EP4704691A1

EP4704691A1 - Method of calibration of a gait and posture analysis system comprising a sensorized insole and the corresponding gait and posture analysis system

Info

Publication number: EP4704691A1
Application number: EP24729910.0A
Authority: EP
Inventors: Italo BELLI; Ines SORRENTINO; Gianluca MILANI; Enrico VALLI; Daniele PUCCI
Original assignee: Fondazione Istituto Italiano di Tecnologia
Current assignee: Fondazione Istituto Italiano di Tecnologia
Priority date: 2023-05-03
Filing date: 2024-04-19
Publication date: 2026-03-11
Also published as: WO2024228079A1; IT202300008673A1; CN121057543A

Abstract

The present invention relates to a method (2000;3000;4000) for calibrating a posture/ gait analysis system (1). The system (1) comprises an insole (2) equipped with a plurality of pressure sensors (4) of the capacitive type and a control unit (5) connected to the pressure sensors (4). The method comprises the following steps. Initially, signals (2000) are acquired provided by the pressure sensors (4) subjected to a pressure varying between a first value and a second value, wherein each signal is indicative of a capacity value of the sensor which depends on the pressure applied thereto. For each pressure sensor, a polynomial relation is to be defined between an estimate of the pressure experienced by the sensor (4) and a change in the sensor's capacity; a plurality of sets of coefficients of the polynomial minimizing a norm of the residual between the pressure estimate and the corresponding pressure applied on the sensor (4) are calculated (3001- 3011), wherein each set of coefficients is calculated for a respective order of the polynomial relation between a minimum order and an optimized maximum order. A mean of the minimisation results obtained by each sensor for the same order of the polynomial relation is then performed (3011). The set of coefficients and the respective order of the polynomial relation associated with the minimum value between the calculated means are selected (3012, 3013) and stored (3014, 3015) in the control unit (5).

Description

METHOD OF CALIBRATION OF A GAIT AND POSTURE ANALYSIS SYSTEM COMPRISING A SENSORIZED INSOLE AND THE CORRESPONDING GAIT AND

POSTURE ANALYSIS SYSTEM

DESCRIPTION

TECHNICAL FIELD

The present invention relates to the field of wearable electronic devices. In greater detail, the present invention relates to a method of calibrating a gait and posture analysis system comprising a sensorized insole equipped with a plurality of capacitive sensors distributed on the sole of the foot, in order to measure pressure distribution and thereby analyse gait and posture. The present invention also relates to a gait and posture analysis system calibrated by such a method.

STATE OF THE ART

Gait and posture analysis are very important in the health sector because of the close relationship between gait and posture with certain diseases and their progression. For example, real-time monitoring of the gait of a Parkinson's patient allows the progress of the disease to be monitored.

In addition, gait analysis has become an important aspect in the sports sector, particularly in competitive sports. In fact, gait analysis makes it possible to correct any defects in the athlete's gait, improving their performance.

In order to provide information on a subject's gait, several solutions have been proposed in literature involving sensorized insoles, in particular those that are sensitive to pressure, i.e. a force applied to a surface of the insole and usually transverse to this surface.

Sensorized insoles are much more versatile than static force plates for the analysis of gait and posture, particularly in the case of the analysis of highly dynamic movements, such as during sports activities, which would require plates with a very large surface area to perform an adequate analysis.

One of the most common technologies used for underfoot pressure sensors in insoles is based on capacitive transducers. Capacitive transducers comprise a pair of plates separated by a dielectric so that their capacitance value varies according to the distance between the capacitor plates. Consequently, the capacitance value of each capacitive transducer comprised in an insole varies as a function of the pressure (i.e. a weight force) applied by the foot on the transducer itself. For example, the capacity can be described as: where C is the capacitance, G is the permittivity of the dielectric, A is the overlap area of the two plates of the generic capacitive transducer and d is the distance between them. Therefore, when the capacitive transducer is compressed, the distance d decreases and the capacitance C increases.

Calibrating capacitive transducers is, however, not a trivial task. In fact, the relationship between capacitance and pressure is only theoretically linear while, in real applications, the relationship between capacitance and pressure is influenced by strongly non-linear effects - for example, hysteresis of the materials constituting the insole.

In an attempt to solve this problem I. Gonzalez, J. Fontecha, R. Hervas, and J. Bravo: "An Ambulatory System for Gait Monitoring Based on Wireless Sensorized Insoles" Sensors 15, no. 7: 16589-16613, 2015, proposes a calibration method in which a user, who is wearing the sensorized insoles, must first keep his foot elevated and then rest his foot on the ground. While the user does this, the minimum and maximum measurement values provided by each transducer in the insole are determined. The measurement values thus obtained are then used as reference values during the use of the sensorized insoles.

The proposed calibration method must be carried out while the sensorized insoles are being used by the user. In other words, the calibration operation can only be performed after the sensorized insoles are worn by the end user and requires that the end user performs the calibration operation and is equipped with appropriate instrumentation to perform the calibration operation and apply modifications to the electronics of the sensorized insoles.

OBJECTS AND SUMMARY OF THE INVENTION

An object of the present invention is to overcome the drawbacks of the prior art.

In particular, it is the object of the present invention to provide a practical and reliable calibration method for a gait/posture analysis system comprising an insole equipped with pressure sensors - for example, capacitive transducers.

A further object of the present invention is to propose a calibration method for a gait/ posture analysis system comprising an insole equipped with pressure sensors that lends itself to both industrial implementation, for large-volume production, and customized implementation, for the production of an analysis system suitable for a specific user.

A further object of the present invention is to present a gait/ posture analysis system comprising an insole equipped with pressure sensors - e.g., capacitive transducers - capable of providing an accurate and reliable indication of the pressure - i.e., weight force - applied at a plurality of insole positions.

These and other objects of the present invention are achieved by a system incorporating the features of the accompanying claims, which form an integral part of the present description.

According to a first aspect, the present invention concerns a method for calibrating a posture/ gait analysis system. The system comprises an insole equipped with a plurality of pressure sensors - i.e. a weight force - of the capacitive type and a control unit connected to the pressure sensors. The method comprises the following steps. Initially, signals provided by the pressure sensors subjected to a pressure varying between a first value and a second value are acquired, wherein each signal is indicative of a capacity value of the sensor which depends on the pressure applied thereto. For each pressure sensor, a polynomial relation is defined between an estimate of the pressure experienced by the sensor and a change in the sensor's capacity; a plurality of sets of coefficients of the polynomial minimizing a norm of the residual between the pressure estimate and the corresponding pressure applied on the sensor are calculated, wherein each set of coefficients is calculated for a respective order of the polynomial relation between a minimum order and an optimized maximum order. A mean of the minimization results obtained by each sensor for the same order of the polynomial relation is then performed. The set of coefficients and the respective order of the polynomial relation associated with the minimum value between the calculated means are selected and stored in the control unit.

The method according to the present invention makes it possible to obtain a particularly accurate estimate of the pressure, i.e. the weight force, acting on the insole sensors. In particular, the use of a polynomial relation involving the definition of the coefficients and degree of polynomials according to the present invention allows a desired accuracy to be achieved with the lowest possible computational complexity.

In an embodiment, the step of defining the polynomial relation involves defining the polynomial relation as: where Pt (C_{i t}) is a pressure estimate based on the signals provided by the z-th sensor at the time instant f, n (C_{i t}) is a first polynomial of order n_p defined as: where G,t is the capacitance value of the sensor i at the instant t and ao,o-np are the polynomial coefficients, is a second polynomial of order n_ps defined as: where Q,t-k is the capacitance value of the sensor i at the instant tk, preceding the instant f, ak,i-nps are the polynomial coefficients and n_s is a memory parameter indicative of the number of polynomials h_k{C_{i t-k}) associated with corresponding instants tk preceding the time instant t comprised in the polynomial relation Pi(C_{i t}).

The polynomial relationship defined in this way makes it possible to estimate the pressure currently exerted on the insole sensors by taking into account the effect of the pressure applied on the insole in a period of time prior to the current instant. In other words, the use of this polynomial relation for the pressure calculation makes it possible to include "memory" effects due to the materials used in the construction of the insole and/ or the pressure sensors comprised therein in the pressure estimation in a simple way.

In an embodiment, the step of calculating a plurality of sets of coefficients of the polynomial involves reducing the polynomial relation to: and determining an optimized maximum order value of the order n_p for each sensor by finding the lesser value of: where ... a_{O np}] is ^a vector of polynomial coefficients, N_SSns is the number of sensors in the insole, P is a vector of effective pressure values exerted on the sensor comprised between the first value and the second value, <t>j is a regression matrix defined as: where d is a number of time instants f in which the effective pressure values are acquired and is the regressor at time instant t defined as: and n_p\max is a predetermined maximum order value. Preferably, the predetermined maximum order value is twenty-five.

The identification of a maximum degree performed in this way makes it possible to simply and effectively reduce the complexity of the complete polynomial relation for which a set of coefficients and their respective order must be defined, simplifying the identification of the coefficients and order that allow for the optimization of the estimate of the pressure on the insole.

In an embodiment, the step of determining an optimized maximum order value for the order n_p involves checking whether for order values n_p comprised in a predetermined sub-range there is a local minimum residual norm value, if the negative case, selecting the minimum residual norm value from among the calculated values as the optimized maximum order value, or in the affirmative case, checking whether for order values n_p outside the sub-range there is a residual norm value lower, by at least one threshold value, than the local minimum of the residual norm value, in the affirmative case, selecting the residual norm value lower than the local minimum residual norm value as the optimized maximum order value, or in the negative case, selecting the local minimum residual norm value as the optimized maximum order value.

Preferably, the predetermined sub-range comprises order values n_p comprised between 1 and 10.

In addition or alternatively, the step of imposing an optimized maximum order value of order n_ps of the polynomial h_k(C_{i t}-k) equal to the optimized maximum order value of the order Tip of the polynomial n

Through this method of determining the optimized maximum orders n_p and n_ps of the polynomials comprised in the polynomial relation, it is possible to substantially reduce the computational complexity required to determine the coefficients and order of the polynomial relation, which allows the optimization of the insole pressure estimate without significantly reducing the accuracy of the estimate.

One embodiment comprises the step of imposing a maximum memory parameter value equal to a number of signals acquired by each sensor in an acquisition time period. Preferably, the acquisition time period is one second.

By selecting the value of the memory parameter in this way, it is possible to introduce the amount of information required to evaluate the impact of the pressure applied on the insole previously on the signals generated by the sensors without substantially increasing the complexity of the polynomial relation and, therefore, the computational load required to determine the coefficients and order of the polynomial relation that allow the optimization of the estimate of the pressure on the insole.

In an embodiment, the step of performing a mean of the minimization results comprises calculating the mean of the minimization results obtained by each sensor for the same order of the polynomial as: where a vector of polynomial coefficients, P is a vector of effective pressure values exerted on the sensor comprised between the first value and the second value, <I\ is a regression matrix defined as: where d is a number of time instants f in which the effective pressure values are acquired and (faCCt) is the regressor at time instant t defined as: and wherein n_p, n_ps, n_s = max indicates that all combinations of the order values n_ps of the polynomial /ifc(Qt-fc) and the order values n_p of the polynomial n (Q _t)are considered, comprised between one and the optimized maximum order value, and of the memory parameter values n_s, comprised between one and the maximum memory parameter value n_s.

Calculating the mean of the minimization results described above makes it possible to determine the coefficients and order of the polynomial relationship that optimize the insole pressure estimate quickly and easily.

According to a different aspect, the present invention concerns a posture/ gait analysis system. The system comprises an insole equipped with a plurality of capacitive pressure sensors and a control unit connected to the pressure sensors. The control unit stores a set of polynomial coefficients and an order of a plurality of polynomial relations each associated with a corresponding sensor. The calculation of each polynomial relation allows the control unit to convert a change in capacity of a respective sensor into an estimate of the pressure experienced by the sensor.

Preferably, the plurality of polynomial relations calculated by the memory unit are of the form: where P_£ (Q_t) is a pressure estimate based on the signals provided by the z-th sensor at the time instant f, n (C_{i t}) is a first polynomial of order n_p defined as: where Ci,t is the capacitance value of the sensor i at the instant t and ao,o-n_P are the polynomial coefficients, is a second polynomial of order n_ps defined as: where Q,t-k is the capacitance value of the sensor i at the instant fit, preceding the instant f, ak.'i-nps are the polynomial coefficients and n_s is a memory parameter indicative of the number of polynomials h_k(C_{i t-k}) associated with corresponding instants tk preceding the time instant t comprised in the polynomial relation

The system according to the present invention achieves the same advantages as outlined above for the corresponding method of calculation and polynomial coefficients and the order of the plurality of polynomial relations.

Further features and objects of the present invention will be more evident from the description of the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

The invention will be described below with reference to some examples, provided for explanatory and non-limiting purposes, and illustrated in the accompanying drawings. These drawings illustrate different aspects and embodiments of the present invention and reference numerals illustrating structures, components, materials and/or similar elements in different drawings are indicated by similar reference numerals, where appropriate.

Figure 1 is a diagram of the posture/ gait analysis system according to an embodiment of the present invention;

Figure 2 is a block diagram of a calibration arrangement of the system in Figure 1;

Figure 3 is a flow chart of a dataset acquisition procedure comprised in a calibration method according to an embodiment of the present invention;

Figure 4A is a graph representing the variation of a pressure sensor capacitance comprised in the insole of the system in Figure 1 as a function of time;

Figure 4B is a graph depicting the variation of pressure as a function of time within a pressurized chamber of the arrangement in Figure 2;

Figure 5 is a flow chart of a process for identifying calibration parameters of the analysis system of Figure 1 comprised in the calibration method according to an embodiment of the present invention;

Figure 6 is a graph of the mean trend of the minimum values of the mean residual norm of the pressure estimated based on the measurements provided by a sensor as a function of the degree of a polynomial used in the pressure estimate performed by the analysis system in Figure 1;

Figure 7 is a graph of the mean trend of the minimum residual norm value as a function of other parameters used in the pressure estimate performed by the analysis system in Figure 1, and

Figure 8 is a flow chart of a verification procedure for a calibration matrix comprised in the calibration method according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

Some preferred embodiments will be described in detail below, although the invention is susceptible to various alternative modifications. It must in any case be understood that there is no intention to limit the invention to the specific embodiment illustrated, but, on the contrary, the invention intends to cover all the modifications and/or use of the elements or equivalent steps that fall within the scope of the invention as defined in the claims.

Unless otherwise defined, all the terms of the art, notations and other scientific terms used herein are intended to have the meanings commonly understood by those skilled in the art to which this description belongs. In some cases, terms with commonly understood meanings are defined herein for clarity's sake and/or ready reference; the insertion of such definitions in the present description must therefore not be interpreted as representative of a substantial difference with respect to what is generally understood in the art.

The terms "comprising", "having", "including" and "containing" are to be understood as open terms (i.e. the meaning "comprising, but not limited to") and are to be considered as a support also for terms such as "essentially consist of", "essentially consisting of", "to consist of" or "consisting of".

The use of "for example", "etc.", "or" denotes non-exclusive alternatives without limitation, unless otherwise indicated. The use of "includes" means "includes, but not limited to" unless otherwise indicated.

Referring to Figure 1, a posture/ gait analysis system, the analysis system 1 in the following, comprises a sensorized insole, hereafter referred to simply as insole 2, according to an embodiment of the present invention and a processing unit 3.

The insole 2 is constructed as a multilayer element comprising a plurality of pressure sensors, hereafter referred to simply as sensors 4. The insole 2 is shaped to fit into a shoe. Therefore, the insole 2 has an elongated shape and comprises with an enlarged front area, intended to receive the forefoot, with respect to the rear area, intended to receive the heel. The sensors 4 are electrically connected to a connector 5 formed in a side tongue 6 of the insole 2. In the example in Figure 1, two connectors 5 are formed in the side tongue 6 of the insole 2. A subset of sensors 4 is connected to each connector 5 via appropriate wiring, omitted for simplicity in Figure 1.

The insole 2 is connected to a processing unit 3 by means of wiring 7 connected to connectors 5 and is adapted to transmit to the processing unit 3 the signals generated by the sensors 4 of the insole 2 during its use.

The processing unit 3 comprises an acquisition module 31 and a data processing module. The acquisition module 31 receives the signals generated by the sensors 4 and converts them into a digital signal which is processed by the data processing module 32 in order to provide useful information for analysing the posture and/or gait of a corresponding user (not illustrated). For example, the acquisition module 31 comprises - in a non-limiting manner - one or more analogue-to-digital conversion (ADC) devices, filters, amplifiers, etc. Conversely, the processing module 52 comprises - in a nonlimiting manner - one or more processors, DSPs, FPGAs, ASICs, etc. configured to perform the desired signal processing.

In summary, the insole 2 and the processing unit 3 form the analysis system 1 adapted to acquire information relating to an applied force (i.e., a pressure) at a plurality of positions on the surface of the insole (i.e., positions in which a sensor 4 is present) and determine useful information about the posture and/ or gait of a user using the insole 2 (e.g., inserted in a shoe).

The analysis system 1 just described is calibrated by means of the calibration method according to an embodiment of the present invention.

In particular, the calibration method comprises a process 2000 for acquiring a dataset using a test arrangement 7 of which Figure 2 is a block diagram.

The test arrangement comprises a pressurized chamber 71 - for example, a watertight steel tank capable of withstanding a maximum pressure of 10 bar -, an air compressor 72 - for example, with a capacity of 24 litres -, a pressure regulator 73 - preferably, closed-loop -, a pressure sensor 74 and a controller 75.

In particular, the air compressor 72 is in fluid communication with the pressurized chamber 71 via a first conduit 76 connected to the pressure regulator 73 and a second conduit 77 connecting the pressure regulator 73 to the pressurized chamber 71, while the pressure sensor 74 is installed on the second conduit, preferably in the vicinity of the pressurized chamber 71. The controller 75 is electrically connected to both the pressure regulator 73 - for setting the desired pressure in the pressurized chamber 1 - and the pressure sensor 74 - for acquiring the effective pressure measurement inside the pressurized chamber 1.

The process 2000, of which Figure 3 is a flowchart, involves placing insole 2 inside the pressurized chamber 1 (step 2001).

The pressure inside pressurized chamber 1 is varied according to a predetermined pattern (step 2002). In a preferred embodiment, the pressure values are varied linearly from a minimum pressure value P_mm to a maximum pressure value PMAX. Preferably, the minimum pressure value P_min is set equal to 0 bar (0 Pa) - i.e., it corresponds to a condition in which no force is applied to the insole 2. Otherwise, the maximum pressure PMAX is selected to simulate an applied force equal to the pressure exerted by a user with a predetermined weight. For example, a maximum pressure value PMAX comprised between 3 bar (0.3 MPa) and 4 bar (0.4 MPa), such as 3.4 bar (0.34 MPa) can simulate a uniform pressure on the surface of the insole 2 corresponding to a weight force of approx. 75 kg to 85 kg, as approximately 80 kg, applied to an area of the insole - preferably, the area of the insole 2 intended for the toes - comprised between approximately 20 cm² and 30 cm², as approximately 25 cm². The Applicant found that this maximum pressure value PMAX allows for the optimization of the calibration of insoles intended for users weighing between 75 kg and 85 kg.

While the pressure is varied between Pmin and PMAX, the measurements provided by each of the sensors 4 of insole 2 are acquired (step 2003). In particular, the measurements provided by the sensors 4 will be proportional to the current value of the capacitance of each sensor 4, each of which depends on the variation of the distance between the plates of the sensor 4 which in turn depends on the force exerted thereon, as can be observed in the qualitative graphs of Figures 4A and 4B, which represent the trend of the capacitance Cs of a sensor 4 as a function of time and the trend of the pressure P as a function of time inside the pressurized chamber 1, respectively, during the execution of the process 2000.

In an embodiment, the insole 2 is connected directly to the controller 75 instead of the controller 3 - for example, by connecting the connectors 5 of the appropriate ports of the controller 75.

Preferably, the measurements provided by each sensor 4 of the insole 2 are processed - e.g. filtered - in order to remove noise and spurious peaks (step 2004).

The measurements provided by each sensor 4 of the insole 2 are stored by the controller 75 associated with the corresponding pressure detected by the pressure sensor 74 of the arrangement 7 (step 2005). For example, when a pressure increment AP is selected, for every pressure increment AP in the pressurized chamber 71 a set of measurements Ms(P) comprising the measurements provided by each sensor 4 of the insole 2 is acquired and stored. In addition, the corresponding pressure value P is also acquired and stored. In particular, each set of measurements Ms(P) is associated with the corresponding pressure value P in the pressurized chamber 71. The sets of measurements Ms(P) and the corresponding pressure values comprised between Pmm and PMAX are grouped together to form a dataset DM of reference measurements. In the example considered, the dataset DM comprises d sets of measurements Ms(P) associated with d corresponding pressure values in the pressurized chamber 71 - where d = (PMAX- Pmm)/AP each measured by the pressure sensor 74 of the arrangement 7 at a corresponding instant of time f (with fe HV).

The method comprises an optimization procedure 3000, of which Figure 5 is a flow chart. In particular, the process 3000 aims to determine the optimal values of a calibration matrix Me for the insole 2. The calibration matrix Me comprises the parameters defining the function linking estimated pressure P by the sensors 4 and capacity, i.e:

In particular, the Applicant has determined that the relationship between estimated pressure P and capacity variation can be represented by means of a polynomial function defined as: where P_t (C_{i t}) is the estimated pressure based on information provided by the i-th sensor 4 of the insole 2 at time instant f, is a polynomial of order n_p that links the instantaneous capacity value Q,t to the estimated pressure P, is a polynomial of order n_ps that improves the accuracy of the estimated pressure P by taking into account information associated with the previous state of the sensor 4 of the insole 2 at time instant tk, where fa = (t-k), and n_s is a parameter that defines the number of previous states of the sensor 4 of the insole 2 that are considered.

The calibration matrix Me thus contains the values of the polynomial coefficients a_jy of the polynomials n and h_k(C_{i t-k}^) - where the generic polynomial coefficient a_jy is associated with the /-th instant of time f (comprised between 0, i.e. the current instant of time, and J, i.e. the most distant instant of time from the current instant of time) and to the i/-th order of the polynomial function (comprised between 0 and n_p or between 0 and n_ps).

In detail, in order to populate the calibration matrix Me, the process 3000 initially involves setting the order value n_ps and the parameter n_s equal to zero (i.e., n_ps = 0, n_s = 0; step 3001), reducing the polynomial function (1) to:

A set of optimization operations is then performed on a range of values of order n_p sufficiently large to determine the polynomial coefficients aj _y of the polynomial n (step 3002). In a preferred embodiment, the range of values of order n_p comprises integers between 1 and 25 (i.e., n_p = [1, 25]). In particular, the Applicant determined that selecting 25 as the maximum value of the order n_p allows for optimal calibration of different types of insoles and regardless of the number of sensors installed on each insole, without requiring particularly high computational resources.

The polynomial coefficients a_} -_y of the polynomial n (C_i/t) are determined using the least-squares approach, in particular, by minimizing a norm of the residual between the pressure estimate calculated by the system and the corresponding pressure value inside the pressurized chamber 71, hereafter residual norm for brevity, for different values of order n_p within the selected range.

In the case considered, where the order value n_ps and the parameter n_s is zero, formula (6) can be rewritten as: where (p_t(C_L) G IR^rap⁺¹ is the regressor at instant t for the z-th sensor 4: whereas G IKⁿP⁺¹ is the vector of coefficients for the z-th sensor 4: Furthermore, considering the dataset DM of reference measurements acquired on d sampling instants, an overdetermined system is defined which defines the relationship between the pressure measured inside the pressurized chamber 71 and the corresponding pressure estimates calculated from the sensor signals 4: (10) where P e HW is a vector comprising the pressure measurements in the pressure chamber 71 measured by the pressure sensor 74, and <I\ G j^a.np+i j_s ^g _reg_ression matrix for the z-th sensor 4 and is defined as:

(11)

Consequently, the polynomial coefficients a.jy, i.e. a vector 2/, are determined by calculating the residual rules: (12)

In other words, for each value of order n_p in the considered interval [1, 25] and for each sensor 4, the set of polynomial coefficients a_} -, _y of the polynomial n is identified, which minimizes the difference - i.e., the residual norm - between the pressure value estimated by the analysis system 1 and the pressure value inside the pressurized chamber 1 measured by the pressure sensor 74 of the arrangement 7.

Preferably, a single order value n_p is identified for all the sensors 4 by determining the mean minimum residual norms with order value n_p variable between 1 and 25:

(13) where N_S£-ns is the number of sensors in the insole 4.

Subsequently, the presence of a local minimum residual norm value is checked at mean order values n_p comprised within a predetermined sub-interval, e.g. comprised between 0 and 10, (decision step 3003).

In the negative case (output branch N of step 3003), the order value n_p is identified and selected as the maximum optimized value for the polynomial n that provides the minimum residual norm among those calculated (step 3004).

When a local minimum of the value of the residual norm is identified (output branch Y of step 3003 and as illustrated in the graph of Figure 6 for n_p = 7), the corresponding order value n_p is selected as the optimized maximum value (step 3005) and it is checked whether for order values n_p greater than the order values included in the sub-range (i.e, with order n_p comprised between 11 and 25 in the example considered) there is a residual norm with a value less than the local minimum of the residual norm value for at least one threshold value - for example, equal to 0.1 bar (0.1 MPa) (decision step 3006).

In the negative case (output branch N of step 3006), the order value n_p associated with the local minimum of the residual norm value is confirmed as the optimized maximum value (step 3007). Otherwise, if an order value greater than 10 associated with a residual norm lower than at least the local minimum threshold value of the residual norm value (output branch Y of step 3006) is identified, this order value n_p is selected as the optimized maximum value (step 3008).

Subsequently, the maximum order value n_ps of the polynomial h_k(C _pt-k) is set equal to the optimized maximum order value n_p just selected for the corresponding sensor 4 (step 3009). In fact, the Applicant has determined that it is possible to use the same maximum order value n_p used for the polynomial TI (C_{i t}), the selected process for the other polynomials and thus the entire calibration procedure without introducing substantial errors in the pressure estimate

The maximum value of the parameter n_s is selected according to a periodicity with which information is to be acquired by the sensors 4 (step 3010). For example, if we consider using the insole to analyse the posture and/or gait of a user, a periodicity of one second, with a sampling frequency of approximately 60Hz we set the parameter n_s to a maximum value of sixty (z.e., n_s \_max = 60).

Subsequently, the estimated pressure P is considered in its complete form defined by the formula (4) and a set of optimization operations is performed on intervals of the order values n_p, n_ps and of the parameter n_s comprised between zero and the maximum values selected above, in order to determine optimal polynomial coefficients ay_y of the polynomials n (C_{i t}) and h_k(C_{i t-k}) (step 3011).

In detail, it is provided to reapply the least squares approach, calculating the minimum residual norms for different order values n_p, n_ps and of the parameter n_s comprised in the respective ranges selected for each sensor (n_p, n_ps = [1; 7] and n_s = [1; 60] in the example considered) and then averaging the results for each sensor 4.

Specifically, in formula (7), when the order value n_ps and the parameter n_s are non-zero, we have that: E ]&ⁿp⁺ⁿps^xns⁺¹ is the regressor at instant t for the z-th sensor 4: whereas 2/ G vector of coefficients for the z-th sensor 4:

In this case, the regression matrix <I*iis comprised in ]R^d’ⁿP⁺ⁿps^xns⁺¹ _and _a p_adding of n_s elements in Ci is performed for consistency.

Consequently, the polynomial coefficients aj_y, i.e. the vector 2/, are determined by calculating the mean value of the minimum residual norms calculated for each sensor 4 of the insole 2:

In particular, the means of the values of the minimum residual norms for each combination of the order values n_p, n_ps and of the parameter n_s define a respective surface of solutions in a three-dimensional space in which two axes represent the order value n_ps and the parameter n_s, and the third axis represents the values of the residual norms, as illustrated in Figure 7.

The polynomial coefficients aj_y of the polynomials 7i (Q_£) and h_k(C_it-k)' associated with the mean of the values of the minimum residual norms with the lowest value of those calculated are selected (step 3012).

In addition, the order values n_p and n_ps and the parameter n_s associated with the mean of the values of the minimum residual norm with the lowest value of those calculated (step 3013) are selected.

Finally, the value of the polynomial coefficients aj _y of the polynomial the polynomial /i_fc(C_{i t-fe}) determined via the formulae (15) are stored in the calibration matrix Me (step 3014). In addition, the order values n_p and n_ps and the parameter n_s are also stored (step 3015).

To summarize, the calibration matrix Me, for each sensor 4 of the insole 2, comprises a data set that includes a set of polynomial coefficients aj,_y of the polynomials n and /i_fc(Q_t-k) for each sensor 4 that optimize the estimated pressure P_£ (Q_t) through the measurements provided by the corresponding sensor 4.

Preferably, the method includes a validation process 4000, of which Figure 8 is a flowchart, which aims to verify whether the performance of the analysis system 1 using the calibration matrix Me defined through the process 3000 conforms to minimum expected performance.

Initially, the insole 2 is positioned inside the pressurized chamber 1 (step 4001).

The pressure inside the pressurized chamber 1 is varied according to a predetermined pattern (step 4002), e.g. linearly between P_min and PMAX, in a manner similar to that described above in connection with the process 3000. As the pressure is varied between Pmm and PMAX_Z the measurements provided by the sensor 4 of the insole 2 are acquired (step 4003) and the corresponding measurement measured by the pressure sensor 74 of arrangement 7.

Preferably, the measurements provided by each sensor 4 of the insole 2 are processed - e.g. filtered - in order to remove noise and spurious peaks (step 4004).

The measurements provided by the sensors 4 are used to calculate the estimated pressure Pj (Q_t) associated with each sensor 4 using relation (2) on the basis of the data comprised in the calibration matrix Me (i.e. the angular coefficients) and the optimal values of lip, /Ips and n_s defined above (step 4005).

The residual norms between estimated pressure (C_{i t}) and measured pressure associated with each of the sensors 4 are then calculated (step 4006) and a mean residual norm associated with each sensor 4 is calculated by averaging all the residual norms calculated from the measurements provided by the same sensor 4 during the pressure variation from P_mm to PMAX (step 4007).

It is checked whether each mean residual norm is less than or equal to a threshold value (decision step 4008), e.g. equal to 0.2 bar (0.2 MPa).

In the affirmative case (output branch Y of step 4008), the calibration matrix Me is confirmed (step 4009) and is then permanently stored in the control unit 3 and used to estimate the pressure exerted by a user on the insole 2.

Conversely, if one or more residual norms are greater than 0.2 bar (output branch N of step 4008), the calibration matrix Me is not considered adequate, therefore, the calibration matrix Me is discarded and procedure 3000 or both procedures 2000 and 3000 are repeated in order to recalculate the data to be entered into the calibration matrix Me (step 4010).

However, it is clear that the above examples must not be interpreted in a limiting sense and the invention thus conceived is susceptible of numerous modifications and variations.

For example, as will be clear to a person skilled in the art, the insole may comprise a number of wiring harnesses other than two, for example a single wiring harness, or three or more wiring harnesses. In general, the number of wiring harnesses is selected according to the number of desired pressure sensor subsets and/or according to the specific implementations of the acquisition module comprised in the processing unit.

In alternative embodiments (not illustrated), the processing unit is integrated into the insole.

In other embodiments (not illustrated), the insole comprises a radio signal transmission module (i.e., a transmitter) or a radio signal transceiver module (i.e., a transceiver) and, similarly, the processing unit comprises a radio signal reception module (i.e., a receiver) or a transceiver. In this case, the insole's transceiver/ transmitter transmits pressure measurements from the sensors to the processing unit via a radio communication channel (e.g., RF, Bluetooth®, WiFi, Zigbee, etc.).

As will be apparent to the person skilled in the art, one or more steps of the procedures described above may be performed in parallel with each other or with an order different from that presented above. Similarly, one or more optional steps can be added or removed from one or more of the procedures described above.

In an alternative embodiment instead of determining a common optimized maximum value as described above, an optimized maximum value is calculated for each insole sensor by calculating minimum residual norm values with order value n_p varying between 1 and 25 using the coefficient values determined by means of formulae (12) a minimum residual norm value for each sensor 4 in a similar manner as described above in relation to steps 3003-3008 mutatis mutandis. In this way, a more accurate estimate is obtained at the cost of a higher computational load.

Furthermore, although the description refers to a single insole and the calibration of a single insole, there is nothing to prohibit performing the entire method or one or more of its component procedures in parallel on a plurality of insoles.

In an embodiment, once a calibration matrix has been calculated and confirmed for an insole, this calibration matrix is stored in a plurality of control units, each of which is connected to a respective insole of the same type.

Naturally, all the details can be replaced with other technically-equivalent elements.

In conclusion, the materials used, as well as the shapes and contingent dimensions of the devices, apparatuses and terminals mentioned above, may be any according to the specific implementation needs without thereby departing from the scope of protection of the following claims.

Claims

1. Method (2000;3000;4000) for calibrating a posture/ gait analysis system (1), the system (1) comprising an insole (2) equipped with a plurality of pressure sensors (4) of the capacitive type and a control unit (5) connected to the pressure sensors (4), where the method comprises the steps of: acquiring (2000) signals provided by the pressure sensors (4) subjected to a pressure varying between a first value and a second value, wherein each signal is indicative of a capacity value of the sensor which depends on the pressure applied thereto; for each pressure sensor: defining a polynomial relationship between an estimate of the pressure experienced by the sensor (4) and a variation of the sensor capacitance; calculating (3001-3011) a plurality of sets of coefficients of the polynomial minimizing a norm of the residual between the pressure estimate and the corresponding pressure applied on the sensors (4), where each set of coefficients is calculated for a respective order of the polynomial relation between a minimum order and an optimized maximum order; performing (3011) a mean of the minimization results obtained by each sensor for the same order of the polynomial relation , selecting (3012, 3013) the set of coefficients and the respective order of the polynomial relation associated with a minimum value among the means of the minimization results, and storing (3014, 3015) the set of coefficients and the respective order in the control unit (5).

2. Method (2000; 3000; 4000) according to claim 1, wherein the step of defining the polynomial relation provides for defining the polynomial relation as: where P; (Q_t) is a pressure estimate based on the signals provided by the /-th sensor (4) at the time instant f, zr (Q_t) is a first polynomial of order n_p defined as: where Ci,t is the capacitance value of the sensor i at the instant t and ao,o-np are the polynomial coefficients, h._k(Ct is a second polynomial of order n_ps defined as: where Q,t-k is the capacitance value of the sensor i at the instant ft, preceding the instant f, ak,i-nps are the polynomial coefficients and n_s is a memory parameter indicative of the number of polynomials /i_fc(Q_t-fc) associated with corresponding instants ft preceding the time instant t comprised in the polynomial relation Pi(Q_t).

3. Method (2000; 3000; 4000) according to claim 2, wherein the step of calculating (3001-3010) a plurality of sets of coefficients of the polynomial provides for reducing (3001) the polynomial relation to: and determining (3002 - 3008) an optimized maximum order value of the order n_p for each sensor (4) by finding the lesser value of: where = [a_{0 0} a_{0 1} ... a_{O np}] is ^a vector of polynomial coefficients, N_SSns is the number of sensors in the insole, P is a vector of effective pressure values exerted on the sensor (4) comprised between the first value and the second value, <I>i is a regression matrix defined as: where d is a number of time instants t in which the effective pressure values are acquired and is the regressor at time instant t defined as: and n_p | max is a predetermined maximum order value, preferably twenty -five.

4. Method (2000; 3000; 4000) according to claim 3, wherein the step of determining (3002 - 3008) an optimized maximum order value for the order n_p involves: checking (3003) the presence of a local minimum residual norm value in correspondence with order values n_p comprised in a predetermined sub-range, in the negative case, selecting (3004) as the optimized maximum order value the smallest residual norm value among the calculated values, or in the affirmative case, checking (3006) if for order values n_p outside the sub-range there is a residual norm value lower, by at least one threshold value, than the local minimum of the residual norm value, in the affirmative case, selecting (3008) the residual norm value less than the local minimum residual norm value as the optimized maximum order value, or in the negative case, selecting (3007) the local minimum residual norm value as the optimized maximum order value.

5. Method (2000; 3000; 4000) according to claim 4, wherein the predetermined subrange comprises order values n_p comprised between 1 and 10.

6. Method (2000; 3000; 4000) according to claim 4 or 5, comprising the step of imposing (3009) an optimized maximum order value of order n_ps of the polynomial hjctCit-k)' equal to the optimized maximum order value of the order n_p of the polynomial

7. Method (2000; 3000; 4000) according to any one of claims 2 to 6, comprising the step of imposing (3010) a maximum memory parameter value equal to a number of signals acquired by each sensor (4) in an acquisition time period, preferably equal to one second.

8. Method (2000; 3000; 4000) according to claim 6 or 7 , wherein the step of performing (3011) a mean of the minimization results comprises calculating the mean of the minimization results obtained by each sensor for the same order of the polynomial as: where a vector of polynomial coefficients, Nsens is the number of sensors in the insole, P is a vector of effective pressure values exerted on the sensor (4) comprised between the first value and the second value, is a regression matrix defined as: where d is a number of time instants t in which the effective pressure values are acquired and <p_tQC_L) is the regressor at time instant t defined as: and wherein n_p, n_ps, n_s = max indicates that all combinations of the order values n_ps of the polynomial h_k(C_{i t-k}) and the order values n_p of the polynomial n (Q_t)are considered, comprised between one and the optimized maximum order value, and of the memory parameter values n_s, comprised between one and the maximum memory parameter value n_s.

9. Posture/gait analysis system (1) comprising an insole (2) equipped with a plurality of pressure sensors (4) of the capacitive type and a control unit (5) connected to the pressure sensors (4), where the control unit (5) stores a set of polynomial coefficients and an order of a plurality of polynomial relationships each associated with a corresponding sensor (4), wherein the calculation of each polynomial relationship allows the control unit (5) to convert a change in capacitance of a respective sensor (4) into an estimate of the pressure experienced by the sensor (4).

10. System (1) according to claim 9, wherein the plurality of polynomial relations calculated by the memory unit (5) are of the form: where P_t (C_{i t}) is a pressure estimate based on the signals provided by the /-th sensor (4) at the time instant t, n (C_{i t}) is a first polynomial of order n_p defined as: where Ci,t is the capacitance value of the sensor i at the instant t and ao,o-np are the polynomial coefficients, h_k(C_{i t-k}) is a second polynomial of order n_ps defined as: where Q,t-k is the capacitance value of the sensor i at the instant tk, preceding the instant f, ak,i-nps are the polynomial coefficients and n_s is a memory parameter indicative of the number of polynomials /i_fc(Q_t-fc) associated with corresponding instants fe preceding the time instant t comprised in the polynomial relation Pi(C_{i t}).