WO2017013570A1

WO2017013570A1 - System and method for the acquisition and processing of data relating to a sports performance

Info

Publication number: WO2017013570A1
Application number: PCT/IB2016/054265
Authority: WO
Inventors: Tommaso GUSEO
Original assignee: Wearit S.R.L.
Priority date: 2015-07-20
Filing date: 2016-07-18
Publication date: 2017-01-26
Also published as: ITUB20152294A1

Abstract

This invention relates to a system and a method for the acquisition and processing of data relating to the sports performance of an athlete for a sport that uses skis, such as, for example, alpine skiing, Nordic skiing, freestyle, jumping and water skiing, comprising. The system comprises at least one inertial measurement unit (IMU) (10, 10', 10") applied to each boot, binding or ski and comprising at least one linear acceleration" sensor (12) and one angular acceleration sensor (14), wireless transmission means (16) installed on at least one boot, binding or ski and suitable to transmit the signals acquired from the respective inertial measurement unit, and processing means (18) suitable to receive and process the signals acquired by the inertial measurement units of both boots, bindings or skis so as to generate data relating to the position, speed, and mutual distance of the two boots, bindings or skis.

Description

DESCRIPTION

"System and method for the acquisition and processing of data relating to a sports performance" .

[0001 ] The invention relates to a system and method for the acquisition and processing of data to reconstruct the sports performance of an athlete in a sport that uses skis, such as, for example, alpine skiing, Nordic skiing, free-style, jumping and water skiing.

[0002] The system is equipped with sensors suitable to acquire the qualitative and quantitative parameters for measuring the level and effectiveness of the sports activity. This system may be embedded in the ski boot or the ski itself, or it can be externally installed as an accessory on a shoe or the sole of a ski boot or on the ski bindings. The system measures parameters such as the style and type of posture, the height and duration of a jump, the attitude of the skis both on land and in flight, the distance travelled and the curve radii.

[0003] It is well known that many skiers during sports practice attempt to reach high speeds, and perform jumps along the descent of a slope in order to improve their sports performance. In fact, the speed of the ski run must be maximised both during the time on the ground and the time in flight. In both cases, the attitude of the , skis maintained by the athlete during the entire sports practice makes an important contribution. An improvement of technique, and thus of the related position of the skis, leads to less friction during the sliding phases and thus a greater efficiency of the technical gesture.

[0004] Measuring sports performance has, until now, been left to the sensations of the athlete and improvement of performance is often confused by a multiplicity of factors over a sufficiently long section. In the best case, the average speed is assessed over a path and there are no quantitative indications for improvements and on the portions of the path in which one can note progress.

[0005] In other cases, such as the freestyle, snowboarding or kite surfing, sports performance is also related to the length and height of the jumps, the flight time and the position of the sports equipment during practice. It often involves performing conventional figures that are assessed in a qualitative manner. This activity leads the athlete to have personal impressions that make him think he has performed a great jump, without knowing how much time he spent in the air or its distance or height.

[0006] Often these athletes can collect qualitative information about sports practice by acquiring pictures with a video camera or sports action camera that provides a qualitative view (most of the time subjective) of the performance itself. [0007] There are then two typical scenarios in which qualitative measurement is necessary: to compare the progress of technique and the sports activity by the same athlete on different paths and to compare the sports practices of different athletes on different paths and at different times.

[0008] Qualitative measurement of the practice is even more important than quantitative measurement for understanding progress in training and to verify, especially in sports that require the development of a technique, the athlete's ability to improve.

[0009] In ^"some sports, such as running or cycling, there are numerous devices such as watches or smartwatches that, when worn by the^" athlete, mainly allow using GPS data to collect quantitative data primarily related to the duration and distance of the sports activity.

[0010] However, there are many other sports where performance is characterised by a multiplicity of specific parameters that prove to be difficult to measure with a general purpose device such as a smartphone or any smartwatch worn by the athlete. In many cases, the quality of the athletic gesture would be better evaluated by starting from the equipment used by the athlete.

[001 1 ] For example, skiing, whether alpine or Nordic, has specific techniques that cannot be measured with GPS- based systems alone. In this case, it becomes necessary to have an inertial system based on MEMS technology to collect information related to linear and angular accelerations.

[0012] However, the MEMS technology used today in the sports field is used most of the time to comprehensively assess the athletic gesture rather than focusing on the particularities within it.

[0013] While athletes and coaches have a need to be able to better correlate the quantitative aspect of the sports activity to the quantitative aspect provided by data obtainable with MEMS technology.

[0014] Patent US7640135 describes a system and a method for determining flight time using various accelerometers and identifying vibrations related to free fall.

[0015] Patent US7983876 describes a system capable of measuring the distance covered and the speed by means of a shoe and connected accelerometers sending data to a watch .

[0016] Patent US7451056 describes a system for monitoring sports activities connected to a person and provided with at least one accelerometer able to measure the shock, power and more generally the percentage of the sports activity within a period.

[0017] Patent US7623987 describes a shoe containing an accelerometer able to simultaneously measure the speed, distance and altitude travelled by the person wearing the shoe and reporting such data to a watch.

[0018] Patent US7457724 describes a shoe containing an accelerometer able to simultaneously measure the speed, distance and altitude travelled by the person wearing the shoe and reporting such data to a watch.

[0019] Documents AU2008202170 and AU2006222732 describe a system for saving sports practice data that includes an accelerometer, a gyroscope, a GPS and other motion and heart rate sensors and allows viewing the activity flata on a display or sending it remotely.

[0020] EP2669875 describes a system for a shoe that sends data to a remote system showing the start and end times of the sports activity and allowing the separation of intermediate times.

[0021] EP2444130 and US3973333 describe a method and a device to wear for skiing correctly.

[0022] US6767313, US3529818 and US2012/02700194 describe different devices that allow you to train yourself by reproducing the typical ski postures, whether cross^¬ country or downhill.

[0023] One of the most difficult challenges of analysing movement is the faithful reconstruction of the athletic gesture. The use of measuring systems based on MEMS and GPS sensors as also described in the previous patents appears to be efficient for the generic description of the sports activity but deficient as regards the typical characteristic description. In fact, the systems and devices described in the cited patents tend to describe the totality of movements using only a single inertial platform.

[0024] In particular, the Applicant has observed that the known devices have the following disadvantages:

[0025] - they are not able to identify the type of activity but only metrics related to the distance measured with GPS that characterise the sports performance qualitatively and only macroscopically;

[0026] - they do not provide a reliable measurement of the relative positioning between the objects, and in particular the attitude of the skis during the ski run;

[0027] - they do not provide the relative orientation between the objects;

[0028] - at the end of an activity, they do not allow you to check the detailed quantitative results related to a specific performance.

[0029] Therefore, the purpose of this invention is provide a system and a method for the acquisition and processing data in order to reconstruct a sports performance from both the quantitative and qualitative point of view, overcoming the limits and drawbacks of the known techniques mentioned above.

[0030] Such purpose is achieved with a system for the acquisition and processing of data according to -claim 1 and with a method according to claim 16.

[0031] Further characteristics and advantages of the invention will be evident from the following description of its preferred embodiments, provided by way of non- limiting examples, with reference to the accompanying drawings, in which:

[0032] - Figure 1 schematically represents the system for the acquisition and processing of data of a sports performance according to the invention;

[0033] - Figure 2 represents a ski boot in the sole of which is embedded an electronic data acquisition and processing device according to the invention;

[0034] - Figure 3 represents the electronic device connected as an external unit to a cross-country ski shoe;

[0035] - Figure 4 is a side view of a ski shoe in which the electronic device is embedded;

[0036] - Figures 5 and 5a show the sole of a ski shoe in which two or three inertial measurement units are, respectively, embedded;

[0037] - Figures 6 to 6c illustrate, in plan view, an electronic device housed in a base to be applied under the sole of a ski boot;

[0038] - Figure 7 illustrates, for a pair of skis, the different relationships between the relative distances of the tips in which the inertial measurement units of the system are located;

[0039] - Figure 8 illustrates several different curve techniques detectable with the system according to the invention;

[0040] - Figure 9 illustrates several different types and styles of curve detectable with the system according to the invention;

[0041 ] - Figure 10 is a flow diagram of the algorithm for the acquisition and processing of sports performance data;

[0042] - Figure 11 is a flowchart of a data processing algorithm according to the invention.

[0043] In the following description, elements common to the various embodiments of the acquisition and data processing system will be indicated with the same reference numbers.

[0044] In these drawings, numeral 1 indicates the acquisition and data processing system according to the invention as a whole.

[0045] In a general embodiment, the system 1 comprises at least one inertial measurement unit 10 (hereinafter, for brevity, also called IMU) applied to each boot, binding or ski and comprising at least one three-axis linear acceleration sensor 12, and a three-axis angular acceleration sensor 14, for example a gyroscope. The system also comprises wireless transmission means 16 on at least one boot, binding or ski and suitable to transmit the signals acquired from the respective inertial measurement unit, and processing means 18 suitable to receive and process the signals acquired by the inertial measurement units of both boots, bindings or skis so as to generate data relating to the position, speed, and mutual distance of the two boots, bindings or skis .

[0046] With reference to Figure 1, in a preferred embodiment, the system 1 comprises two electronic devices 2, equal to each other, each associated with a boot, binding or ski.

[0047] Each electronic device 2 includes an IMU 10 that, in addition to the linear acceleration sensor 12 and angular acceleration sensor 14, also includes a magnetic field sensor 20, useful for providing an absolute position reference .

[0048] In a preferred embodiment, each electronic device 2 is equipped with other sensors, for example a pressure sensor 22, a force sensor 24 suitable to detect the force exerted by the athlete on the ski, an altimeter 26 and a receiver of a GPS signal 28.

[0049] Each electronic device 2 is equipped with a processing unit 18 operatively connected to all the sensors of the electronic device.

[0050] Preferably, each electronic device also comprises a memory 30 capable of storing the data acquired by the sensors, a logic unit 32 capable of processing the data coming from the sensors and reconstructing the attitude of the device, a power supply battery 34, and a battery charger 36, for example of the induction type. As illustrated in Figure 4, the battery charger 36 is provided with a pair of pins 36' for electrical connection to a socket of an external power source.

[0051 ] The processing unit 18 is operatively connected to the wireless transmission means 16..

[0052] In a preferred embodiment, the system 1 also comprises a remote terminal 38, 40, for example a smartphone, a smartwatch or a tablet, suitable to exchange information with one or both of the electronic devices 2 by means of the wireless transmission means 16. The remote terminal 38, 40 can have simple functions of displaying, saving and sharing the data received from the electronic devices 2, or it can also be provided with means of processing the data received from the electronic devices, for example of a computer program suitable to implement the algorithms described below.

[0053] Such a computer program may then reside in the processing unit 18 on-board the boot, binding or ski, in the remote terminal 38, 40, or in both.

[0054] In a preferred embodiment, each electronic device 2 is provided with two or more inertial measurement units 10, 10', 10", spaced spatially from each other.

[0055] The usefulness of having several I Us is to create a system a higher number of constraints than the number of variables, in order to describe in a best numerical manner, and without calibrations and set-points by the user, the behaviour in space of the segment or segments that connect the inertial measurement units.

[0056] For example, Figures 2 and 5 illustrate a possible application of an electronic device 2 directly inside the shell 102 of a ski boot 100. This electronic device 2 is equipped with two IMUs 10, 10' , one located in a proximal area and one in the distal part of the sole of the foot. This position ensures the maximum distance between the components of the measurement units, in this way maximising the length of the segment between the two and minimising the error factors that are present in the determination of the attitude of the segment itself. [0057] In the case, for example, three IMUs 10, 10', 10", an efficient arrangement that maximises the area covered is to be preferred, as shown in Figure 5a. The maximisation of the area available is closely related to the minimisation of the residual errors.

[0058] Returning to the embodiment of Figure 2, inside the sole of the boot 100 is embedded the electronic device 2 comprising two IMUs 10,10', a processing unit 18, an inductive charging system 36, a battery 34, and a connection bus 42 between the processing unit 18 and the second IMU 10' .

[0059] In one embodiment for cross-country skiing illustrated in Figure 3, to a cross-country ski shoe 200 is rigidly connected an electronic device 2, for example housed in a protective enclosure, comprising a processing unit 18, a first IMU 10 and a second, spaced, IMU 10' .

[0060] In a further embodiment variant illustrated in Figures 6 to 6c, each electronic device 2 is housed in a base 202 (Figure 6), applicable to a. ski boot 100 (Figure 6b). Such a base 202 comprises a front portion 204, a rear portion 206 and an intermediate portion 208 of adjustment for different sizes of boot. This base 202 is fixed to the boot, for example by means of fixing screws 210 (Figure 6c). Inside the base 202 (Figure 6), the electronic device 2 comprises a processing unit 18, a first IMU 10, an inductive charging system 36, a battery 34, a second IMU 10', and a connection bus 42 between the processing unit 18 and the second IMU 10'.

[0061] Advantageously, the base 202 is removably connectable to the sole of the ski boot 100.

[0062] As said, the system according to the invention comprises processing means capable of extracting information relating to the linear and angular accelerations, speed and position of the individual inertial measurement units. The use of an over- constrained system allows, through the imposition of greater constraints, extracting a solution that is more stable and less affected by measurement errors than could, in any case, be obtained with a single inertial measurement unit.

[0063] For example, Figure 7 illustrates the different relationships between the relative distances of the nodes in which the inertial measurement units 10, 10' are present. In particular, in parallel skis (Figure 7(a)) the distances 701 and 702 are equivalent. In the case of a snowplough position of the skis (figure 7(b)), the distance 703 is less than the distance 704. In the case of offset skis (figure 7(c)), the distance 705 is greater than the distance 706.

[0064] Based on these measurements of the positions of the IMUs and ^' of the relative distances it is therefore possible to determine the relative position of the skis^'.

[0065] We will now describe the algorithms with which the processing means acquire data from the sensors and process it so as to calculate and provide qualitative and quantitative parameters that characterise the sports practice .

[0066] As said, the processing means can reside on only one or on both of the electronic devices associated with the boot, binding or ski, or even be present in the remote terminal .

[0067] With reference to the block diagram of Figure 10, in a first phase, the processing means 18 (CPU) acquire the linear and angular acceleration data and possibly of the magnetic field from the individual IMUs 10, 10' . The processing means 18 also receive other data from further sensors (GPS 28, altimeter 26, pressure 22, force 24), if present .

[0068] The processing means 18 receive data from remote IMUs, i.e., present on the other boot, binding or ski, through the connectivity module 16.

[0069] The processing means process the information in order to obtain the acceleration (121), speed (122) and position (123) of the individual IMUs.

[0070] At this point, the processing means 18 detect the athlete's activities (125) and, from the analysis of this data, correlate the data relating to the sensors of a boot, binding or ski (local data) with the data relating to the other boot, binding or ski (remote data), so as to obtain the relative position between these objects (127), and apply this information to a filter that represents a model of the sports discipline that the athlete is practising (126), in order to identify the detected event (126') such, for example, technique · in cross-country skiing or the curve or jump event in alpine skiing or the style of the action performed by the athlete (126") and then provide a qualitative description of the performance.

[0071 ] In other words, the processing means 18 have the first purpose of reconstructing, through levels of integration, the physical values of speed and position starting from acceleration measurements. These values can, in fact, be derived by simple integration or, in a preferred implementation, estimated using a Kalman filter. This solution proves to be more precise because it allows adapting the parameters of the model during the estimation of the outputs in order to reduce estimation error. Once the relative position of the objects has been obtained, the distances between the objects themselves can be estimated and then classified, for example using hidden Markov model classifiers. Classifying these distances allows associating, for example, a curve to the right, to a radius of curvature, to a relative attitude between the skis, to a positioning of the centre of mass, to estimate the precise speed and allowing a judgement of merit if the curve seen inside the entire' path was optimal.

[0072] In a preferred implementation illustrated in the flow chart of Figure 11, the algorithm comprises a first step of acquiring local data 300, i.e., coming from sensors on board the boot, binding or ski in which the processing means are present and a second step 302 of normalisation with the elimination of the phenomena of variability in the data due to the presence, for example, of the earth's magnetic field. The conditioning of the local data created in step 301 is necessary in order to avoid disturbances of the absolute values.

[0073] In one embodiment, the determination of the basic parameters (acceleration, speed and position representing the attitude) occurs through a non-linear system that is generalised through the use of EKF quaternions (step 304) and a Kalman filter (step 306) .

[0074] For the^' acquisition of the remote data (step 308), i.e., related to the IMU of the other boot, binding or ski, where processing is not performed, the processing means use a synchronisation system provided by the communication protocol (step 310) to be temporally aligned to the local data. In a similar manner to what is done with the local data, also the remote data is reconstructed using EKF quaternions (step 312) . Then, the reconstructed remote and local data are applied to an extended Kalman filter to be combined with each other (step 314) .

[0075] The joint estimate of the position then takes place by means of a classic Kalman filter that is corrected by the geometric constraints imposed by the relative position between the state variables

[0076] X ( k+ i ) = A_kX k + B _kV k + W k

[0077] while as regards the output:

[0078] Z k = H k k + v_k

[0079] Where X k is the state vector at time k and Z k is the output. The matrices Ak , B k , H k are respectively the state matrices, the input matrix and the output matrix. The signals wt and V k are white noise, Gaussians with covariance matrix Q and R, respectively. The vector x contains at each moment, all state information, but this is not directly measurable. What one can do is to estimate simultaneously the state Xk and the covariance P k calculated from the matrices Ak , Bk and H k according to the formula [0080] P_k = A_k B_k

Hk 1

[0081 ] The Kalman filter, through the two steps of prediction and updating of the measurement allows minimising the estimate of the unknown variables as R approaches zero.

[0082] Note that, in a preferred embodiment, the system appears to be over-constrained due to the introduction of an a priori geometric constraint between the data coming from the first IMU and from the -second IMU to exploit the consistency between the data within the system. The over- constrained system allows the resolution of the variables by Least Squares systems and thus minimising the reconstruction error and sparing the user from having to make particular calibrations.

[0083] As regards the data coming from the GPS sensor (step 316), whether internal or coming from the outside, these must undergo an absolute temporal alignment (step 318) with respect to the data locally synchronised by the extended Kalman filter (step 314).

[0084] In a preferred embodiment, after synchronisation, a path segmentation model divides the paths into homogeneous sections (step 320) . Then, the characteristic parameters of the homogeneous sections are calculated, such as, for example, radii of curvature and variations in height (step 322).

[0085] At the end of these steps, one thus obtains a packet of data suitable to be processed through classifiers suitable to recognise the action (step 324), the recognition of the events (step 326) and the recognition of the styles (step 328) .

[0086] These classifiers using known technologies, for example based on neural networks, support vector machines and hidden Markov models.

[0087] In practice, the classifiers associate to parameters and quantitative variables detected by the sensors, a qualitative value of style and performance. The classifiers are specific to each sport activity, containing within them the modelling of the sport activity itself.

[0088] In other words, the algorithm, in its classifications steps (324-328), is mainly catalogues the athletic action within typical basic postures of the descent, as for example the tack (Figure 8(a)), where, during the change of direction, the skis assume the particular positioning a snowplough, the base curve (Figure 8(b)), where during the change in direction, the skis are practically parallel and the centre of gravity is centred with respect to the axis of the skis, or the Christie (Figure 8(c)), where during the change in direction, the skis are practically parallel and slightly offset, and the centre of gravity is significantly decentralised with respect to the axis of the skis.

[0089] The processing means are able to recognise the relative position of the skis (for example by calculating the distances between the IMUs as shown in Figure 7) and the load difference, and thus the relative positioning of the vertical that makes the base curve different, for example, from the Christie.

[0090] Among the events are grouped, for example, changes of direction and their amplitude and shape, as shown for example in Figure 9. In fact, the change of basic direction (Figure 9(a)) leads to a very wide path and the position of the centre of gravity is particularly displaced with respect to the line of trajectory of the skis, which tend to follow the trajectory. This type of path is typical of slopes that are not too steep.

[0091 ] When, for example, the slope increases the skier tends to compress the athletic movement in the direction of advancement (Figure 9(b)) to avoid excessively increasing the speed. In this case, the curve turns out to be much more pronounced. This curve is necessary to minimise the time in which the skis are aligned to the inclination of the slope.

[0092] In a similar but complementary manner, in false flat sections, especially in the presence of slalom competitions, the ski run takes on the typical characteristics of a wagging tail (Figure 9(c)), where the skier tends to maximise the time with skis aligned with the direction of the ski slope to maximise the speed of advancement.

[0093] For example, by combining the trajectory detected by the GPS to the style and attitude measurements obtained from the data coming from the sensors (for example, speed and angle of attack) and processed with the classifiers, one is able to extract the quality of the athletic gesture .

[0094] One can then combine the data coming from the GPS with, for example, a jumping event and the attitude of the athlete. In this way, what is obtained is not only the profile in plan or elevation (given by the profile of the mountain) , but also the amplitude and efficiency of the jump and the maximum height reached.

[0095] The change of style event for example is important in the technique of cross-country skiing where associating the athlete's position from the GPS with the data coming from the IMUs will allow understanding what the style was chosen by the athlete in relation to the slope of the hill. It is then possible to compare different athletes with equal style or compare different styles and techniques of the same athlete.

[0096] Other types of events that are detected by the event classifier are jumps or curves made by bringing the skis out of fresh snow, for example, in the off-track.

[0097] To the forms of embodiment of the system and method for the acquisition and processing of data according to the invention, a technician in the field, to satisfy contingent requirements, may make modifications, adaptations and replacements of members with others functionally equivalent, without departing from the scope of the following claims. Each of the characteristics described as belonging to a possible embodiment can be achieved independently from the other embodiments described .

Claims

1. System for the acquisition and processing of data relating to the sports performance of an athlete for a sport that uses skis, such as, for example, alpine skiing, Nordic skiing, free-style, jumping and water skiing, comprising:

- at least one inertial measurement unit (IMU) (10, 10', 10") applied to each boot, binding or ski and comprising at least one linear acceleration sensor (12) and one angular acceleration sensor (14);

- wireless transmission means (16) installed on at least one boot, binding or ski and suitable to transmit the signals acquired by the respective inertial measurement units ;

- processing means (18) suitable to receive and process the signals acquired by the inertial measurement units of both boots, bindings or skis so as to generate data relating to the position, speed, and mutual distance of the two boots, bindings or skis.

2. System according to claim 1, further comprising a remote terminal (38, 40) having at least a display function of the data generated by the processing means.

3. System according to claim 1 or 2, wherein said processing means are installed on board of at least one boot, binding or ski.

4. System according to claim 2 or 3, wherein said processing means are, at least partially, housed in the remote terminal.

5. System according to any of the preceding claims, wherein each inertial measurement unit comprises at least one accelerometer (12), one gyroscope (14) and one magnetometer (20) .

6. System according to any of the preceding claims, wherein each inertial measurement unit comprises at least one force sensor (24) suitable to detect the force exerted by the athlete on the ski.

7. System according to any of the preceding claims, wherein at least one boot, binding or ski, or the remote display terminal, is equipped with a receiver to an absolute position signal (28), such as a GPS signal.

8. System according to any of the preceding claims, wherein to each boot, binding or ski are applied at least two inertial measurement units (10, 10') spatially distanced from each other and operatively connected to the processing means.

9. System according to any of the preceding claims, wherein the processing means are suitable to calculate parameters representative of. the path taken by the athlete, for example, radii of curvature and altimetric variations.

10. System according to any of the preceding claims, wherein the processing means are suitable to associate to the data obtained by the sensors qualitative parameters representative of a technique or of an event such .as a jump, curve and style obtained from a modelling of the sports discipline performed by the athlete.

11. System according to the preceding claim, wherein the processing means comprise classifiers, based for example on neural networks and/or support vector machines and/or hidden Markov models, suitable to recognise actions, events, and styles of the performance of the athlete by comparing the data obtained from the inertial measurement units with the modelling of the sports discipline.

12. System according to any of the preceding claims, wherein at least one inertial measurement unit, the processing means and the transmission means form an electronic device housed in an enclosure suitable to be coupled to the boot, binding or ski.

13. System according to any of claims 1 to 11, wherein at least one inertial measurement unit, the processing means and the transmission means form an electronic device (2) housed incorporated in the sole of a ski boot.

14. System according to any of claims 1 to 11, wherein at least one inertial measurement unit, the processing means and the transmission means form an electronic device configured as a base to be applied, for example by screwing, under the sole of a ski boot.

15. System according to any of claims 12 to 14, wherein said electronic device further comprises a battery (34) and a charger (36) for charging the battery by connection to an external power supply.

16. Method for the acguisition and processing of data relating to the sports performance of an athlete for a sport that uses skis, such as, for example, alpine skiing, Nordic skiing, free-style, jumping and water skiing, comprising:

a) acguiring data related to acceleration, speed, position of each ski, boot or binding connected to said s ki ,

b) submitting said data to a filter representative of the sports discipline performed by the athlete, in order to obtain qualitative data related to the technique, style and performance of the athlete.

17. Method according to the preceding claim, further comprising, before step b) , a step of acquiring information relating to the path followed by the athlete, such as radii of curvature and altimetric variations.

18. Method according to claim 16 or 17, wherein step b) comprises a step of submitting the data to classifiers of actions, events and styles, realised for example with neural networks and/or support vector machines and/or hidden Markov models.

19. A computer product directly loadable into the memory of a computer and comprising portions of program readable by the processor to execute the method according to any one of claims 16 to 18.