Gait segmentation method

Abstract

A gait segmentation method is described, comprising the steps of determining (1) by means of a position sensor the positions of a horizontal path (5) of a foot (6), calculating (2) horizontal displacements of the foot (6) from the positions of the path of the foot, and gait segmentation (3) from the horizontal displacements. The segmentation (3) identifies desired temporal events, phases or subphases of gait by comparing the horizontal displacements of the foot with a constant foot displacement threshold value independent of the gait mode without the need for prior training data or specific thresholds for each type of gait, allowing the analysis, study and characterization of movement in fields such as human gait research, physiotherapy, rehabilitation, sport, dance and medicine.

Description

[0001] GAIT SEGMENTATION METHOD

[0002] OBJECT OF THE INVENTION

[0003] The invention relates to a method of gait segmentation based on the advancement of the foot in the horizontal plane, which can be applied to a large number of areas or sectors in which a precise analysis of the movement of any person or living being walking is required, regardless of the speed of movement.

[0004] BACKGROUND OF THE INVENTION

[0005] Gait analysis is a technique used to study human body movement. A gait cycle is commonly divided into two phases: stance phase (when the foot is in contact with the ground, approximately 60% of the cycle) and swing phase (when the foot is in the air, occupying the remaining 40% of the cycle). The stance phase is further divided into three subphases: loading phase, mid-strike phase, and heel-lift phase. Gait analysis techniques involve identifying the instant in time where temporal events occur (i.e., events located on a time axis) that define the different phases and subphases of gait and obtaining the characteristic parameters that allow for their characterization.

[0006] There are numerous devices or systems known for gait analysis, both portable (inertial sensors, force sensors, electromyography, etc.) and non-portable (optical systems, pressure corridors, etc.). However, in recent years, the use of Inertial Measurement Units (IMUs) has increased due to their advantages: low cost, portability, precision, etc.

[0007] Different systems require methods that segment each step and identify the different gait phases. These methods provide satisfactory results for common gait patterns, but are compromised when attempting to analyze different gait patterns, such as in the case of people with gait disorders.

[0008] Gait segmentation methods based on peak detection (maximums or minimums), thresholds, and zero crossings are available in the literature. However, these present difficulties when analyzing unusual gait patterns. Other methods based on artificial intelligence and machine learning techniques, while showing satisfactory results, require prior training based on large amounts of sample data, and their performance is restricted to pre-trained gaits.

[0009] DESCRIPTION OF THE INVENTION

[0010] The present invention provides a gait segmentation method, wherein a gait is defined by a horizontal path described by a foot, comprising the steps of:

[0011] -Obtaining, using a position sensor, the positions of the horizontal trajectory of the foot over a period of time.

[0012] -Calculation of horizontal displacements of the foot from the positions of the foot's trajectory.

[0013] -Gear segmentation based on horizontal movements.

[0014] The positions of the foot trajectory are obtained by any means known to those skilled in the art, such as inertial or optical sensors.

[0015] The horizontal displacement of the foot is defined as the distance traveled by the foot in the horizontal plane, obtained from the positions of the foot's trajectory, where the X and Y positions correspond to the horizontal plane and the Z positions to the vertical plane.

[0016] From now on, it is defined that a sample / is taken at an instant t¡, so when determining the sample at which a temporal event occurs, the instant at which it occurs can also be determined.

[0017] Specifically, the calculation of each horizontal displacement associated with an instant i, Dt, is performed by calculating the difference between the positions in the horizontal plane at two predefined instants. This pair of instants can be two consecutive instants or separated by other instants.

[0018] Specifically, it is calculated between two predefined positions of the trajectory according to the following formula: where the subscripts / ej refer to the samples / ej taken at times t¡ and t¡ respectively, AD ¿ is the horizontal displacement of the foot associated with sample i, X¿ e7 are the positions on the X and Y axis t and Y¡ are the positions on the Y axis.

[0019] The segmentation stage identifies desired temporal events, phases, or subphases of gait based on horizontal foot displacements, which are an objective measure of step execution. It is performed by comparing horizontal foot displacements with a foot displacement threshold value. This value is constant and independent of gait type, and does not require prior training data or specific thresholds for each gait type.

[0020] Specifically, in the segmentation stage, the following temporal events of a gait cycle are obtained: heel contact, plantar support, heel toe-off and toe toe-off, and the following phases and subphases delimited by these events:

[0021] -Support phase, defined between heel contact and foot take-off.

[0022] -Swing phase, defined between foot take-off and the following heel contact.

[0023] -Loading subphase, defined between heel contact and plantar support.

[0024] -Middle subphase, defined between plantar support and heel take-off.

[0025] -Heel lift subphase, defined between heel take-off and toe take-off.

[0026] Thus defined, a gait cycle begins with a preceding heel strike and ends with a following heel strike.

[0027] In the segmentation stage, segmentation is performed as follows:

[0028] -The temporal events of foot take-off and heel contact are determined by a displacement threshold th and a sample window W defined as: th e [0.003 to 0.006] m

[0029] W = 15 x FS samples where FS is the sampling rate of the sensor or system in samples / second.

[0030] The chosen threshold is fixed and valid for any type of gait because it only indicates the existence of movement. The threshold must have an optimal value that guarantees maximum step detection capacity (e.g., >98% detection rate) and minimum error in the gait cycle phases (e.g., <3% error rate) for all types of gait steps.

[0031] -For a sample to be considered as a temporary standing take-off event, the following two conditions must be met:

[0032] AD¿ > th

[0033] These conditions require that in sample / the displacement increment is greater than th (i.e. there is displacement), and that in the previous sample window starting at sample i-Wy and ending at sample i-1 the horizontal displacement between each pair of predetermined samples in the sample window is less than th, i.e. the foot is considered to be supported and not moving.

[0034] -For a sample to be considered a temporary heel contact event, the following two conditions must be met:

[0035] AD¿ < th

[0036] These conditions require that at sample / the displacement increment be less than or equal to th (there is no displacement) and in the previous sample window starting at time iW and ending at sample i-1 the horizontal displacement between each pair of predetermined samples in the sample window be greater than th, i.e. the foot moves.

[0037] Using the above time events, the stance and swing phases can be determined.

[0038] Once the phases have been determined, the subphases are determined as follows:

[0039] -A more restrictive threshold is applied to stance phase displacements. Specimens belonging to the mid-subphase meet: DCT.-DP — th / 6 where CT is the heel strike sample and DP the toe-off sample. This condition implies that displacements between each pair of consecutive samples between heel strike and toe-off that are equal to or less than th / 6 are considered to belong to the mid-subphase. Furthermore, the first sample associated with a displacement that meets this condition is defined as plantar support AP and the last sample as heel toe-off DT.

[0040] Using the above time events, the loading and heel lift subphases can then be defined.

[0041] Finally, the results provide the instants of the temporal events of heel contact, plantar support, heel toe-off and toe-off, the stance and swing phases, and the loading, mid- and heel lift subphases.

[0042] The scope of application focuses on the analysis, study, and characterization of movement in various fields: human gait research, physiotherapy, rehabilitation, sports, dance, medicine focused on gait disorders such as fall risk or frailty in the elderly, as well as in companies that provide commercial solutions, for which this method represents a more robust and precise alternative.

[0043] DESCRIPTION OF THE DRAWINGS

[0044] To complement the description being made and in order to help better understand the characteristics of the invention, in accordance with a preferred example of practical implementation thereof, a set of drawings is attached as an integral part of said description, in which the following has been represented for illustrative and non-limiting purposes:

[0045] Figure 1.- Shows a diagram of the stages of the gait segmentation method.

[0046] Figure 2 shows a diagram of a horizontal trajectory of a foot on Cartesian axes. Figure 3 shows a diagram of temporal events, phases, and subphases of a preferred embodiment of the invention on a time axis.

[0047] Figure 4.- Shows a graph of the number of steps detected with respect to the displacement threshold value (th) according to the method object of the invention using two systems.

[0048] Figure 5.- Shows a graph of the percentage of support subphases detected and the value of the average subphase duration as a percentage with respect to the threshold value applied to the support phase displacements.

[0049] Figure 6 shows a graph of the percentage of steps detected relative to the displacement threshold value (th). Figure 7 shows a graph of the percentage of error in estimating the gait cycle duration and support time relative to the displacement threshold value (th).

[0050] PREFERRED EMBODIMENT OF THE INVENTION

[0051] A preferred embodiment of the gait segmentation method object of the invention is described below, where the gait is defined by a horizontal trajectory.

[0052] (5) described by a foot (6), as shown in figure 2.

[0053] As Figure 1 illustrates, the gait segmentation method comprises:

[0054] -A first stage of obtaining (1) positions of the horizontal trajectory (5) of the foot

[0055] (6).

[0056] -A second stage of calculation (2) of horizontal displacements of the foot (6) from the positions of the horizontal trajectory (5) of the foot (6) at two consecutive moments.

[0057] -A third stage of gait segmentation (3) from horizontal movements.

[0058] -A fourth stage of output of results (4) of gait segmentation.

[0059] In this embodiment, an inertial sensor is used placed on the instep of the foot (6) holding it with elastic bands, but other sensors such as optical ones, other locations and other holding means can be used, as is known to the person skilled in the art.

[0060] In the first stage (1) of the method, the inertial sensor provides angular velocity and acceleration data that are processed by integration, correction and filtering algorithms known to the person skilled in the art to obtain the positions of the horizontal trajectory (5) with a frequency that is the sampling frequency FS of the sensor or system. An example of the horizontal trajectory (5) of the foot (6) is shown in Figure 2.

[0061] From now on, it is defined that a sample / is taken at an instant t¡, so when determining the sample at which a temporal event occurs, the instant at which it occurs can also be determined.

[0062] In the second stage (2), the horizontal displacements of the foot (6) are calculated according to the difference between the positions of two consecutive instants. Specifically, they are calculated between two consecutive positions of the trajectory according to the following formula: where the subscripts / and i-1 refer to the samples / and i-1 taken at the times t¡ and t¡.i respectively, AD ¿ is the horizontal displacement of the foot (6) associated with sample i, X t e ji are the positions on the X axis, e Yj and Y i-í are the positions on the Y axis.

[0063] As shown in Figure 3, in this embodiment, in the segmentation stage (3) the following temporal events of the gait cycle (17) are obtained: heel contact (9), plantar support (10), heel take-off (11) and foot take-off (12), and the following phases and sub-phases delimited by said events:

[0064] -Support phase (7), defined between heel contact (9) and foot take-off (12).

[0065] -Rolling phase (8), defined between foot take-off (12) and the following heel contact (16).

[0066] -Loading subphase (13), defined between heel contact (9) and plantar support (10). -Middle subphase (14), defined between plantar support (10) and heel toe-off (11).

[0067] -Heel lift subphase (15), defined between heel take-off (11) and foot take-off (12).

[0068] Reference 16 indicates a heel contact following the previous heel contact (9).

[0069] Thus defined, a gait cycle (17) begins with a preceding heel contact (9) and ends with a following heel contact (16).

[0070] In the segmentation stage (3), segmentation is performed as follows:

[0071] -The temporal events of foot take-off (12) and heel contact (9) are determined by a displacement threshold th and a sample window W defined as: th e [0.003 to 0.006] m

[0072] W = 15 x FS samples where FS is the sampling rate of the sensor or system in samples / second.

[0073] These values ​​identify whether a displacement exists and whether it is continuous within a sample window corresponding approximately to 10% of the gait cycle (17). The threshold chosen is fixed and valid for any type of gait because it only indicates the existence of movement. The threshold must have an optimal value that guarantees maximum capacity to detect steps (for example, >98% detection) and minimum error in the phases of the gait cycle (17) (for example, <3% errors) for all types of gait steps. Using standard testing procedures, those skilled in the art can meet these conditions by choosing a value in the closed interval of 0.003 to 0.006 meters.

[0074] -For a sample / to be considered as a temporary standing take-off event (12), the following two conditions must be met:

[0075] AD¿ > th

[0076] These conditions require that in sample / the displacement increment is greater than th (i.e. there is displacement), and that in the previous sample window starting at sample i-Wy and ending at sample i-1 the horizontal displacement between each pair of consecutive samples in the sample window is less than th, i.e. the foot (6) is considered to be supported and not moving.

[0077] -For a sample / to be considered as a temporary heel contact event (9), the following two conditions must be met:

[0078] AD¿ < th

[0079] These conditions require that in sample / the displacement increment is less than or equal to th (there is no displacement) and in the previous sample window that starts at time iW and ends at sample i-1 the horizontal displacement between each pair of consecutive samples in the sample window is greater than th, that is, the foot moves.

[0080] Using the previous time events (9, 10, 11, 12), the support (7) and swing (8) phases can be determined.

[0081] Once the phases (7, 8) have been determined, to determine the subphases (13, 14, 15) proceed as follows:

[0082] -A more restrictive threshold is applied to stance phase displacements (7). Samples belonging to the middle subphase (14) meet: DCT-DP — th / 6 where CT is the heel contact sample (9) and DP the toe-off sample (12). This condition implies that displacements between each pair of consecutive samples between heel contact (9) and toe-off (12) that are equal to or less than th / 6 are considered to belong to the middle subphase (14). Furthermore, the first sample associated with a displacement that meets this condition is defined as plantar support (10) AP and the last sample as heel toe-off (11) DT.

[0083] Using the previous time events (9,10,11,12), the loading (13) and heel lift (15) subphases can then be defined.

[0084] In the fourth stage (4), the instants of the temporal events of heel contact (9), plantar support (10), heel take-off (11) and foot take-off (12), the support (7) and swing (8) phases and the loading (13), middle (14) and heel lift (15) subphases, and other calculated parameters are provided as a result.

[0085] Additionally, in the segmentation stage (3), some temporal parameters of the gait can be calculated as defined below:

[0086] - Gait cycle duration (17), DC, which is the time elapsed between the previous heel contact (9) and the next heel contact (16):

[0087] DC — t CT next CT previous where t CT former es the instant of contact of the previous heel (9) and t CT next is the instant of the next heel contact (16).

[0088] -Duration of the phases:

[0089] -Percentage of support time with respect to the duration of the gait cycle (17): nr. 100 where t DP It is the moment of standing take-off (12).

[0090] -Percentage of swing phase time relative to the duration of the gait cycle (17): swing time (%) = 100 — support time (%)

[0091] -Percentage of time of the subphases (13, 14, 15):

[0092] . . ^CP cT previous nr. load time (%) = - - — - - x 100 balancing time

[0093] ÍET tcp mean time (%) = - - — - - x 100 balancing time 100 where t ET It is the moment of elevation of the heel.

[0094] Additionally, with the data on positions and times, parameters such as walking speed and angle of rotation of the foot (6) at the moment of heel contact (9) and foot take-off (12) can also be estimated.

[0095] To illustrate the advantages of the object of the invention, the results of a test carried out according to the preferred embodiment of the method object of the invention are set out below.

[0096] In this test an inertial sensor was used, which was placed on the instep of the foot (6) and fastened to the shoe with elastic bands. During the experiment, tests were carried out with three different subjects and six walking modes, both usual at different speeds (medium, slow and fast), and unusual forms (with high steps, shuffling and with foot drop). The data provided by the inertial sensor (angular velocity and acceleration) were processed by integration, correction and filtering algorithms to obtain the positions of the foot (6) in the plane, and the rest of the stages of the gait segmentation method were applied. Subsequently, a set of gait parameters were estimated and evaluated using the Optitrack optical system as a true reference with 26 optical markers placed on the lower extremities of the subjects.

[0097] The Optitrack optical system provides the three-dimensional position and orientation of each of the body markers and segments as output.

[0098] The results of this study show that the proposed gait segmentation method identifies more than 98% of all steps across the different gait modes. These results are shown in Table 1, which breaks down the number and percentage of steps detected for each gait mode used.

[0099] In addition, the mean relative errors in the estimation of the set of temporal parameters of the gait have been calculated, shown in table 2. As can be seen, the relative error is 1.46% for the duration of the gait cycle (17), 2.95% for the percentage of support time and 6.08% for the swing time. The mean absolute errors have also been calculated and are found in table 3. The results show that the difference in the estimated gait cycle time (17) is only 0.006 seconds, while for the support (7) and swing (8) phases it is approximately 1.5% of the duration with respect to the duration of the total gait cycle (17).

[0100] The relative error is defined as:

[0101] Reference value — Estimated value

[0102] Relative error = - - - - - 100

[0103] Reference value

[0104] Where Reference Value is a value provided by the Optitrack system and Estimated Value is a value provided by the gait segmentation method.

[0105] The absolute error is defined as:

[0106] Absolute error = \Reference value — Estimated value

[0107] These results validate the proposed segmentation method, providing a robust solution for different walking methods without the need for prior training data.

[0108] Table 1. Number and percentage of steps detected by the proposed method, taking Optitrack as a true reference, for the different walking modes used.

[0109] Table 2. Mean Relative Error % (MEAN + TYPICAL DEVIATION) for temporal parameters by gait mode:

[0110] 1. Normal steps at medium speed; 2. Normal steps at slow speed; 3. Normal steps at fast speed; 4. High steps; 5. Shuffled steps; 6. Soft steps

[0111] Table 3. Mean absolute error (MEAN + TYPICAL DEVIATION) for temporal parameters by gait mode: 1. Normal steps at medium speed; 2. Normal steps at slow speed; 3. Normal steps at fast speed; 4. High steps; 5. Shuffled steps; 6. Shallow steps

[0112] Comparison with other trading methods

[0113] In order to analyze the improvement that the method object of the invention represents compared to those existing on the market, it has been compared with the commercial system Physilog 6 (Gaitllp) of the company MindMaze. The Physilog 6 system is a gait analysis tool based on two inertial sensors that are placed on the feet (6) and a software that provides a set of gait parameters.

[0114] The study was carried out according to the preferred embodiment of the method object of the invention, taking as a true reference the Optitrack optical system and six different walking modes.

[0115] Table 4 shows the percentage of steps detected by the method object of the invention and by the commercial system Physilog 6. As can be seen, the method object of the invention detects 6.8% more steps. Table 5 shows the average relative error obtained from the temporal parameters of the method object of the invention and the commercial system, as well as the error increase between both methods (last row). The results show an error improvement of approximately 69% for the duration of the gait cycle (17), support time and swing time, and between 12-38% for the load, middle and heel lift times. These results support the use of the method object of the invention for gait identification and segmentation.

[0116] Table 4. Number and percentage of steps detected by the proposed method and by the commercial system Physilog 6, taking Optitrack as the true reference, for the different walking modes used.

[0117] Table 5. Mean Relative Error % (MEAN + TYPICAL DEVIATION) for the temporal parameters committed by the proposed method and by the commercial system Physilog 6. Increase in error between both methods.

[0118] Study of threshold values

[0119] To study the threshold values, the method object of the invention was applied to the foot trajectory obtained from an optical system (Optitrack) and an inertial sensor (IMU). The position signal provided by each of the systems has a signal-to-noise level of 50 dB for Optitrack and 90 dB for the IMU.

[0120] Selecting the foot displacement threshold th for step detection: A study has been conducted using the foot displacement threshold (th) values ​​from the following list: th = {0.00025; 0.0005; 0.001; 0.002; 0.003; 0.004; 0.005; 0.006; 0.007; 0.008; 0.009; 0.010} m

[0121] Figure 4 shows the value of the threshold th on the X axis and the number of steps detected after applying the proposed method on the data from the Optitrack system (squares) and an IMU (circles) on the Y axis.

[0122] As can be seen, in the case of Optitrack, the number of steps reaches its maximum and remains stable for th values ​​between 0.003 and 0.006 m. For values ​​outside this range, the number of steps detected is lower. In the case of the IMU, the number of steps detected is very similar to that obtained when applying the method to Optitrack in the range of 0.003 to 0.006 m. For th values ​​equal to 0.001 m and 0.002 m, there is a greater number of steps detected. However, this is due to the fact that the segmentation is not performed correctly, detecting "false steps" whose length and duration are greater than the detected one.

[0123] Therefore, the threshold th set for the detection of steps is framed within the interval [0.003, 0.006] m.

[0124] Selecting the threshold for detecting gait subphases:

[0125] In order to detect the gait subphases (13, 14, 15), it is necessary to apply a new, more restrictive threshold. In order to determine the value of this threshold (th'), the method object of the invention has been applied to the data of the trajectory (5) of the foot (6) from Optitrack. The threshold th has been set to th = 0.006 m (since it belongs to the interval described in the previous section) and using the value of the threshold th' for the detection of the gait subphases (13, 14, 15) from the following list: th' = [0.0000625; 0.000125; 0.00025; 0.0005; 0.001; 0.002; 0.003; 0.004; [0.005] Figure 5 shows the percentage of detected support subphases (13,14,15) (squares) and the duration of the average support subphase (14) (circles) expressed as percentage of support time (7), versus the threshold value th' on the Y axis.

[0126] As can be seen, the percentage of subphases (13,14,15) detected is maximum for values ​​of th' between 0.0005 and 0.001, so the threshold must be between these values. For smaller values ​​of th' the percentage of subphases (13,14,15) detected tends to decrease and for larger values ​​of th' the duration of the average support subphase (14) is equivalent to the complete support phase (7).

[0127] Therefore, the threshold th' established for the detection of subphases (13, 14, 15) of the gait is framed within the interval [0.0005, 0.001] m, which is equivalent to th' = — .

[0128] In summary, it is concluded that:

[0129] -The step detection threshold value is: th = [0.003, 0.006]

[0130] -The threshold value for detecting subphases (13, 14, 15) is:

[0131] Comparison of the results obtained with both systems (Optitrack and IMU)

[0132] The method object of the invention has been applied to the data of the foot trajectory obtained by means of the optical system (Optitrack) and by means of the inertial sensor (IMU).

[0133] The following values ​​have been used for the threshold th

[0134] (0.00025; 0.0005; 0.001; 0.002; 0.003; 0.004; ) th = > m

[0135] ( 0.005; 0.006; 0.007; 0.008; 0.009; 0.010 J

[0136] Figures 6 and 7 show the results obtained by applying the method object of the invention to the IMU data, considering the results obtained with the Optitrack system as the true reference. The X axis represents the threshold value th, while the Y axis represents the percentage of steps detected (Figure 6) and the percentage of error made in the estimation of the duration of the gait cycle (17) and the support time (Figure 7). As can be seen, the percentage of steps detected is maximum (greater than 99%) and the error minimum (less than 5%) for th values ​​between 0.003 and 0.006.

[0137] Admissible noise level

[0138] The chosen thresholds are valid assuming signal-to-noise levels in the path greater than 4.1 dB. The signal-to-noise level is defined as: signal

[0139] Signal / noise level = 10 • log ( - — ) noise

[0140] It should be noted that the studies carried out in this document were performed with signal-to-noise levels of 50 dBB for Optitrack and 90 dBB for the IMU. Other signal-to-noise level values ​​on the foot position data (6) may significantly affect the results obtained.

[0141] Below is an experiment on how the signal / noise level affects the application of the method object of the invention.

[0142] For this study, a random white noise has been applied to the XY position signal of the foot (6), varying the amplitude of said noise, so that the results shown in Table 6 have been obtained. It includes the signal-to-noise ratio value and the percentage of steps detected. As can be seen, the proposed method detects more than 96% of the steps for signal-to-noise level values ​​greater than 4.1 dBB.

Claims

MODIFIED CLAIMS received by the International Office on 21 March 2025 (21.03.2025). 1-. A gait segmentation method comprising the stages of: - obtaining (1) by means of a position sensor the positions over a period of time of the horizontal trajectory (5) of a foot (6), -calculation (2) of horizontal displacements of the foot (6) (A£>) from the positions of the horizontal trajectory (5) of the foot (6), determining the difference between the positions of two predefined instants, - gait segmentation (3) from horizontal displacements, determining temporary events of foot take-off (12) and heel contact (9) where the calculated horizontal displacements of the foot (6) are compared with a foot displacement threshold value (6) (th) using a sample window value (W), where the threshold value is within the range [0.003, 0.006] meters and the sample window value (W) is equal to 15 times the sensor sampling frequency in samples / second, determining that: -a standing take-off temporary event (12) exists if: AZ) ¿ > th ^^iw-.í-i — th where AD¿ is the horizontal displacement of the foot (6) associated with sample i, and ADj-jy.ji are the displacements in a previous sample window starting at sample i- W and ending at sample i- 1, -a temporary heel contact event (9) exists if: AZ) ¿ < th the method characterized in that the segmentation stage (3) also provides at least one of the following temporal events of plantar support (10) and heel take-off (11) of a gait cycle (17), applying a more restrictive threshold to the displacements of the support phase (7), determining that the samples belonging to the middle subphase (14) comply: ^DCT-.DP — th / 6 where CT is the heel contact sample (9) and DP the toe-off sample (12), determining that the first sample associated with a displacement that meets this condition is defined as plantar support (10) AP and the last sample as heel-off (11) DT. 2-, The gait segmentation method of claim 2, wherein, from said temporal events (9, 10, 11, 12), the segmentation stage provides at least one of the support (7) and swing (8) phases or of the loading (13), middle (14), heel lift (15) subphases, defining -the support phase (7), between heel contact (9) and foot take-off (12). -the swing phase (8), between the foot take-off (12) and the following heel contact (16). -the loading subphase (13), between heel contact (9) and plantar support (10). -the middle subphase (14), between plantar support (10) and heel take-off (11). -the heel lift subphase (15), between heel take-off (11) and foot take-off (12). 3-, The gait segmentation method of claim 1, wherein the horizontal displacements of the foot (6) are calculated according to the difference between the positions of two consecutive instants according to the following formula: where the subscripts / and i-1 refer to the samples / and i-1 taken at times t¡ and ti-i respectively, AZ) ¿ is the horizontal displacement of the foot (6) associated with sample i, X t ex t-±are the positions on the X, Y, and Y axes i-1 are the positions on the Y axis. 4-, The gait segmentation method of claim 3, wherein the segmentation step (3) calculates temporal parameters of the gait defined as -cycle duration, DC, which is the time elapsed between a heel contact (9) previous and the following heel contact (16): DC — t C T next ~ ^CT previous where t CT former es the instant of previous heel contact (9) yt CT siguiente is the instant of the following heel contact (16), -duration of the phases: -Percentage of support time with respect to the duration of the gait cycle (17): d „ „ 100 where t DP is the instant of standing take-off (12), -Percentage of swing phase time (8) relative to the duration of the gait cycle (17): swing time (%) = 100 — support time (%) -Percentage of time of subphases (13, 14, 15): „ 0 100 where t ET is the instant of heel lift (15). 5-, The gait segmentation method of claim 1, wherein in the segmentation stage (3) with the position and time data, parameters of gait speed and rotation angle of the foot (6) are estimated at the moment of heel contact (9) and foot take-off (12). 6-, A gait segmentation system comprising a position sensor and a processing module configured to carry out the steps of the method according to any of claims 1 to 6. 7-, The gait segmentation system of claim 7 wherein the position sensor is an inertial sensor or an optical sensor. DECLARATION ACCORDING TO ARTICLE 19.1 According to Article 19.1 of the PCT Treaty, concerning the modification of claims before the International Bureau: A new set of demands is submitted, replacing those originally presented: - The modified claim 1 is the result of adding the technical features of the originally filed claim 2 to the originally filed claim 1. The original claim 2 has been deleted. - The modified claims 2, 3, 4, 5, 6, 7 are the result of renumbering the originally presented claims 3, 4, 5, 6, 7, 8, as a consequence of the elimination of the original claim 2.

Images