M. Chen et al. / Journal of Biomechanics 43 (2010) 2017–2021
which enabled direct visual observation of the forefoot’s plantar surface using a high speed camera. This set-up ensured easy identification of a single anatomical landmark (e.g., 2nd MTH) at the forefoot’s plantar surface in contact with the force sensor. 2.1. Force sensor design The force sensor was designed by utilizing the shear-web principle and strain-gauge based techniques (Fig. 1). It was fabricated using a single piece of aluminum bar (2024-T351), with a flat top sensing surface that measured a square contact area of 1.9 1.9 cm2. During operation, the load applied at the sensing surface produced a nearly uniform strain field at each shear web structure located at the sensor body, onto which a set of miniature 901 Tee strain gauge rosettes (Vishay J2A-13-S254R350) were bonded. The sensor was instrumented with a total of five sets of gauge rosettes, positioned accordingly to capture the three orthogonal forces independently. The first set (A1/A2) sensed the vertical force component. The subsequent second and third sets (B1/B2, C1/C2) measured the AP force component, and finally the fourth and fifth sets (D1/D2, E1/E2) measured the ML force component. This arrangement provided three sensor channels measuring the normal force, the AP and ML shear forces at the foot’s plantar surface (Fig. 1B). Signals from each strain gauge set were fed into a Wheatstone-bridge circuit with temperature compensation. Signal conditioning included five precision instrumentation amplifiers (Tokyo Sokki Kenkyujo CO., Ltd.) for differential signal processing. Sensor outputs were collected at a sampling rate of 100 Hz and stored in a portable scope recorder (Yokogawa SL1400).
correlation coefficients (R2 value) were greater than 0.999 for all channels. Cross-talk effects were found to be less than 0.6% in all cases. In order to determine whether the sensor has adequate frequency response to transient force fluctuations during gait, the sensor was driven with a mechanical sinusoidal wave vibrator (Beta Corp.). The amplitude response of the sensor was found to be constant within 2% up to 50 Hz of sinusoidal vibration. 2.3. Construction of the new gait platform The sensor was mounted flush onto a transparent acrylic (polymethyl methacrylate) gait plate, which could accommodate the foot of the human subject (Fig. 3). Beneath the gait plate, a reflective mirror was positioned at 451 to the vertical direction, permitting direct visualization of the plantar aspect of the forefoot using a Photron Fastcam Super 10 K (Tokyo) high-speed camera, placed at right angle to the platform. The mirror has a laser cut rectangular opening to accommodate the sensor body. As shown in Fig. 3B, the image captured clearly indicates the sensor location in relation to a particular anatomical landmark (e.g., the 2nd MTH). The high-speed camera was set to record images at 100 frames per second (fps) with a resolution of 512 480 pixels, providing real-time images of the forefoot plantar surface during walking. The assembled platform (load cell and gait platform) was embedded in a straight, 7-m long, 1-m wide walkway. The entire surface of the walkway was covered with a slip-resistance material. Visual checks ensured that the gait platform and walkway did not alter the normal gait during subject walking.
3. Pilot study using the gait platform 2.2. Sensor calibration Each of the three sensor channels was independently calibrated using an Instron machine (Model 5848). A loading and unloading cycle of 4 Hz and up to 200, 50, and 50 N were applied along the vertical, AP, and ML axes, respectively. Cross-talk effects within the sensor were checked by sequentially loading the specific channels and recording the outputs from the other channels. Fig. 2 shows the individual loading and unloading plots for vertical (Channel A), AP (Channel BC), and ML (Channel DE) force components. A simple linear regression equation was sufficient to determine the respective calibration factors for each channel. The
A 26-year-old male subject, height of 169 cm and body weight of 65.1 kg, with no foot pathology volunteered for the pilot trial using the gait platform. Informed consent was obtained according to the procedures of the National University of Singapore Institutional Review Board. Prior to data collection, the location of the subject’s 2nd MTH was identified with a black ink dot after palpating the underlying metatarsal and its tuberosity. Nevertheless, this marked location would be blocked by the sensor itself whenever the 2nd MTH came into contact with the force sensor (see camera’s view in Fig. 3B), making identification of a single MTH difficult. Furthermore, maintaining a consistent placement of the foot in relation
Fig. 1. (A) Schematic diagram of sensor showing the positions of the strain gauges and (B) photograph showing the attachment of strain gauges to the front surfaces of the sensor body. Gauges (not shown) are also bonded to the rear surfaces. The positions of the vertical and shear channels are also shown.
ARTICLE IN PRESS W.-M. Chen et al. / Journal of Biomechanics 43 (2010) 2017–2021
2019
Fig. 2. Calibration graphs showing linearity and minimal mechanical cross-talk of (A) the vertical, (B) AP and (C) ML shear channels.
Fig. 3. (A) Schematic diagram of the gait platform. (B) Image of the foot making contact with the gait plate, sensor location (seen as white dash line) in relation to the placement of the foot plantar surface can be easily monitored from the reflected mirror image by a high speed camera positioned to capture the side view.
to the force sensor was important in order to obtain reproducible results for successive trials. These problems were solved by using a reference image called the ‘‘MTH template’’ as shown in Fig. 4. In order to create image templates for the 2nd MTH, two forefoot plantar images were captured with and without the sensor in position (Fig. 4A and B). As shown in Fig. 4A, the target metatarsal site could be determined within a 1.9 1.9 cm2 area centered at the black ink dot which marked the 2nd MTH. Whilst in Fig. 4B, the actual sensor location could be determined. Using a customized Matlab (Mathworks Inc.) code, the two images were
matched by translations in the 2-D plane according to the three non-collinear ink dots marked on hallux (R1, R2, R3) as shown in Fig. 4C. ‘‘MTH template’’ represents the registered images with minimal errors induced due to the discrepancy between the target MTH and the actual sensor locations as evaluated by Xi and Yi (Fig. 4C). Subsequently, the subject initiated gait from a stationary posture and landed the right foot on the platform at the first step of walking. The validity of using the ‘‘one-step’’ protocol in collecting gait variables has been reported by Peters et al. (2002).
ARTICLE IN PRESS 2020
W.-M. Chen et al. / Journal of Biomechanics 43 (2010) 2017–2021
Fig. 4. Method to create the ‘‘MTH template’’: images of the plantar surface of the forefoot (A) without and (B) with sensor during barefoot walking. Bony mark was used to identify the target metatarsal site. (C) R1 R3 represents three non-collinear markers for image registration between A and B. Xi and Yi are errors due to metatarsal site and sensor location discrepancy.
Table 1 The Xi and Yi offset errors for each of the five successive trials at the 2nd MTH. Offset values (mm)
T1
T2
T3
T4
T5
Avg. (7 Std.)
Xi Yi
0.481 0.191
0.712 0.415
0.165 0.179
0.332 0.806
0.477 0.253
0.433 (70.203) 0.369 (70.262)
Averages (7 standard deviations) of the offset values are indicated in the last column.
Fig. 5. Simultaneous recordings of force traces generated in (A) the vertical, (B) AP and (C) ML directions from the 2nd metatarsal site of five walking trials. The five overlaid waveforms’ data corresponded to the five trials. Note that positive values of AP channel denote anterior-directed AP shear forces, while positive values of ML channel are lateral-directed ML shear forces.
ARTICLE IN PRESS W.-M. Chen et al. / Journal of Biomechanics 43 (2010) 2017–2021
Data collection began as subjects started to plantarflex the right foot at the late stance phase and subsequently push-off, further completing a series of 5 to 6 steps to the end of the walkway at a self-selected speed. In order to obtain measurements at the 2nd MTH, the ‘‘MTH template’’ displayed on a laptop screen was compared with the real-time image from the FASTCAM system, so as to guide the subject to an appropriate starting point (i.e., heelstrike position). In this pilot study, an average of three walking trials is required to ensure that the 2nd MTH would exactly strike the force sensor. Five successive walking trials recording the GRF components at the 2nd MTH for one subject were collected. With the current gait protocol, the system could achieve an average Xi and Yi errors of 0.433 and 0.369 mm, respectively, for five trials, indicating that the offset variability was minimal (Table 1). The vertical, AP, and ML forces underneath the 2nd MTH during walking were presented in Fig. 5. The repeatability, quantified as the standard deviation of the maximum local force amplitude, was found to be 3.7%, 9.2%, and 8.9% in the vertical, AP, and ML directions, respectively. The average peak vertical forces of 33.86% b.w. are within the range of a previous study (28.376.9% b.w.) conducted by using an commercial EMED (Novel) capacitance sensor system (Jacob, 2001).
4. Conclusion The combined load sensor and gait platform successfully measured the three-dimensional dynamic contact forces underneath the 2nd MTH. The advantage offered by the current system is that only a single force sensor is required to accurately measure isolated plantar metatarsal forces during walking. Such local variations in the vertical and shear force components, including their magnitude and temporal characteristics, are of significant research and clinical interest (Tappin and Robertson, 1991; Perry et al., 2002; Yavuz et al., 2008).
Conflicts of interest statement The authors declared that no conflict of interest exists.
2021
Acknowledgement This work was supported by a Grant from the Temasek Defence Systems Institute (Project no.: TDSI/09-009/1A), Singapore. References Abuzzahab Jr., F.S., Harris, G.F., Kidder, S.M., 1997. A kinetic model of the foot and ankle. Gait & Posture 5, 148. Cavanagh, P.R., Simoneau, G.G., Ulbrecht, J.S., 1993. Ulceration, unsteadiness, and uncertainty: the biomechanical consequences of diabetes mellitus. Journal of Biomechanics 26 (Suppl. 1), 23–40. Davis, B.L., Perry, J.E., Neth, D.C., Waters, K.C., 1998. A device for simultaneous measurement of pressure and shear force distribution on the plantar surface of the foot. Journal of Applied Biomechanics 14, 93–104. Giacomozzi, C., Sawacha, Z., Uccioli, L., D’Ambrogi, E., Avogaro, A., Cobelli, C., 2008. The role of shear stress in the aetiology of diabetic neuropathic foot ulcers. Journal of Foot and Ankle Research 1, 3. Hicks, J.H., 1955. The foot as a support. Acta Anatomica 25, 34–45. Jacob, H.A., 2001. Forces acting in the forefoot during normal gait—an estimate. Clinical Biomechanics 16, 783–792. Mackey, J.R., Davis, B.L., 2006. Simultaneous shear and pressure sensor array for assessing pressure and shear at foot/ground interface. Journal of Biomechanics 39, 2893–2897. Perry, J.E., Hall, J.O., Davis, B.L., 2002. Simultaneous measurement of plantar pressure and shear forces in diabetic individuals. Gait & Posture 15, 101–107. Peters, E.J., Urukalo, A., Fleischli, J.G., Lavery, L.A., 2002. Reproducibility of gait analysis variables: one-step versus three-step method of data acquisition. The Journal of Foot and Ankle Surgery 41, 206–212. Roy, K.J., 1988. Force, pressure, and motion measurements in the foot: current concepts. Clinics in Podiatric Medicine and Surgery 5, 491–508. Sanders, A.P., Snijders, C.J., van Linge, B., 1992. Medial deviation of the first metatarsal head as a result of flexion forces in hallux valgus. Foot & Ankle 13, 515–522. Tappin, J.W., Robertson, K.P., 1991. Study of the relative timing of shear forces on the sole of the forefoot during walking. Journal of Biomedical Engineering 13, 39–42. Uccioli, L., Caselli, A., Giacomozzi, C., Macellari, V., Giurato, L., Lardieri, L., Menzinger, G., 2001. Pattern of abnormal tangential forces in the diabetic neuropathic foot. Clinical Biomechanics 16, 446–454. Yavuz, M., Botek, G., Davis, B.L., 2007. Plantar shear stress distributions: comparing actual and predicted frictional forces at the foot–ground interface. Journal of Biomechanics 40, 3045–3049. Yavuz, M., Hetherington, V.J., Botek, G., Hirschman, G.B., Bardsley, L., Davis, B.L., 2009. Forefoot plantar shear stress distribution in hallux valgus patients. Gait & Posture 30, 257–259. Yavuz, M., Tajaddini, A., Botek, G., Davis, B.L., 2008. Temporal characteristics of plantar shear distribution: relevance to diabetic patients. Journal of Biomechanics 41, 556–559.
