INTERACTIONS

FIELD OF THE INVENTION

The invention relates to the field of horseback riding, and in particular a system allowing improved performance .

Horseback riding is one of the most practiced sports in the world, and may be considered as a complex biomechanical system involving two paired entities (the horse and the rider), with constant interaction. Thus, a rider on the horse's back must move in synchronization with the movement of the horse in order to minimize the burden thereupon, all while controlling its movements. The movements of the horse and of the rider are thus crucial indicators for the optimization of the performance of pair during training and competition.

Horseback riding comprises many different disciplines - jumping, racing, dressage, endurance competitions, and so forth. Endurance competitions, very difficult for both the horse and rider, take place over several hours and over distances up to 160 kilometers. Nevertheless, as highlighted recently by the study "Speed and Cardiac Recovery Variables Predict the Probability of Elimination in Equine Endurance Events" by Younes et al . of official results of endurance competitions (80 to 160 kilometers), the rate of elimination is 39 % of horses that started, of which 64 % is due to lameness.

This study highlights a major shortcoming, by the riders, in assessing the performance/health trade-off for the horse over the length of the course. The rider is in control- only he or she knows the distance to be run and the pitfalls of the course, and thus should adapt the speed and gait of his or her horse in consequence. The rider is also a burden for the horse, which greatly increases the horse's risk of injury. It is thus important for the rider to present the least burden possible on the horse all while controlling the horse in an optimal manner.

Strategic decisions by the rider (and/or trainer) , as well as how well he or she holds /positions himself or herself on the horse, are deciding factors upon the performance and health of the horse, as demonstrated by the article "Effect of the rider position during rising trot on the horse's biomechanics" by Martin et al. Though the fatigue caused by endurance competitions is well-known, the studies are focused on the evaluation of the horse only by veterinary check-ups before, during, and after the competition, without taking into consideration the effect of the rider.

The French patent FR 3 007 662 discloses a method allowing a signature representing the movements of the rider and of the horse in a simultaneous manner to be determined. To this end, a first sensor is preferably attached to a belt to be worn on the back of the rider' s waist, and a second sensor is preferably attached to the girth of the saddle, under the horse's belly. Each sensor supplies data about sensed movement, and Lissaj ous-type curves are created from the supplied data. The analysis of these curves may be complicated and time-consuming, which is not ideal when the rider wishes to monitor the performance of his/her horse in real-time.

It may therefore be desired to overcome the above- mentioned drawbacks, in particular by providing precise performance information in real-time.

SUMMARY OF THE INVENTION Embodiments of the invention relate to a system for measuring mount-rider interactions comprising:

at least a first sensor configured to be attached to the head of a first entity, which is the mount,

a second sensor configured to be attached to the head of a second entity, which is the rider, and

a data treatment unit,

wherein the first and second sensors are configured to :

capture data about the movement of the entity, and send the data to the data treatment unit, and

wherein the data treatment unit is configured to calculate parameters resulting from the movement of each entity .

According to one embodiment, the first and second sensors are inertial measurement units.

According to one embodiment, the data treatment unit is configured to calculate the parameters in a conjoint and synchronized manner.

According to one embodiment, the first and second sensors are configured to communicate with the data treatment unit by a wireless communication protocol.

According to one embodiment, the system further comprises a visual or audio signal communicating pertinent information to the second entity.

According to one embodiment, the mount is a horse capable of wearing a harness, and the first sensor

comprises a clip for attachment to the harness.

According to one embodiment, the rider is a human capable of wearing a helmet, and the second sensor

comprises means for attachment of the second sensor to the helmet . Embdoiments of the invention also relate to a method of analyzing mount-rider interactions by means of a system according to one of claims 1 to 7, comprising the steps of: placing the first sensor on the head of the first entity, which is the mount,

placing the second sensor on the head of the second entity, which is the rider,

receiving, by the data treatment unit, data from the first and second sensors while the mount is in motion,

calculating, by the data treatment unit, parameters resulting from the movement of each entity.

According to one embodiment, the step of calculating, by the data treatment unit, of the parameters is performed in a conjoint and synchronized manner.

According to one embodiment, the method further comprises a step of obtaining a graph of vertical

accelerations with respect to time from the sensor data.

According to one embodiment, the method further comprises a step of determining the position of the rider while the mount is in motion.

According to one embodiment, the method further comprises a step of determining the gait of the mount .

According to one embodiment, the method further comprises a step of obtaining an amplitude difference between two consecutive strides of the mount.

According to one embodiment, the method further comprises the steps of supplying:

a pairing indicator, which indicates, in real-time, to what point the pair is synchronized and in symbiosis; and

a regularity indicator, which indicates whether the pairing is regular or not over the course of the session. BRIEF DESCRIPTION OF THE DRAWINGS

Other particularities and advantages of the invention will also emerge from the following description, illustrated by the accompanying drawings, in which:

Figure 1 is a side view of the horse-rider movement sensing system according to one embodiment;

Figure 2 is a perspective view of a first sensor according to one embodiment;

Figure 3 is a perspective view of a second sensor according to one embodiment;

Figure 4 is a comparison graph of signals obtained while a horse is walking;

Figure 5 is a comparison graph of signals obtained while a horse is trotting;

Figure 6 is a comparison graph of signals obtained while a horse is accelerating, here from walking to trotting;

Figures 7 (7A, 7B, 7C} are graphs of craniocaudal accelerations measured of a rider during a rising trot, standing trot, and sitting trot respectively;

Figures 8 (8Ά, 8B, 8C, 8D} are graphs of sensed data with respect to different positions of the rider while the horse is trotting; and

Figures 9 {9Ά, 9B) are graphs of sensed data with respect to different positions of the rider while the horse is cantering.

DETAILED DESCRIPTION OF THE INVENTION

Embodiments of the invention relate to a system for analyzing mount-rider interactions comprising at least a first sensor fixed to the mount's head (hereinafter considered to be a horse but not limited thereto) , a second sensor fixed to the rider's head (hereinafter considered to be a human but not limited thereto) , and a data treatment unit. The sensors are able to communicate with the data treatment unit, which is able to synchronize the data and determine various factors relating to the synchronicity of the pair.

Figure 1 is a side-view of a system according to one embodiment of the invention. The system SYS comprises at least a first sensor 1, a second sensor 2, and a data treatment unit 3. The first sensor 1 is configured to be attached to the head of a first entity 4, and the second sensor 2 is configured to be attached to the head of a second entity 5. In the following, it will be considered that the first entity is a horse, but not limited thereto (could be a camel, donkey, bull, or any other animal that can be ridden) , and that the second entity is a human being but not limited thereto (could be for example a monkey, a dog, or any other animal that could ride the first entity) .

The sensors 1, 2 may be inertial measurement units (IMU) , each comprising an accelerometer, a gyroscope and a magnetometer. The data treatment unit 3 receives the data supplied by each sensor, and calculates, in real-time or near real-time, parameters resulting from the movement of each entity (horse and rider) , in a conjoint and synchronized manner, in order to supply pertinent information to the rider or to a third entity such as a trainer for example. The rider and/or trainer may then use the information to make decisions during the exercise, and for later analysis. By "real-time" it is considered, in the context of sport practice monitoring, that the delay between the monitored action and the result of the processing and information display is sufficient to allow the user (here the rider and/or trainer) efficient adjustments of the action as it unfolds. Here, the data processing can start as soon as the data is received by the data treatment unit, and is performed in a continuous manner for the entire session.

The horse 4 is generally fitted with a harness 6 on the head and a saddle 7 on the back, attached around the thorax, while the rider typically wears at least a helmet 8 and pants 9 (other equipment, such as boots and a crop, are not referenced here) .

The choices of placement of the sensors 1, 2 on the heads of the horse and of the rider respectively are not arbitrary choices, but rather the result of careful research and development. The reasons for these choices are explained herebelow.

The measurement of the locomotion of a living entity imposes certain constraints on the placement of the sensor. The sensor must be firmly attached to the entity so that the measured data reflect only the movement of the entity, and not movement of the sensor with respect to the entity, such as vibrations due to the sensor being loosely attached. Furthermore, the placement of the sensor on the entity should be carefully selected such that the supplied data are representative of the parameters/indicators to be calculated or described.

Taking into account these two main constraints, as far as the first sensor 1 on the horse is concerned, two placement possibilities of the sensor may be considered - the harness 6 and the saddle 7, as this equipment is generally always present for horseback riding events. These two possibilities have been studied, and it was decided to go with the first solution (on the harness 6} for the following reasons: In order for sensed data to best represent the movements of the horse, the sensor should be mediolaterally centered, which, in the case of placement on the saddle, reduces the possibilities to the pommel 10 (front of the saddle) , cantle 11 (rear of the saddle) , or the girth 12 (strap under the horse's belly, holding the saddle in place) ,

Placing the sensor in a non-integrated manner on the pommel 10 of the saddle is not preferred, mainly for technical reasons. As the rider is seated directly behind the pommel, the sensor would be subjected to various effects of the rider, such as being touched by the rider's hand or the rider' s seat moving, which prevents proper data measurement. For similar reasons, it is not preferred to place the sensor on the cantle 11. Placing the sensor on the girth 12 makes it vulnerable, for example to being hit by the hooves or barrier when the horse is jumping, being splashed with water, and so forth.

Furthermore, the girth may be considered to be the center of mass COM of the animal and thus representative of the movement of the entire animal. However, this is not necessarily the objective here, wherein it is desired to analyze, in a precise manner, the movement of the horse's legs. The horse's neck, described in the literature ("Signal decomposition method of evaluating head movement to measure induced forelimb lameness in horses trotting on a treadmill" by Keegan et al.) as the "oscillator" of the horse' s body, allows the locomotion of the body to be balanced. The neck thus reflects the movements of the horse's legs. On the contrary, under the girth may be one of the most mechanically stable points of the system (horse) , less impacted by the movement in order to propel the body in a longitudinal manner. The center of mass thus efficiently represents the global movement of the system, but not the technical details of the system's locomotion.

Finally, the configurations and dimensions of saddles may vary greatly due to the sport being practiced, the rider's preferences, the morphology of the horse, and so forth. Thus, the saddle is not an ideal placement.

In almost all equine activities, the horse is equipped with a harness 6 on the head, whether it is during training or competition, but also during grooming and while out to pasture. The harness is thus a major element and may have several forms (bridle, halter, and so forth} but always or at least almost always has strap or headpiece passing behind the horse's ears, to keep it in place. Furthermore, the horse' s head is an area isolated from most or all disturbances, and is unlikely to be disturbed by movements of the rider or by environmental elements such as tree branches .

The sensors 1, 2 may communicate with the data treatment unit 3 by any known means, preferably wireless such as Bluetooth®. The data treatment unit may be a dedicated unit or equipment comprising an application paired to the sensors. Many possibilities exist for communicating pertinent information to the rider. The rider can have a smartphone held in an armband or being wearing a smart watch for ease of consultation, an application may be integrated into a helmet visor or googles worn by the rider for visual consultation without having to look away from the course, or by an audio signal delivered to the helmet or headphones placed in the rider's ears for example.

The communication protocol should be reliable, able to handle a relatively large amount of data, and could be encrypted or protected if necessary, for example to prevent other riders gaining valuable information during a competition. The communication protocol should be able to handle a large amount of data, so as to provide feedback in the real-time to the rider.

Assuming the rider carries the unit 3 on an armband for ease of use, when the sensor 1 is placed on the horse' ^’ s girth 12, the mass of the horse's body is between the sensor 1 and the unit 3, disturbing the communication of data. In order to obtain the best possible communication between the sensor 1 and the unit 3, the optimal position is when it is placed on the horse's head, as demonstrated by various placement tests, since the path is direct without obstacle interposed (as shown by the arrow) .

Figure 2 is a perspective view of the first sensor 1. The sensor 1 comprises a bottom portion 20 and a top portion 21, the bottom portion forms an attachment means and the top portion encloses the electronics of the sensor. The attachment means is preferably a clip, to tightly hold the sensor when placed, yet also be easily placed and removed as necessary by the rider. Other attachment means could be envisaged, such as a hook-and-loop fastener system, buckles, snaps, double-sided tape and so forth. Furthermore, the sensor 1 could be integrated within the material of the harness 6, such as within a pocket formed in the material of the harness. However, the solution of a separable sensor with a clip has been retained here, to be able to be used on different horses, and to easily place and remove as needed, in a repeatable manner.

The top portion 21 comprises an arrow 22 pointing towards the front of the horse as a reminder of proper placement, for calibration and repeatability purposes. The sensor electronics can include an internal battery (chargeable for example by means of a USB port) , light emitting diodes (for example to indicate when the sensor is turned on, when the battery is low, or other indications), an internal storage (memory) , an antenna for communication with the data treatment unit, as well as the movement sensing elements indicated above.

Figure 3 is a perspective view of the second sensor 2. The sensor 2 is preferably attached by a hook-and-loop fastener system between its bottom face and the top of the helmet, but other means may be used, such as double-sided tape. The sensor 2 essentially comprises the same electronics as that of the first sensor 1, and will not be described in more detail here.

The sensors 1, 2 may have on/off switches (not shown), or may be "always on", in a sleeping mode when not in use and waking up when an application on the data treatment unit is launched.

Now will be discussed the results of tests of the system, with the sensor 1 placed on the horse's harness 6 versus the sensor 1 placed on the saddle 7, more specifically on the girth 12. The horse is being ridden, and moves at three gaits- walk, trot, and canter. It may be noted that the system may also be used for a horse while galloping, but as galloping is generally only used for short distances and periods of time, the cantering gait is discussed here in detail.

The system allows the parameters stride frequency, stride length, movement regularity, gait, and pairing horse/rider to be calculated and studied, from the vertical accelerations of the horse and of the rider as measured by the sensors. The data were acquired here at a sampling frequency of 200 Hz (that is to say, 100 Hz per sensor) .

Figure 4 is a comparison graph of signals obtained while a horse is walking. First signals SI (solid line) are obtained from a headpiece sensor (that is to say, placed on the horse's head) and second signals S2 (dashed line) are obtained from a girth sensor (that is to say, placed on the horse's girth). The y-axis relates to the vertical accelerations VA in millimeters per seconds squared, and the x-axis relates to time T in seconds.

The sensors are of the same make and placed on the same horse during a same session, here walking. It may be clearly noted that the first signals SI are amplified and have less noise with respect to the second signals 32. It may also be noted that it is easier to determine the oscillations corresponding to each step from the first signals .

Figure 5 is a similar comparison graph of signals obtained while the horse is trotting. First signals Sll (solid line) are those obtained from a headpiece sensor, and second signals S12 {dashed line) are those obtained from a girth sensor. It may be noted that the first signals Sll and second signals S12 have similar amplitudes, but the second signals 312 have more noise. The noise causes oscillations at frequencies that are greater than the oscillations caused by the strides, and thus disrupt the stride oscillations, falsifying the analysis.

Figure 6 is a similar comparison graph of signals obtained while the horse is accelerating, here from walking to trotting. First signals S21 {solid line) are those obtained from a headpiece sensor, and second signals S22 (dashed line) are those obtained from a girth sensor. It may be noted that the second signals S22 have smaller amplitudes during the first part of the graph and the oscillations take longer {eight more strides) to match the increased speed. As the horse's neck is an "amplifier" of the movement of the horse's legs, one can see here the horse's anticipation of and preparation for the change of pace. This is an important aspect and should be studied, as it is an asset in the description of the quality of the movement of both the horse alone and of the pair horse- rider .

Thus, due to these various reasons, it may be established that placing the sensor 1 on the horse's head presents numerous advantages, both from a practical point of view (universality, ease of placement, improved communication with the data treatment unit) , but also for the data collection and analysis, for a precise and pertinent study of the locomotion of the pair. Indeed, placing the sensor on the harness provides results unmatched by other systems on the market.

Now will be discussed the choice of placement of the sensor 2 on the rider. The sensor placed on the rider should, due to requirements of the sport, be lightweight, portable, easy to use, firmly attached, ergonomic, and so forth. A rider typically wears a helmet, riding pants, boots or gaiters, and a crop. It is evidently preferable to not place a sensor somewhere on the rider that will move excessively and asymmetrically, in particular the arms or legs, such as on an armband, a "smart" watch, or on the crop .

For the same reasons as with the horse, the most reasonable placements from technical and practical points of view are on the helmet and at the center of mass, at the waist. The front of the waist is probably not desired, as it might disturb the rider when he or she bends over the saddle .

The placement at the waist, such as on a belt around the waist, was the solution described in the patent FR 3 007 662. However, the style of dressing may change depending on the type of riding activity (jumping, dressage, endurance races...) . For example, pants worn for jumping or dressage may have a waistband or loops whereas pants worn for endurance competitions are without in order to be more comfortable as they are worn for longer periods.

From the various observations and tests, the solution that appeared was thus to place the sensor on the rider's helmet, as it is required equipment for most types of riding. Moreover, the shape and material of the helmets does not vary widely depending on the riding discipline, and the helmets are generally strapped tightly in place. Once placed on the helmet, the sensor is not subjected to disturbances due to interactions between the rider and horse, and can be « forgotten » in the sense that it should not bother the rider's movements.

Furthermore, the same technical reasons (center of mass versus the ''oscillator" of the neck and head) as for the horse apply here for the human.

The sensor placed on the rider allows the measurement, analysis, and study of the different locomotion parameters of the pair, such as the degree of pairing, the position used by the rider, and the technique.

Now will be discussed the results of tests of the system, with the sensor 2 placed on the rider's helmet 8 versus the sensor 2 placed at the rider's center of mass, such as attached to the back of the pants 9.

Figures 7 (7A, 7B, 7C) are graphs of craniocaudal accelerations measured of a rider (placed on top of the helmet) during a rising trot, standing trot, and sitting trot respectively.

While the horse is trotting, the rider can be in the rising position (posting) , the standing position (also known as "half-seat" or "two point") , or the sitting position. The rising position differs from the standing and sitting positions in that it alternates between a moving period (up) and a stable period. A large difference of vertical accelerations may be observed in Figure 7A between two successive peaks (two half-strides } , which is not the case for the standing or sitting positions, wherein the vertical accelerations are more regular and stable. Nevertheless, the standing position in Figure 7B may be distinguished from the sitting position in Figure 7C due to the amplitude of the acceleration peaks. While in the standing position, the rider's legs absorb a large portion of accelerations from the seat to the top of the head (craniocaudally) , and thus the accelerations are smaller.

While the rider might know that he or she is or was in a given position (rising, sitting, standing) during the session, it is important to determine the position and/or technique used, not only to give feedback to the rider but to determine what portion of data is related to what position. For example, it may show that the horse performed less well when the rider was using a certain position, or that the position is not well synchronized with the horse's movements .

Figures 8 (8A, 8B, 8C, 8D) are graphs of sensed data with respect to different positions of the rider while the horse is trotting. The y-axes of Figures 8A, 8C relate to the amplitude difference AD between two consecutive strides, in millimeters per seconds squared, whereas the y- axes of Figures 8B, 8D relate to the vertical amplitudes of the strides, in millimeters per seconds squared. The x-axes relate to the number of strides NS.

Figures 8A, 8B provide the results when the sensor is placed at the rider's center of mass, for example on a belt worn at the waist, while Figures 8C, 8D provide the results when the sensor's is placed on the rider's helmet. The plus signs ( + ) indicate a rising position (posting- right hand) , the x signs ( x ) indicate a rising position (posting - left hand) , the open boxes ( □ ) indicate the sitting position, and the open circles ( O ) relate to the standing position.

The absolute value of the acceleration amplitude differential between two successive peaks is shown by each symbol (abs (Amp2-Ampl) ) , which allows a discrimination between the rising position and the sitting/standing position in Figures 8A, 8C, as well as the values of the peak amplitudes, which helps to differentiate between the other positions (sitting or standing), in Figures 8B and 8D .

More specifically, from Figures 8A, 8C it may be determined that the rider is in the rising position (left hand or right hand) or in one of the other positions (sitting or standing) . If it has been thus determined that the rider is in one of the other positions (sitting or standing) , then the analysis of Figures 8B, 8D is performed to determine which other position is being used, sitting or standing .

Thus, from Figures 8A, 8B as compared to Figures 8C, 8D respectively, it may be noted that it is more difficult to discriminate the positions (rising, sitting, standing) when the sensor is placed at the center of mass. Indeed, for the measurements at the center of mass, the values of the amplitude differentials between two successive strides are equal to or only slightly greater for the rising position as opposed to the other positions, which makes it difficult to discriminate these values. On the contrary, when the sensor is placed on the rider's head, the values of the amplitude differentials between two successive strides is clearly greater for the rising position than for the other positions, which allows the position used by the rider to be determined.

Similarly, during a canter, the rider can assume a standing or sitting position, which may be distinguished from each other by the amplitudes of the rider' s vertical accelerations .

Figures 9 (9A, 9B) are graphs of sensed data with respect to different positions of the rider while the horse is cantering. The y-axes relate to the vertical amplitudes of the strides, in millimeters per seconds squared, while the x-axes relate to the number of strides NS.

Figure 9A provides the results when the sensor is placed at the rider's center of mass, for example on a belt worn at the waist, while Figure 9B provides the results when the sensor's is placed on the rider's helmet. The open boxes ( □ ) indicate the sitting position, and the open circles ( 0 ) relate to the standing position.

Again, from Figure 9A as compared to Figure 9B, it may be noted that it is more difficult to discriminate the positions (sitting, standing) when the sensor is placed at the center of mass. Indeed, for the measurements at the center of mass, the vertical amplitudes are equal to or only slightly greater for the sitting position as opposed to the standing position, which makes it difficult to discriminate these values. On the contrary, when the sensor is placed on the rider's head, the vertical amplitudes are clearly greater for the sitting position than for the standing position, which allows the position used by the rider to be determined.

While trotting or cantering, it may be noted that the data from a sensor at the rider' s center of mass provides less discrimination between the different types of position that may be used (seated, rising, standing) . The data are intermixed and are difficult to discriminate, whereas those provided by a sensor on the rider's helmet provide a clear separation between the different types of position.

This study clearly shows that in order to determine in a precise and reliable manner the position used by the rider, the sensor should be placed on the rider's head rather than at the center of mass. Indeed, the data acquired from the rider's head allow a more precise estimation of the rider's locomotion parameters, as indicated by but not limited to the positon of the rider.

The exact placement of sensor on the horse's and/or rider' s equipment may vary slightly depending on the type of material (for example type of harness, the form of the helmet..,) . A correction/calibration phase may be implemented in order to compensate for any differences, before analysis of the data. The positions of the sensors are identified, and the data are corrected accordingly, while correcting any effect of gravity upon the data. The data from the sensors are then projected again unto a Cartesian coordinate system, for which one of the axes corresponds to gravity.

A parmticular attention is paid to the synchronization of the data from the two sensors. When they are received by the data treatment unit, they are time-stamped. Alternatively or in complement, the data may be time- stamped by the sensors. If the data treatment unit loses or poorly receives data, for example due to the data connection being lost, a re-synchronization operation may be performed so that the analysis corresponds to the synchronized data.

Gait The horse' s gait is first determined by analysis of the data.

The data are first filtered by a low-pass filter, for example by a second-order Butterworth filter, with a cutoff frequency normalized at 0.7 for an acquisition frequency of 100 Hz, as an example. After the signal has been filtered, the oscillation peaks that have an amplitude greater than 1000 millimeters per second squared (mm/s ^A 2) and the positive zero-crossings (from negative vertical amplitude to positive vertical amplitude) are identified. Further operations may be performed to erase duplications of data, to suppress outliers, and so forth.

The position (Pos) of the stride (F) is termed PosF (n) (n being an index from 1 to N of the stride) and is determined by the position of the first oscillation peak arriving after the positive zero-crossing of the oscillation. As each datum is tagged with a time vector, the temporal instants corresponding to the positions of the strides can be identified: TimeF - Time {PosF (n) ) .

The stride frequency FreqF corresponds to the frequency of the horse's step, which is determined by studying the vertical acceleration as detected by the sensor placed on the horse. The stride frequency FreqF corresponds to the inverse of the time between two successive strides, that is: FreqF (n) =1/ {TimeF (n+1) - TimeF (n) ) .

During this step, the position and the value of the maximum amplitude of the oscillation may also be noted.

The horse's gait {walking, trotting, or cantering} is determined by comparing the stride frequency, the speed (Speed) and/or the maximum oscillation amplitude of the vertical acceleration of the horse's head. The conditions are defined as follows:

IF FreqF(n) < 2.2 AND FreqF(n) <= 1 AND

(Speed (n) < 7 OR AmpMax < 25000)

Gait (n) = Walk

ELSE IF FreqF (n) < 2.2 AND FreqF(n) <= 1 AND

(Speed (n) >= 7 OR AmpMax >= 25000)

Gait (n) = Canter

ELSE IF FreqF (n) >= 2.2 AND FreqF (n) < 4.2 AND

Speed (n) >= 2

Gait (n) = Trot

ELSE

Gait (n) = N/A

END

That is to say, if the stride frequency is comprised between 1 and 2.2 Hz (Hertz) and the speed is less than 7 km/h or the maximum amplitude is less than 25000 mm/s ^A 2, then the gait is that of walk. Otherwise, if the stride frequency is comprised between 1 and 2.2 Hz and the speed is greater than 7 km/h or the maximum amplitude is greater than 25000 mm/s ^A 2, then the gait is that of canter. Otherwise, if the stride frequency is comprised between 2.2 and 4 Hz the speed is greater than 2 km/h, the gait is that of trot. If none of these are true, then the gait is unable to be defined.

Depending on the gait, the horse uses movement patterns that are widely different. For example, trotting and cantering are very different from each other as far as the lifting/setting of feet.

Trotting

If the gait has been determined to be a trot, the analysis can now determine the rider' s position during the trot. As previously indicated, when the horse is trotting, the rider can be in three positions- sitting, standing, or rising. To determine which position the rider is in, the vertical acceleration data from the sensor on the rider's helmet is used.

After filtering, the maximas of the oscillation peaks

(MaxRider) of the vertical acceleration are determined, which should be greater than 1500 mm/s ^A 2 and with frequencies less than 0.2 Hz. The differential between two successive maximas (diff_MaxRider) is then determined. If the differential is greater than 5000 mm/s ^A 2, then the rider is in the rising position. If the differential is less than 5000 mm/s ^A 2 and the maxima of the stride is greater than 12000 mm/s ^A 2, then the rider is in the sitting position. Finally, if the differential is less than 12000 mm/s ^A 2, then the rider is in the standing position. Otherwise, if none of the above are true, then the position is unable to be determined.

These conditions may be expressed as follows:

diff_MaxRider = MaxRider (n+1 ) - MaxRider (n) ;

IF |diff MaxRider! > 5000

Trot _tech = Rising

ELSE IF idiff MaxRider! < 5000 AND Max Rider (n) >

12000

Tech_trot = Sitting

ELSE IF |diff MaxRider! < 5000 AND Max Rider (n) <

12000

Trot _tech = Standing

ELSE

Tech trot = N/A

END If it has been determined that the rider is using the rising position, then the type of rising position can also be determined. The rising position (or posting) during trotting consists of the rider rising up then resting in a balanced position for every half-stride of the horse. This technique can be used in two different manners- either the rider rises and remains balanced by following the right foreleg of the horse, called left hand trot, or by following the left foreleg of the horse, called right hand trot .

In order to determine which rising technique is being used, left hand or right hand, the data received from each sensor (horse and rider) should be analyzed.

For the algorithm, after having determined whether the rider is using the rising position, it is desired to determine whether he or she is using the right hand or left hand technique. The vertical acceleration of the rider is first analyzed over the last two half-strides , in order to determine the « rising phase » and the « balancing phase ». For this, the maximum values of two vertical acceleration oscillations of the rider corresponding to two half-strides (Pmaxl, Pmax2), the greater of these two values corresponding to the rising phase and the lesser of these two values corresponding to the balancing phase.

Next, it is determined which of the horse's feet (which diagonal pair) strike the ground at each halfstride, that is to say the left diagonal - left fore and right hind, or the right diagonal - right fore and left hind. For this, the mediolateral acceleration of the horse is determined for every two half-stride.

Three points are determined with respect to a current half-stride n:

Pn : Position of the half-stride n ; Pn-1 : Position of the half-stride n-1 ; and

Pn-2 : Position of the half-stride n-2.

The areas under the medio-lateral acceleration curves (AccM) are integrated, from the half-stride n-2 to the half-stride n-1 as area Al, and from the half-stride n-1 to the half-stride n as area A2.

If area Al is less than area A2 , the right diagonal is set before the left diagonal. If area A2 is less than area Al, then the left diagonal is set before the right diagonal .

By combing this information, it is possible to determine which hand the rider is using, as follows:

IF A2 < Al AND Pmax2 > Pmaxl

Hand = Left

ELSE IF A2 < Al AND Pmax2 < Pmaxl

Hand = Right

ELSE IF A2 > Al AND Pmax2 < Pmaxl

Hand = Left

ELSE IF A2 > Al AND Pmax2 > Pmaxl

Hand = Right

ELSE

Hand = N/A

END

Canter

As opposed to trotting, cantering is an asymmetric gait. The horse may use two techniques for cantering, either on the right or on the left, termed « right lead » or « left lead » respectively, depending on which foreleg is farther forward. Between these two configurations, the order As opposed to trotting, cantering is an asymmetric gait. The horse may use two techniques for cantering, either on the right or on the left, termed « right lead » or « left lead » respectively, depending of lifting and setting the feet changes, and thus the direction of movement as well .

For example, when the horse canters on the right lead, its movement is towards the right. It is thus easier to turn in this direction, and inversely. To determine on which lead the horse is cantering, the medio-lateral acceleration data is used. When the horse is cantering on a given lead, the distribution of its medio-lateral acceleration around zero for a stride changes. For example, when the horse is cantering on the left lead, at the moment of impact (setting the feet) , the mediolateral acceleration is first positive and then negative, whereas for the right lead, the mediolateral acceleration is first negative and then positive.

To begin, the moments for which the vertical acceleration (VA) is positive (that is, between two zero- crossings), are defined as P01 and P02. Then a third point is defined at the mid-point POm between P01 and P02. The area under the mediolateral acceleration curve (AccM) is calculated between the first zero-crossing P01 and the midpoint POm as area A3, then between the mid-point POm and the second zero-crossing P02 as area A4. The values of these two areas A3, A4 are compared and allow which lead (Lead) the horse is cantering upon to be determined.

The conditions are defined as follows:

IF A3 > A4

Lead = Left

ELSE IF A3 < A4

Lead = Right

ELSE

Lead - N/A

END

While the horse is cantering, the rider may be in a seated position or in a standing position. Likewise, to determine which position the rider is in, the vertical acceleration data from the sensor on the rider's helmet is used.

The oscillation range (Range_Rider} of the rider's vertical acceleration corresponding to the stride is calculated, which corresponds to the difference between the maximum and minimum values of the oscillation amplitude, Range__Rider = max (VA_Rider) - min (VA_Rider) .

If the oscillation range is greater than or equal to a threshold value Thresh {for example 29000 mm/s ^A -2) , then the rider is sitting, and if it is less than the threshold value, then the rider is standing. Otherwise, the position is unable to be determined, for example if the VA_Rider is poorly defined, the algorithm cannot determine the maximum and minimum values and thus the range, so the position is unable to be determined.

These conditions may be expressed as follows:

Range_Rider = max (VA_Rider) - min (VA_Rider} ;

IF Range_Rider >= Thresh Tech_canter = Seated

ELSE IF Range_Rider < Thresh

Tech_canter = Standing

ELSE

Tech canter = N/A

END

The threshold value Thresh depends on each pair, and is generally determined after a short calibration session with the horse rider. For example, in the Figure SB, the threshold value is 20000 mm/s ^A 2. The value is determined by averaging the ranges while cantering in the seated and standing positions, and selecting the mid-point value.

Other parameters

Before calculating the movement parameters of the pair horse/rider, it should first be determined whether the pair is in fact moving. To do so, the vertical accelerations of both the horse and the rider are analyzed for a given period of time, for example three seconds. The length of the given period of time should be selected so as to be sure that enough information concerning at least two strides is gathered. If at least three regular oscillations are detected, with frequencies less than 0.23 Hz and amplitudes greater than 800 m /s ^A 2 for the rider and 2000 mm/s ^A 2 for the horse, then the pair is moving. One of the entities (horse or rider) could be moving without the other, but the aim here is to study the movement of the pair .

The system allows data from the horse and from the rider to be recorded, analyzed, and stored in a synchronous manner. When the horse is not being ridden, the horse and the rider are separate and independent physiological/biomechical systems. When the horse is being ridden however, the two entities interact constantly, forming a sole complex system. The system supplies, to the rider and/or a trainer, besides data commonly used in sports monitoring {such as speed and distance) , descriptors allowing the mechanical and physiological harmony (pairing or coordination) between the rider and the horse to be objectified. This information is important for improving the performance of the pair, as well as to maintain the health and prevent injury of the horse, which is an advantage for any rider/trainer concerned about the horse's well-being .

A « pairing » indicator indicates to the rider, in real-time, to what point the pair is synchronized and in symbiosis, for example in general since the beginning of the riding session and/or for a precise period, such as since starting the present gait (walk, trot, canter) , or over the last two strides .

A « regularity » indicator indicates to the rider whether the pairing with the horse is regular or not over the course of the session.

The pairing indicator is determined by analysis of a session, which should be longer than a predetermined minimum number of seconds (for example 15 seconds) in order to have a sufficient amount of data points yet also providing regular updates to the rider.

Then, it is verified that at least 85% of the expected data have been received, and that the data from the horse and rider sensors have been correctly synchronized. If they have not been correctly synchronized, then the data are resynched and any missing data are interpolated.

The pairing indicator is then calculated if a number of strides is greater than or equal to a predetermined number, for example twelve strides, with a same gait and position (such as rising trot, or seated canter) . It is advantageous to study the pairing for stable, similar phenomena, as the pairing varies depending on the gait and position .

Once this sequence has been determined, a crosscorrelation of the vertical accelerations of the horse (VA_Horse) and of the rider (VA-Rider) is performed. The cross-correlation measures the similarity between the horse's acceleration and of the delayed copies of the acceleration by the rider as a function of the delay. The maximum allowed delay can also be fixed.

The maximum of the cross-correlation coefficient is then obtained, and the associated delay is determined. The pairing indicator is determined from this delay, and indicates to the rider the synchronization of the pair - the greater the delay, the lower the pairing indicator, and vice versa. As the aim is to obtain the highest pairing possible, the pairing indicator should be as high possible.

Once the pairing indicator has been calculated, the regularity indicator is determined as follows:

The regularity indicator is estimated from a calculation of the standard deviation of the previously- calculated time delays. The calculation is performed only if there are a certain number (for example ten) of previously-determined time lags, for a given gait and position .

As may be expected, the lower the average difference, the greater the regularity indicator, which signifies that the pair is regular and stable in its pairing.

Embodiments of the invention as described above provide a reliable means of monitoring, in a precise and synchronous manner, the movement of the horse/rider pair during the practice of various equine activities, or other activities comprising a rider and his or her mount. Nevertheless, the implementation of the invention is not limited to such use, but can be used in veterinary medicine to detect joint/bone/muscle problems for example, causing lameness. As the invention allows the detection, during the locomotion cycle, the moment of abnormal placing of the feet, it is possible to detect the affected leg or foot. Due to the ease of use and portability of the system, (two small sensors and a data treatment unit) , it may easily be taken to a distant site for analysis of an animal.

The invention thus provides the rider with pertinent information about his or her performance in order to preserve the well-being of the horse. Some of this information comes from the horse and some comes from the rider, and the study of both provides a complete picture of the horse/rider pair in movement.

Although the present invention has been described hereinabove with reference to specific embodiments, the present invention is not limited to the specific embodiments, and modifications which lie within the scope of the present invention will be apparent to a person skilled in the art. In particular, different features from different embodiments may be interchanged, where appropriate. Many further modifications and variations will suggest themselves to those versed in the art upon making reference to the foregoing illustrative embodiments, which are given by way of example only and which are not intended to limit the scope of the invention.