2.1.

Insole Incorporation of the Fiber Bragg Grating Network

The optical platform is composed of a cork sole, with 1.0-cm thickness, in which FBG sensors were incorporated in critical points of analysis [heel, midfoot, metatarsal, and toe areas, Fig. 1(a)] to monitor the VGRF,^10,23 as shown in Figs. 1(b) and 1(c).

The cork was the material chosen to embed the sensors due to its excellent properties for this application, namely thermal isolation, malleability, and a near-zero Poisson ratio.²⁴ The first is extremely relevant, once the body temperature may influence the FBGs sensing feedback. By using a thermal isolation material, no compensation is needed in the acquired Bragg wavelength. Also, the malleable properties of the cork make it easy to incorporate the sensing elements in the required points of analysis, which additionally facilitates the sole implementation in any type of shoe configuration. Moreover, and although malleable to handle, it also provides the necessary resistance to prevent the optical fiber network to be damaged or broken during the gait movement. Finally, its near-zero Poisson ratio is an advantage to its use, once the pressure applied in one area of the plantar surface will not influence the other areas, allowing in that way to isolate the pressure points, preventing sensor’s cross talk.²⁴

The FBG sensing elements were inscribed and multiplexed into a GF1-photosensitive optical fiber (Thorlabs^®) cable, using an UV KrF Bragg Star™ Industrial-LN pulsed KrF excimer laser operating at 248 nm, applying pulses with energy of 5 mJ and repetition rate of 500 Hz. Phase masks customized for 248-nm UV laser were chosen for the 1550-nm spectral region grating inscription (Ibsen Photonics). In order to protect and provide extra resistance to the optical fiber sensing elements, the FBGs were encapsulated in epoxy resin (Liquid Lens™) cylinder structures (1.0-cm diameter and 0.5-cm height). Each sensing element consists of such cylindrical epoxy structure with the FBG at the middle position. The entire FBG sensing network (epoxy structure included) was incorporated within the cork structures, in order to obtain a compact (no air gap within it) insole.

Two solutions were developed, one to be used as a fixed platform [Fig. 1(b)] containing five FBG-based sensors and another to act as an instrumented insole to be adapted into a shoe [Fig. 1(c)] with six FBG sensors. The second solution was intended to be able to monitor all the critical points^10,23 for VGRF analysis, but due to equipment restrictions, only six points were considered.

Fig. 1

Schematic representation of (a) the foot plantar main areas¸ (b) the sensors network implemented for a fixed platform, and (c) the sensors network implemented for the insole adapted in a shoe.

For the insole, an FBG temperature sensor was also inserted into the cork near the first sensing point. This procedure was implemented to guarantee that the sensing points are not influenced by any temperature change, and if they are, it can be properly compensated.

2.2.

Monitoring System Assemble and Calibration

For the monitoring and calibration of both systems, the FBG sensing network was connected to a portable interrogation system constituted of a battery, a miniaturized broadband optical ASE module (B&A Technology Co., As4500), an optical circulator (Thorlabs, 6015-3), and an optical spectrometer (Ibsen, I-MON 512E-USB). The latter operates at a maximum rate of 960 Hz, with a wavelength resolution of 5 pm, responsible for the acquisition of the Bragg wavelength shift, which is proportional to the vertical force applied on the sensing element points, as shown in Fig. 2.

Fig. 2

Fixed platform monitoring system: (a) schematic representation and (b) photograph.

Prior to its implementation for VGRF determination, the developed platform and insole were calibrated. This procedure was made with one load sequence on each isolated sensor. The contact area was of 1.0 cm (diameter of the loading prove used during the calibration process). For that purpose, different load values, ranging from 0 to 170 N, in 10-N steps increment, were induced in the FBG sensing elements, using a Shimadzu^® AGS-5 kND mechanical test machine. The reflected Bragg wavelength shift of each FBG in the network was registered for each load value. The calibration values for the five sensing points in the fixed platform are presented in Fig. 3.

Fig. 3

Bragg wavelength shift as function of the applied vertical force. Points are correspondent to the experimental data and the lines to the linear fits ($0.953<R^2<0.995$).

A linear dependence of the Bragg wavelength shift with the applied loads was obtained, with sensitivity coefficients ranging from $3.38\mathrm{pm}/\mathrm{N}$ (FBG 4) to $11.06\mathrm{pm}/\mathrm{N}$ (FBG 2). The differences in the strain sensitivities may be related to the methodology used to incorporate the FBGs in the insole. As the cork used is a porous material, different amounts of the epoxy resin may have infiltrated into its surroundings. Also, although major effort was made to keep all the FBGs at the same depth of the cylinder epoxy resin, some deviations might have occurred, due to the resin’s initial liquid properties. Nonetheless, such facts do not influence the obtained final results, once all the sensing elements were independently calibrated. In future works, special attention will be given to this aspect.

The sensitivity coefficients, for each FBG sensor, in the platform and insole implementations, were later applied to retrieve the vertical force value from the acquired Bragg wavelength shift data.

3.1.

Platform Testing

The VGRF induced in the sensing elements during a normal gait movement was analyzed with the platform fixed at the ground, as shown in Fig. 4. The response of each sensing element to the VGRF during a gait cycle was repeated and acquired five times. The feedback of the platform to the displacement of the BCM was also evaluated.

Fig. 4

Schematic diagram of the protocol implemented for gait analysis using the fixed platform.

In Fig. 5, the acquired data are presented, from which it is possible to verify that the sensing network response is similar for the five passages, confirming the repeatability of the sensor’s response.

Fig. 5

VGRF obtained during the five steps and the resulting curve of all the sensors response sum for each step.

A detailed view of the stance period of the gait cycle is shown in Fig. 6. It is possible to retrieve the sequence in which the sensors are activated and the VGRF associated at each point under analysis. The observed maximum amplitude obtained for each sensor was temporally registered according to what is expected in a gait movement,⁴ verifying a typical VGRF curve during the stance period of the gait cycle.⁴

Fig. 6

Average VGRF obtained during the stance period (62%), from the individual steps analyzed (the regions of the foot touching the platform in each gait step are in dark blue).

The stance period is initiated when the heel touches the ground, HS; after the contact of the heel with the ground, the foot is moved toward a stable support position for the body in which the hip joint becomes aligned with the ankle joint; this phase is called midstance phase (MS). The purpose of the following movements is to propel the BCM forward, which is initiated by the rise of the heel (HR) and its loss of contact with the ground. In this movement, the last point of contact of the foot with the ground should be the toe at the moment known as the TO, which also corresponds to the end of the stance phase in a gait cycle.⁴

According to the layout of the sensing network, it can be observed that the beginning of the HS moment is marked by the FBG1 with greater response, which is expected, as this sensor is placed at the heel. This response is maintained until the MS stage (in which all the sensors are activated) finishes. It can also be verified that at the HR’s initial moment, the VGRF values registered for the sensor FBG 1 considerably drop, and the most accentuated forces are registered for the sensors placed in the metatarsals areas (FBG 2, 3, and 4). Also, at this moment, the VGRF registered at the toe location by the sensor FBG 5 becomes more accentuated and hits its maximum as the toe becomes the last point of support in the stance phase at the TO stage.

From the positive feedback of the fixed platform during the performed tests, it becomes evident that the method implemented is an adequate solution for VGRF monitoring during gait. Moreover, from the analysis of the forces registered during the stance phase, it is also possible to infer and monitor the plantar pressures of individuals.^4,10,23

The BCM displacements, in the body sagittal and frontal planes of motion, were also analyzed using the same platform. For that purpose, an 85-kg male, was asked to place his foot on the sensing platform and to execute a series of BCM movements (with a 5-s duration each), starting by standing still with the BCM centered (C), followed by an anterior (A) position and then back to the original position (C) from which goes to posterior (P) position and then resting again at the center (C). After the sagittal displacement, a frontal displacement was executed, in which the subject moved the BCM first to the left (L), back in the center (C) and then to the right (R), and finally back in the center (C). In Fig. 7, the implemented protocol is schematized.

Fig. 7

(a) Schematic diagram of the protocol implemented for the analysis of the BCM displacement (sagittal and frontal displacements in blue and gray, respectively) and (b) descriptive protocol (the subject remained in each position for 5 s).

During the protocol implementation, the Bragg wavelength shift induced in the sensing network was acquired and the corresponding VGRFs were retrieved. Figure 8 presents the response of each sensor, during the different moments of the tests performed.

Fig. 8

Representation of the vertical forces detected during the BCM displacements [the VGRF intensity on each foot location is colored, in the scheme, from light (less intensity) to dark (more intense) blue].

In the anterior movement, an increase of the vertical forces registered by the sensors positioned in the metatarsal and toe areas is evident, while the sensor placed in the heel section indicates a decrease of the VGRF. On the other hand, during the posterior displacement of the BCM, the VGRF at the heel area is more accentuated, while the force values at the toe and metatarsal areas decrease.

Regarding the centered position of BCM (C), the areas that are mostly actuated in the platform are the metatarsal and midfoot areas, indicating that those zones are the more pressed ones in the ground in the sustainment of the subject’s body weight, while standing.

The behavior monitored by the FBGs network is within the expected; hence, in the anterior movement of the BCM, the body weight will be mainly distributed by the metatarsal and toe areas, while in the posterior movement, it will be mostly supported by the heel area.

Regarding the frontal plane displacements, the sensors located at the extremities of the platform (FBG 2 and 4) should be the ones presenting a higher variation of the VGRF. In fact, it is noticeable that during such movements, the sensor placed in position 2 registers the increase of vertical force, when the BCM is displaced to the left, while the sensor placed in position 4 shows a similar behavior when the BCM is moved to the right. Regarding the sensors positioned at the central of the metatarsal area, the value registered through the frontal protocol implementation is mostly constant, which indicates that in such BCM displacement, the vertical force is mostly shifted between the left and the right metatarsal areas.

3.2.

Instrumented Shoe

In order to perform a dynamic and autonomous VGRF monitoring during gait, an insole with six FBG sensors network was designed and adapted to a shoe, as shown in Fig. 9.

Fig. 9

(a) Schematic representation of the designed insole, (b) photograph of the insole with the sensing network and temperature sensor, and (c) photograph of the instrumented shoe.

In addition to the sensing network for VGRF monitoring, an FBG temperature sensor¹⁶ was also adapted to the insole in order to monitor the temperature variations and guarantee that the Bragg wavelength shift registered during the gait cycles is only the result of the VGRF actuating on the sensing elements. In fact, due to the cork thermal isolation characteristics,²⁴ it is expected that the body temperature will not influence the Bragg wavelength shift of the sensing network, once its sensing elements are placed within the cork sole and not in direct contact with the body. The wavelength shift values obtained for the temperature sensor confirms what was previously stated, as its value remained constant during the gait cycles experiments, as presented in Fig. 10.

Fig. 10

Temperature and the vertical forces registered in the insole sensing network for three gait cycles.

Regarding the gait cycles analysis, the subject, in this case, a 45-kg female, was instructed to walk in a normal pace, during which the Bragg wavelength was acquired by the interrogation system. In Fig. 10, the vertical forces, registered in the insole sensing network for three gait cycles, are presented, in addition to the temperature recorded during the test.

Similar to the behavior registered in the fixed platform (Sec. 3.1), the elements initially actuated are the ones placed within the heel area (FBG 1 and 6), followed by the metatarsal area (FBG 5, 4, and 2) and ending with the element within the toe area (FBG 3) at the conclusion of the stance phase.

Also, from the gait cycles monitoring, the different phases in the gait cycle, namely the stance phase that characterizes $\sim 68\%$ of the gait cycle for the subject tested, and the swing phase, during which the foot is not in contact with the ground, are clearly distinguished. During the latter phase, the VGRF is not present, and therefore, the Bragg wavelength shift of all the sensing elements is mostly constant during that phase.