Foot Plantar Pressure Measurement System: A Review

3.1. Platform Systems

Platform systems are constructed from a flat, rigid array of pressure sensing elements arranged in a matrix configuration and embedded in the floor to allow normal gait. Platform systems can be used for both static and dynamic studies but are generally restricted to research laboratories. One advantage is that a platform is easy to use because it is stationary and flat but has the disadvantage that the patient requires familiarization to ensure natural gait. Furthermore, it is important for the foot to contact the centre of the sensing area for an accurate reading [23]. Limitations include: space, indoor measurement, and patient's ability to make contact with the platform, Figures 1 and 2 show a platform-based sensor [24,25].

3.2. In-Shoe Systems

In-shoe sensors are flexible and embedded in the shoe such that measurements reflect the interface between the foot and the shoe. The system is flexible making it portable which allows a wider variety of studies with different gait tasks, footwear designs, and terrains [23].

They are, therefore, highly recommended [11,23] for studying orthotics and footwear design but there is the possibility of the sensor slipping. Sensors should be suitably secured to prevent slippage and ensure reliable results. A further limitation is that the spatial resolution of the data is low compared to platform systems due to fewer sensors [11,23]. Figures 3 and 4 illustrate in-shoe based systems [24,26].

4.1. Target Implementation Requirements

Real-time measurement of natural gait parameters requires that sensors should be mobile, untethered, can be placed in the shoe sole, and can sample effectively in the target environment. The main requirements of such sensors are as follows:

• Very Mobile: To make a sensor mobile, it must be light and of small overall size [27,28], the suggested shoe mounted device should be 300 g or less.

• Limited Cabling: A foot plantar system should have limited wiring, wireless is ideal. This is to ensure comfortable, safe and natural gait [28].

• Shoe and Sensor Placement: To be located in the shoe sole the sensor must be thin, flexible [29] and light [27]. It is reported that a shoe attachment of mass 300 g or less does not affect gait significantly [27]. Shu et al. [30] mentioned that the sole of foot can be divided into 15 areas: heel (area 1–3), midfoot (area 4–5), metatarsal (area 6–10), and toe (area 11–15), as illustrated in Figure 5. These areas support most of the body weight and are adjusted by the body's balance; therefore, ideally the 15 sensors are necessary to cover most of the body weight changes based on the Figure 5 anatomy.

• Low Cost: The sensor must be affordable for general application [28] to benefit from inexpensive, mass-produced electronics components combined with novel sensor solutions.

• Low Power Consumption: It should exhibit low power consumption such that energy from a small battery is sufficient for collecting and recording the required data.

4.2. Plantar Pressure Sensor Requirements

The key specifications for sensor performance include: linearity, hysteresis, sensing size, pressure range and temperature sensitivity [27,29,31]. Brief discussion of these is important as a basis for the selection of a sensor for specific applications.

• Hysteresis: Hysteresis can be determined by observing the output signal when the sensor is loaded and unloaded. When the applied pressure is increased by loading or decreased by unloading, two different responses are observed (Figures 6 and 7).

• Linearity: The response of the sensor to the applied pressure, when plotted, will show the linearity figure of merit, i.e., how straight the plotted line is. Linearity indicates how simple or complicated the signal processing circuitry will be, a highly linear response requires very simple signal processing circuitry and vice versa, a linear pressure sensor is, therefore, preferred.

• Temperature Sensitivity: Sensors may produce different pressure readings as the ambient temperature changes. This may be due to the materials that are part of the sensor body as they respond differently to temperature change. A sensor with low temperature sensitivity in the 20 °C to 37 °C range is preferred [29].

• Pressure Range: The pressure range is the key specification for a pressure sensor. As different applications require different operating pressures application-specific sensor development is normally adopted in the design. Maximum pressure is the upper limit that the pressure sensor can measure and vice versa. It is also important to note that burst pressure is the maximum pressure that the sensor can withstand before breakage as opposed to maximum pressure. Foot plantar pressure values of up to 1,900 kPa are typically reported in the literature but an extreme pressure of up to 3 MPa has been documented by Urry [31]. One of the foot plantar pressure sensor designs considers 3 MPa as burst pressure value, for comparison when a healthy person of 75 kg body mass is standing on only one forefoot, if pressure is evenly distributed, the interfacial pressure for every 31.2 mm² foot plantar area approximates 2.3 MPa [33].

• Sensing Area of the Sensor: Size and placement of the sensor are also critical, as shown in Figure 8. As a large sensor may underestimate the peak pressure and it is suggested that a minimum sensor of 5 mm × 5 mm should be used, whereas sensors smaller than this must be designed as array sensors.

• Operating Frequency: It is recommended [31] that to measure foot plantar pressure precisely for running activities the sensors must be capable of sampling at 200 Hz. This frequency is generally considered sufficient for sampling most everyday gait activities.

• Creep and Repeatability: Creep is the deformation of material under elevated temperature and static stress. It directly relates to the time dependent permanent deformation of materials when subjected to a constant load or stress [34], as in Figure 9. Low creep sensors are one of the key requirements in foot pressure measurement. Repeatability refers to the ability to produce reliable result even after long period of time [29]. High cyclic loads may cause deformation or fatigue [34]. Repeatability problems can be eliminated if the sensor exhibits no creep or deformation over repetitive or high cyclic loads.

5.1. Capacitive Sensors

The sensor consists of two conductive electrically charged plates separated by a dielectric elastic layer. Once a pressure is applied the dielectric elastic layer bends, which shortens the distance between the two plates resulting in a voltage change proportional to the applied pressure [11,31]. Figure 10 shows the capacitive sensor construction. Commercial products based on this system are the emed^® platform systems (Novel, Germany) [24] and Pedar^® in-shoe systems (Novel, Germany) [24].

5.2. Resistive Sensors

Force-Sensing Resistor (FSR) is a good example of the resistive sensor. When pressure is applied the sensor measures the resistance of conductive foam between two electrodes. The current through the resistive sensor increases as the conductive layer changes (i.e., decreases resistance) under pressure. FSRs are made of a conductive polymer that changes resistance with force, applying force causes conductive particles to touch increasing the current through the sensors [11,31]. Figure 11 shows the resistive sensor construction and commercial products based on this principle are MatScan^® platform systems (Tekscan, USA) [26] and F-Scan^® in-shoe systems (Tekscan, USA) [26].

5.3. Piezoelectric Sensors

The sensor produces an electric field (voltage) in response to pressure. Piezoelectric devices have high impedance and therefore susceptible to excessive electrical interference that leads to an unacceptable signal-to-noise ratio.

The most suitable material for clinically oriented body pressure measurement is polyvinylidene fluoride (PVDF) because it is flexible, thin and deformable [11,31]. Figure 12 shows the piezoelectric sensor construction. Commercial products based on this system are Measurement Specialties, USA [38] and PCB Piezotronics, Inc., USA [39].

5.4. Piezoresistive Sensors

This sensor is made of semiconductor material. In Piezoresistive material the bulk resistivity is influenced by the force or pressure applied, when the sensor is unloaded resistivity is high and when force is applied resistance decreases [11]. Figure 13 shows the piezoresistive sensor construction. When there is pressure on the piezoelectric element (quartz crystal) it produces electric charges from its surface. These charges create voltage proportional to the applied force. Commercial products based on this system are FlexiForce^® (Tekscan, USA) [26] and ParoTec (Paromed, Germany) [40].

The requirements of the pressure sensor for the specific application are low hysteresis, linearity of output, and pressure range [27,29,31]. A recommended pressure range for gait analysis (walking) is approximately 1,000 kPa [31] but for sports the pressure range should be larger due to the nature of the movements. There are a number of commercially available foot-pressure sensors on the market but they generally do not fulfil the requirements of many biomechanical applications due to specification and performance limitations. The limitations include but not limited to, the specified hysteresis [29], pressure span and physical sensor dimensions [42].

In comparison with traditional foot plantar pressure sensors such as capacitive sensors, resistive sensors, piezoelectric sensors and piezoresistive sensors the MEMS pressure sensors have many advantages. For example, easy communication with electrical elements in semiconductor chips, small size, lower power consumption, low cost, increased reliability and higher precision. To provide a better alternative developments of a specifically designed miniature foot pressure sensor based on MEMS technology have been explored. In response to the needs of such sensors, Wahab et al. [43] successfully designed, fabricated and tested a miniature foot pressure sensor based on MEMS technology that can be inserted in the insole of a shoe [43]. As reported by Wahab et al. [43] significant performance enhancements have been achieved, for example, the sensor is small, has high-pressure range measurement capability, and excellent linearity both at low and high pressures and possesses negligible hysteresis.

Currently available in-shoe pressure sensor parameters are compared to Wahab et al. [43] in Table 1. Sensors from Vista Medical, Novel and Tekscan show some performance limitations as they are made of sheets of polymer or elastomer leading to issues such as repeatability, hysteresis, creep and non-linearity of the sensor output [29]. In addition to the above limitations some sensors (e.g., Parotec) have limited pressure range and relatively large dimensions. Figure 14 demonstrates the linearity of the MEMS based pressure sensor and the fabricated sensor as displayed in Figure 15.

6.1. Wired Systems Application

Over the past two years there has been increasing interest in developing in-shoe foot plantar pressure systems and recently there have been applications to plantar pressure using both wired and wireless systems. In 2011, a paper employed dynamic plantar pressure for human identification using a FlexiForce^® (Tekscan, USA) in-sole pressure sensor [5]. They compared the pressure at different positions of key points then identified and classified them using a support vector machine (SVM) running on a PC. The system uses wire to transfer data from the sensor to a data acquisition card on a PC (Figure 16) and it is reported that the system has 96% identification accuracy.

Yamakawa et al. [6] also proposed their own biometric identification in-shoe system based on both feet pressure change and reported that the system could recognize over 90% of the test subjects. The system used F-scan (Nitta Corp, Japan) as the pressure sensor (see Figure 17).

Another innovative application is an in-shoe system to measure triaxial stress in high-heeled shoes [50]. The paper investigated the distribution of contact pressure and sheer stress simultaneously in high-heeled shoes utilizing five in-shoe triaxial force transducers commercialized by Anhui June Sport, China.

Shear stresses can cause blisters, callosities and trophic ulcers. The size of transducer is 17 mm × 18 mm × 10 mm and has 870 kPa full scale pressure range. In the system the transducer in mounted under the hallux, the first, second and fourth metatarsal heads as well as the heel as shown in Figure 18. As can be seen from the figure, peak sheer stress occurs at the second metatarsal; this type of information can be useful for future high-heeled shoe design.

Work undertaken by Healy et al. [51] claimed that their in-shoe system has a better repeatability compared to other commercially available systems. The sensor is practically similar to F-Scan^® (Tekscan, USA) which uses resistive force sensor.

The system is named “WalkinSense” and consists of a data acquisition and processing unit and eight individual sensors. It appears that only the sensor part is their own development, the rest of the system is similar to F-Scan^® (Tekscan, USA) hardware and software. The location of the sensors is illustrated in Figure 19, whilst the WalkinSense® System is shown in Figure 20.

All of the works described above [5,6,50,51] have a common feature that is wired to the processing unit or PC. All of them have certain benefits but for a wired system the major limitation is application in everyday monitoring. The wired system may encumber the test subject causing trip hazards or even a fall, and it can also affect the normal gait patterns. Therefore, it is recommended to make the system mobile for everyday usage, and the system must adapt to a wireless system. Research presented in [50,51] developed their own transducers but seems to have limited number of sensor placement. As mentioned in section IV, pressure recording from 15 locations is regarded as ideal for gait analysis. Papers [5,6] on the other hand used off-the-shelf sensors, which has also limitations as highlighted in Section 5.

6.2. General Wireless Systems Application

As mentioned earlier, there is a need for a system that can provide a wireless, real-time and reliable result to measure foot plantar pressure. There have been quite a few works undertaken in both research and commercial platforms that focused on developing a more mobile method of measuring foot pressure. Shoe based systems have been increasing, in terms of the number of publications and commercial products, due to shrinking in size of sensors, processing unit communication device and data storage. Further, the obvious reason for its usefulness is that such a system could measure the pressure distribution directly beneath the foot.

The work undertaken by Bamberg et al. 2008 [27] from Massachusetts Institute of Technology (MIT), had been mentioned in a large number of the papers that were reviewed. The main reason why Bamberg et al. have received so much attention in the literature is that they had come up with arguably the most complete wireless in-shoe system for gait analysis to date. They called it GaitShoe. In their system, the sensors included three orthogonal accelerometers, three orthogonal gyroscopes, four force sensors, two bidirectional bend sensors, two dynamic pressure sensors and electric field height sensors. The devices were capable of detecting heel-strike, estimating foot orientation and position and toe-off. The microcontroller (Silicon Laboratories), RF Monolithic (as the transceiver), antenna and power supply were also attached to the shoes. Figure 21 displays the GaitShoe with all the hardware mounted on the shoes.

In 2009, Benocci et al. [52] from University of Bologna, Italy developed a wireless system for gait and posture analysis. The wearable system utilised 24 hydrocells (by Paromed) to measure the plantar pressure and inertial measurement unit (IMU) in each shoe insole. The IMU integrated a 3-axes accelerometer and a digital 3-axes gyroscope. To control the system, Texas Instrument MPS430 microcontroller was implemented and Bluetooth acted as the transceiver. The collected data from the sensor allowed the user to recognize walking phases such as swing and stance, step and stride duration, double support and single duration. Figure 22 shows Benocci et al. wireless gait shoe.

Ming Young Biomedical Corp., Taiwan published a state-of-the-art digital textile sensor for measuring gait analysis [53]. Four dome shaped sensors were knitted on each sock. The dome shape sensors were able to record spatio-temporal plantar pressure patterns which were used to calculate the centre of pressure (COP) excursions. Five clip type sensors were sewn to the pant to record lower limb movement. The system was reported to measure the duration of stride cycles and left/right steps, cadence, walking speed, and COP. The microcontroller (Texas Instrument MPS430) and Bluetooth were attached to the wearer's belt. Figure 23 portrays the digital textile sensors in action.

Shu et al. [30] developed an in-shoe plantar pressure measurement and analysis system based on fabric pressure sensing array in collaboration with Hong Kong Research Institute of Textiles and Apparel Ltd. The sensors used were textile fabric sensor array, which is soft, light and has high pressure sensitivity. The sensors were connected with a soft polymeric board through conductive yarns and integrated into the insole.

Sensors were attached to six locations in the insole, as shown in Figure 24. The microcontroller PIC18F452 and the Bluetooth module were attached to the ankle of the patient. The system could interface with desktop, laptop and smart phone and was able to calculate parameters such as mean pressure, peak pressure, COP and shift speed of COP. The results were presented for both static and dynamic measurement conditions. Figure 25 shows the in-shoe plantar pressure measurement and analysis system based on fabric pressure sensing array.

The developments of wearable wireless sensor system for measuring foot plantar pressure have been encouraging. There is no doubt about their application potentials, especially the biomechanics communities. Nearly all use off-the-shelf sensors, microprocessors and wireless transmitters, so the end product is bulky and not comfortable to wear by the patients. The digital textile sensors by Chang-Ming et al. [53] are small and really flexible but it is wired to a Bluetooth based transmitter device at the belt. Both Benocci et al. [52] and Lin Shu et al. [30] used Bluetooth modules to attach to the ankle. Even more uncomfortable would be the system proposed by Bamberg et al. [27] where the whole sensor and wireless communication tools are attached to the top of the shoe. Although Bamberg et al. has developed the in-shoe gait analysis system but the system is not wearable for daily activities monitoring.

6.3. Major Application Areas

6.4. Sport Applications

Another application that relies on a wireless system is sport application. Noteworthy mentions are research by Salpavaara et al. [55] and Holleczek et al. [56]. These two papers employed innovative new sensors for their application using custom made laminated capacitive sensor matrix and textile pressure sensors respectively. Salpavaara et al. system can be utilized to monitor the timing and movement of the legs of the athlete during throwing, jumping and running in various sports events. The obtained data can be used for improving sports coaching. They opt for javelin throwing for their case study. In their case study they conclude that the timing of the steps, support and release phase has a great importance in the performance and the pace of steps should increase towards the end of the throw event. In their system they employ a capacitance-to-digital converter (Analog Device, USA), microcontroller (Atmel, USA) and a Zigbee-compliant radio. Figure 29 shows the five sensors placement. Holleczek et al. developed “SnowPro”, a wearable sport trainer, capable of supporting snowboarders in improving their skill. The system is able to analyze the dynamics of the weight distribution inside the boots. This type of information is essential for identifying the wrong weight shifting techniques, which usually lead to painful crashes in snowboarding sport. The system gives feedback to the user in real-time or after the activity about user performance and support user during learning process. This system utilizes three integrated textile pressure sensors, six capacitance-to-digital converters and a Bluetooth module. Figure 30 displays the final design by Holleczek et al. [56].

The obvious deficiency in the system is the number of sensors used. Salpavaara et al. [55] used five and Holleczek et al. [56] used only three. For many sport biomechanics applications this number may not be adequate. Besides that, for sport applications the system should not obstruct the athlete's movement. From Figure 30(b) it is apparent that it is not very practical. This is due to the fact that both systems utilized off-the-shelf equipment that is usually very bulky.