WO2018065957A1

WO2018065957A1 - Method for detecting the ruminal motility of farm animals

Info

Publication number: WO2018065957A1
Application number: PCT/IB2017/056188
Authority: WO
Inventors: Giorgio MARCHESINI; Paolo BALASSO; Lorenzo SERVA
Original assignee: Università Degli Studi Di Padova
Priority date: 2016-10-07
Filing date: 2017-10-06
Publication date: 2018-04-12
Also published as: IT201600100910A1

Abstract

Method for detecting the ruminal motility of farm animals, which comprises: a step for applying an accelerometer (1) at the fossa of one side of a ruminant animal; a step for detecting a time series of acceleration measurements (Xa, Ya, Za), carried out by means of the accelerometer (1); a step for discrimination in order to obtain, from the aforesaid acceleration measurements (Xa, Ya, Za), a selection group of such acceleration measurements (Xa, Ya, Za) detected in a first time interval (Tl) and indicative of a first behavior condition of the animal, such as a decubitus condition of the animal; a step for processing the acceleration measurements (Xa, Ya, Za) detected in the aforesaid first time interval (Tl), which calculates corresponding parameters of ruminal motility (PM) indicative of the detection of ruminal contractions of the animal.

Description

METHOD FOR DETECTING THE RUMINAL MOTILITY

OF FARM ANIMALS

DESCRIPTION

Field of application

The present invention refers to a method for detecting the ruminal motility of farm animals.

The present method is inserted in the zootechnics field and is advantageously intended to be employed in intensive breeding of ruminant animals, such as cows, sheep and goats.

In particular, the method, object of the present invention, allows providing data relative to the ruminal motility and to the behavior of the animals of a farm, providing information that can be employed for obtaining indications on the state of health and wellbeing of the animals.

State of the art

As is known, in the scope of intensive breeding of productive livestock, specific laws are in force which require farmers to ensure suitable conditions of health and wellbeing for the animals through appropriate stalling conditions.

Nevertheless, both for farmers and for control bodies, it is particularly difficult to evaluate the health and wellbeing conditions of all the animals, especially in large-size farms.

For example, one evaluation of the health of the animals is made through behavior analyses which provide for the observation (direct or by means of cameras) of the animals 24 hours per day over long periods (e.g. for multiple days), in order to detect irregular behaviors of the animals themselves which can be symptoms of diseases or inadequate health conditions. Such analysis methodology is complex to actuate, since it requires a lot of time and numerous personnel, in particular in intensive breeding operations where the high number of animals makes the observation of every single animal particularly difficult.

For the purpose of at least partly resolving the aforesaid drawback, apparatuses have been introduced on the market that are capable of continuously detecting data relative to the movement of each animal, allowing the obtainment of information relative to the time the animal has spent in the manger, to the time the animal has been lying down or to the time spent ruminating.

Such apparatuses of known type comprise multiple movement sensors, each of which intended to be applied to a corresponding animal, for example to a collar, to a foot or to an ear, and is provided with a radio transmitter capable of sending the detected data to a central unit constituted by a computer, for example.

Several examples of such apparatuses of known type are described in the patent documents US 2013/0138389, WO 2010/108496 and WO 2013/005038.

In particular, the patent application WO 2013/005038 describes an apparatus of known type for monitoring cattle, which comprises an accelerometer with three axes mounted on a collar intended to be applied to the neck of a corresponding animal.

The accelerometer is arranged for detecting data relative to the movement and to the position of the neck of the animal and for sending, by means of radio signals, such data to a reception station.

As a function of the data provided by the accelerometer, it is possible to determine the activity of the animal in order to for example identify if the animal is standing, lying down, walking, eating etc. The aforesaid apparatus of known type, even if it is able to provide data relative to the behavior of the animal, is unable to provide further parameters indicative of the health conditions of the animal itself, such as the level of Cortisol in the blood, the cardiac frequency, the respiratory frequency, the ruminal motility, etc. In particular, the continuous detection of some physiological parameters, such as the ruminal motility, allows obtaining information usable for identifying or preventing pathologies tied to an unbalanced diet, characterized by altered ruminal motility, malfunctions of the digestive system, which are accompanied by metabolic problems (hypocalcemia, ruminal acidosis and ketosis) which lead the animals to collapse and even to death.

At the state of the art, the ruminal motility can be manually measured by using a stethoscope or an echograph. Such instruments only allow specialized personnel (e.g. veterinarians) to carry out the measurements, in an extemporaneous manner. Apparatuses and methods are also known that are employed for detecting the activity of rumination of the animals in order to identify, for example, the moments in which the animals are eating. Such apparatuses and methods, like those described in the patent applications WO 2014/201039, WO 2016/086248 and WO 2016/036303, provide for the use of movement sensors adapted to detect the movement of the jaw in order to identify the activity of mastication of the animal.

For example, the document WO 2016/086248 describes a method which provides for detecting a series of measurements by means of an accelerometer applied to the head of the animal (e.g. to an ear), in order to obtain information relative to the deglutition movements of the animal itself during rumination. The document WO 2016/036303 describes a method which provides for detecting a time series of measurements through an accelerometer applied to the neck of the animal (e.g. to a collar) in order to identify the rumination times from the detection of the movements of the jaw of the animal itself.

Also these latter findings of known type, nevertheless, do not allow detecting further significant physiological parameters such as the ruminal motility.

Presentation of the invention

The main object of the present invention is therefore to overcome the drawbacks manifested by the solutions of known type, by providing a method for detecting the ruminal motility of farm animals, which is able to provide - in a highly reliable manner - information relative to the ruminal contractions of the animals of a farm.

Another object of the present invention is to provide a method for detecting the ruminal motility of farm animals, which is able to provide, in a continuous manner, information relative to the ruminal contractions.

Another object of the present invention is to provide a method for detecting the ruminal motility of farm animals, which is able to simultaneously provide data relative to ruminal contractions and data relative to the time the animal has passed in different behavior conditions, such as in decubitus position, in stationary position or in deambulation.

Another object of the present invention is to provide a method for detecting the ruminal motility of farm animals that can be actuated in a simple manner in intensive breeding operations.

Brief description of the drawings

The technical characteristics of the invention, according to the aforesaid objects, are clearly seen in the contents of the below-reported claims and the advantages thereof will be more evident from the following description, made with reference to figure 1 which represents an example of an animal to which an accelerometer is applied for the actuation of the present method.

Detailed description of a preferred embodiment The present method is intended to be employed in intensive breeding operations in order to detect information relative to the ruminal contractions, and advantageously information relative to the behavior of the animals, usable for estimating the degree of wellbeing and health of the animals themselves.

In particular, the information relative to the ruminal motility is particularly useful for detecting, by means of subsequent analysis steps, possible dysfunctions of the forestomachs in an early manner, such dysfunctions underlying many pathologies encounterable in ruminant animals and in particular in white-meat calves and in milk-producing cows.

Ruminal contraction, as is known in the field of reference, is an involuntary action of the smooth musculature that affects the rumen (one of the forestomachs that characterize the ruminants) of the animal during the digestive process.

In particular, ruminal motility regards involuntary contractions of the rumen musculature. Such contractions allow remixing the food at the ruminal level, improving the fermentative capacity of the ruminal microorganisms, they allow the eructation of the gas produced by such microorganisms and they allow the food stored in the rumen to continue along the digestive tract in order to be digested at the abomasus and intestine level. In particular, the frequency of the ruminal contractions is tied to the state of ruminal repletion, to the fibrosity of the food and to the presence of pathologies (e.g. dislocation of the abomasus, etc.). The present method comprises a step for applying an accelerometer 1, preferably of the three axes type, on one side of the trunk of the body of a ruminant animal, such as a cow, in particular on the left side of the animal.

In particular, by trunk of the animal it is intended the main mass of the body lacking the head, the neck and the limbs.

In particular, the side of the animal comprises the lateral region of the abdomen of the trunk of the animal positioned to the front with respect to the iliac crest. The aforesaid application step preferably provides for applying an accelerometer 1 to each animal of the farm or to each animal of a significant sample of the animals of the farm.

Advantageously, the accelerometer 1 is positioned at the fossa of the side of the animal, for example by means of a harness 2.

In particular, with reference to the example illustrated in figure 1, the harness 2 comprises a belt 3 which is wound around the trunk of the animal and carries, mounted therein, the accelerometer 1 arranged at the side fossa of the animal.

Advantageously, the accelerometer 1 is arranged inside a container fixed to the belt 3 of the harness 2 and abutted against the skin of the animal on the fossa of the side of the animal itself.

Preferably, the harness 2 comprises a weight 4 fixed to the belt 3 and arranged below the abdomen of the animal in order to stably maintain the harness 2 itself. In particular, the harness 2 also comprises a collar 5 wound around the neck of the animal and connected to the belt 3 by means of one or more connection bands.

Advantageously, the accelerometer 1 is provided with three Cartesian measurement axes X, Y, Z, of which: - a first axis X positioned substantially orthogonal to the ground when the animal stands in a quadrupedal position,

- a second axis Y substantially parallel to the ground when the animal stands in a quadrupedal position and substantially parallel to the longitudinal axis AL of the animal,

- and a third axis Z orthogonal to the first axis X and to the second axis Y and substantially orthogonal to the longitudinal axis AL of the animal.

By a longitudinal axis AL of the animal it is intended the axis that is extended from the rear of the animal itself and arranged in particular substantially parallel to the ground when the animal stands in a quadrupedal position.

In particular, the above-reported positions of the axes X, Y, Z of the accelerometer 1 refer to the main components of the axes X, Y, Z themselves; in the actual application of the accelerometer 1, the latter axes can present deviations from the conditions of parallelism and/or orthogonality described above, as a function for example of the particular physiology of the animal to which the accelerometer 1 is applied.

In accordance with a particular embodiment, the accelerometer has a range of measurement of the accelerations of about ±10 G with an accuracy of ±0.15 G. For example, the present method can be actuated by employing, as accelerometer 1, the device named MSR145W (of PCE Italia s.r.l. with registered office at Capannori (LU), Italy).

According to the present invention, the present method comprises a step for detecting a time series of acceleration measurements Xa, Ya, Za carried out by means of the accelerometer 1, in particular in a specific detection time interval TR. Advantageously, the aforesaid detection time interval TR lasts multiple days/weeks/months, in particular the detection step being continuously executed 24 hours per day.

Preferably, the accelerometer 1 detects the acceleration measurements Xa, Ya, Za with a specific detection frequency F, e.g. with at least 5 detections per second (i.e. with detection period TP = 1/F).

Advantageously, in the detection step, the accelerometer 1 detects the acceleration measurements Xa, Ya, Za on the three Cartesian axes X, Y, Z. In particular, the acceleration measurements Xa, Ya, Za comprise: first axial acceleration measurements Xa along the first axis X, second axial acceleration measurements Ya along the second axis Y, and third axial acceleration measurements Za along the third axis Z.

The present method also comprises a step for the discrimination, from the time series of acceleration measurements Xa, Ya, Za, of a selection group of acceleration measurements Xa, Ya, Za detected in a first time interval Tl (or in a set of multiple first time intervals Tl) advantageously comprised in the detection time interval TR. The acceleration measurements Xa, Ya, Za of such selection group are indicative of a first behavior condition of the animal assumed in the aforesaid first time interval Tl .

The method also comprises, after the step for the discrimination, a step for processing the aforesaid group of acceleration measurements Xa, Ya, Za detected in the first time interval Tl, which calculates corresponding parameters of ruminal motility PM indicative of the detection of ruminal contractions of the animal.

Advantageously, the discrimination step and the processing step of the present method are implemented by means of at least one electronic processing unit, for example comprising a computer.

Preferably, the electronic processing unit is arranged in a remote position with respect to the accelerometers 1, for example in a room of the farm or in a processing center.

Advantageously, the acceleration measurements Xa, Ya, Za detected by each accelerometer 1 are transmitted to the electronic processing unit in order to implement the steps of discrimination and processing.

In particular, each accelerometer 1 is provided with a transmission module adapted to send the detected acceleration measurements Xa, Ya, Za to the electronic processing unit.

Preferably the transmission module is of wireless type, in a manner such to automatically transmit, e.g. by means of radio waves, the acceleration measurements Xa, Ya, Za to the electronic processing unit.

Optionally, the transmission module is provided with a transcription interface in order to store the acceleration measurements Xa, Ya, Za in a portable electronic medium (such as a USB key), so as to bring such measurements to the electronic processing unit.

Advantageously, the aforesaid step for the discrimination identifies multiple behavior categories for the animal pertaining to corresponding different activity levels of the animal itself. In particular, such behavior categories comprise: a first category relative to a decubitus condition of the animal, in which the latter sleeps or rests; a second category relative to a stationary condition of the animal, in which the latter stands in a quadrupedal position (e.g. for eating, drinking etc.); a third category relative to a deambulation condition of the animal, in which the latter walks, for example in order to move itself into the stall or to graze.

The aforesaid step for the discrimination provides, for each of the aforesaid behavior categories, the time spent by the animal in each category during the first time interval Tl of the acceleration measurements Xa, Ya, Za.

Advantageously, the first behavior condition of the animal - in which the selection group of acceleration measurements Xa, Ya, Za and the corresponding first time interval Tl are identified - corresponds with the aforesaid first behavior category in which the animal is lying down. In substance, the step for the discrimination selects the time (first time interval Tl) in which the animal is lying down and the measurements (of the selection group) detected when the animal is in such conditions.

In accordance with a different embodiment, the first behavior condition of the animal corresponds with the aforesaid second stationary category of the animal, or it comprises the first and the second behavior category of the animal.

In particular, the acceleration measurements Xa, Ya, Za detected in the aforesaid first time interval Tl (corresponding in particular to the decubitus and/or stationary condition of the animal) have proven surprisingly suitable for providing, by means of the aforesaid processing step, particularly reliable indications relative to the detection of ruminal contractions of the animal.

Advantageously, the discrimination step of the present method comprises a first step for transforming the series of acceleration measurements Xa, Ya, Za (detected by the accelerometer 1) into a corresponding series of transformed variables VT, which are advantageously suitable for being subsequently processed by means of a linear discriminant analysis. The step for the discrimination also comprises a second step for transforming the transformed variable series VT into a corresponding series of postural parameters PP indicative of the motor activity of the animal in the detection time interval TR, and a step for selecting, from such series of postural parameters PP, first postural parameters PP1 indicative of the first behavior condition of the animal. Such first postural parameters PP1 are therefore derived, by means of the first and second transformation step, from the selection group of the acceleration measurements Xa, Ya, Za detected in the first time interval Tl in which the animal assumes the first behavior condition (e.g. decubitus condition).

Advantageously, the second transformation step and the selection step are obtained by means of a linear discriminant analysis.

In particular, the weight coefficients employed in the aforesaid linear discriminant analysis are obtained in an experimental manner and have values which in particular depend on the animal species to which the method is intended to be applied.

For example, the linear discriminant analysis is implemented in R language by means of the statistical function LDA of such R language (known to the man skilled in the art of the field).

Advantageously, in accordance with a particular embodiment, the first transformation step (for the calculation of the transformed variables VT from the acceleration measurements Xa, Ya, Za) provides for calculating a series of differential variables ΔΧ, ΔΥ, ΔΖ, each of which obtained from the difference between two consecutive measurements of the time series of the acceleration measurements Xa, Ya, Za. In particular, for each Cartesian axis X, Y, Z for detection of the acceleration measurements Xa, Ya, Za, the difference is calculated between the axial acceleration measurement Xa(t), Ya(t), Za(t) detected at the instant "t" and the subsequent axial acceleration measurement Xa(t+TP), Ya(t+TP), Za(t+TP) detected at the instant "t+TP". More in detail, for each Cartesian axis X, Y, Z for detection of the acceleration measurements Xa, Ya, Za, the following are calculated:

first differential variables ΔΧ = Xa(t+ TP) - Xa(t);

second differential variables ΔΥ = Ya(t+ TP) - Ya(t);

third differential variables ΔΖ = Za(t+ TP) - Za(t).

In addition, the first transformation step provides for calculating a series of average variables Xp, Yp, Zp, each of which obtained from the average of at least two consecutive measurements of the time series of the acceleration measurements Xa, Ya, Za. In particular, for each Cartesian axis X, Y, Z for detection of the acceleration measurements Xa, Ya, Za, the average is calculated between the axial acceleration measurement Xa(t), Ya(t), Za(t) detected at the instant "t" and the subsequent axial acceleration measurement Xa(t+TP), Ya(t+TP), Za(t+TP) detected at the instant "t+ TP". More in detail, the following are calculated:

first average variables Xp(t) = (Xa(t+ TP) + Xa(t))/2;

second average variables Yp(t) = (Ya(t+ TP) + Ya(t))/2;

third average variables Zp(t) = (Za(t+ TP) + Za(t))/2.

Therefore, for each acceleration measurement Xa, Ya, Za, six variables are obtained: ΔΧ, ΔΥ, ΔΖ, Xp, Yp, Zp.

Then, the first transformation step provides for calculating a series of clean differential variables X.med, Y.med, Z.med each of which obtained from the median of multiple differential variables ΔΧ, ΔΥ, ΔΖ in a calculation time interval TC.

More in detail, for each Cartesian axis X, Y, Z for detection of the acceleration measurements Xa, Ya, Za, the following are calculated:

first clean differential variables X.med = median(AX(t), ... , AX(t+TC));

second clean differential variables Y.med = median(AY(t), ... , AY(t+TC));

third clean differential variables Z.med = median(AZ(t), ... , AZ(t+TC)).

Preferably, the aforesaid calculation time interval TC is experimentally determined on the base of specific factors of the animals, such as the species, the category of productivity, the age, etc.

Advantageously, the aforesaid calculation time interval TC is comprised between about 10 seconds and 120 seconds.

In addition, the first transformation step provides for calculating a series of clean average variables Xp.med, Yp.med, Zp.med, each of which obtained from the average of the average variables Xp, Yp, Zp in the calculation time interval TC. More in detail, for each Cartesian axis X, Y, Z for detection of the acceleration measurements Xa, Ya, Za, the following are calculated:

first clean average variables Xp.med = average(Xp(t), ... , Xp(t+TC));

second clean average variables Yp.med = average(Yp(t), ... , Yp(t+TC));

third clean average variables Zp.med = average(Zp(t), ... , Zp(t+TC)).

Then, the first transformation step provides for calculating overall differential variables X.sum, Y.sum, Z.sum, each of which obtained from the sum of the clean differential variables X.med, Y.med, Z.med in a sum time interval TS, and for calculating in Xp.sum, Yp.sum, Zp.sum, each of which obtained from the sum of the clean average variables Xp.med, Yp.med, Zp.med in the sum time interval TS. More in detail, for each Cartesian axis X, Y, Z for detection of the acceleration measurements Xa, Ya, Za, the following are calculated:

first overall differential variables X.sum = (X.med(t) + ... + X.med(t+TS));

second overall differential variables Y.sum = (Y.med(t) + ... + Y.med(t+TS));

third overall differential variables Z.sum = (Z.med(t) + ... + Z.med(t+TS));

first overall average variables Xp.sum = (Xp.med(t) + ... + Xp.med(t+TS);

second overall average variables Yp.sum = (Yp.med(t) + ... + Yp.med(t+TS)); third overall average variables Zp.sum = (Zp.med(t) + ... + Zp.med(t+TS)).

In this manner, each of the aforesaid transformed variables VT comprises corresponding overall differential variables X.sum, Y.sum, Z.sum and overall average variables Xp.sum, Yp.sum, Zp.sum, according to the relation:

VT = (X.sum, Y.sum, Z.sum, Xp.sum, Yp.sum, Zp.sum).

Each transformed variable VT is used as an input variable in the aforesaid linear discriminant analysis in order to calculate the corresponding postural parameter PP associable with one of the aforesaid three behavior categories.

In particular, the aforesaid discrimination step produces, at the output, the behavior categories assumed by the animal during the detection interval TR, advantageously identifying:

- the first time interval Tl (or set of first time intervals Tl) in which the animal assumes the first behavior category (decubitus),

- a second time interval T2 (or set of second time intervals T2) in which the animal assume the second behavior category (stationary),

- and a third time interval T3 (or set of third time intervals T3) in which the animal assumes the third behavior category (deambulation).

Once the behavior categories of the animal in the detection interval TR have been verified, provision is made for aforesaid step for processing the selection group of the acceleration measurements Xa, Ya, Za detected in the first time interval Tl (or group of first intervals Tl) in which the animal assumes the first behavior condition (e.g. decubitus), in order to obtain the parameters of ruminal motility PM indicative of the ruminal contractions of the animal, such as the number of ruminal contractions, the average duration of the ruminal contractions, etc.

Advantageously, the processing step provides for calculating the parameters of ruminal motility PM as a function of the values of the second axial acceleration measurements Ya (detected along the second axis Y) and of the third axial acceleration functions Za (detected along the third axis Z), in accordance in particular with that described in detail hereinbelow.

Advantageously, the processing step provides for calculating a series of primary clean variables XI, Yl, Zl obtained as a result of a first smoothing function (implemented in particular with a median function) on the selection group of the acceleration measurements Xa, Ya, Za. Suitably, the aforesaid first smoothing function is calculated with a first smoothing interval TS1, which is preferably comprised in the range between about 1 second and 120 seconds.

In particular the aforesaid primary clean variables XI, Yl, Zl are calculated for each Cartesian axis X, Y, Z for detection of the acceleration measurements Xa, Ya, Za. More in detail, the primary clean variables XI, Yl, Zl comprise:

- first primary clean variables XI, obtained from corresponding first axial acceleration measurements Xa, according to the relation:

Xl(t) = median (Xa(t), ... , Xa(t+TSl));

- second primary clean variables Yl, obtained from corresponding second axial acceleration measurements Ya, according to the relation:

Yl(t) = median (Ya(t), ... , Ya(t+TSl));

- third primary clean variables Zl, obtained from corresponding third axial acceleration measurements Za, according to the relation:

Zl(t) = median (Za(t), ... , Za(t+TSl)).

Then, the processing step provides for calculating a series of secondary clean variables Y2, Z2 obtained as a result of a second smoothing function (implemented in particular with a median function) on the corresponding primary clean variables Yl, Zl. Suitably, the aforesaid second smoothing function is calculated with a second smoothing interval TS2, preferably greater than the first smoothing interval TSl and in particular comprised in the range between 10 seconds and 1200 seconds.

In particular, the aforesaid secondary clean variables Y2, Z2 are calculated relative to the second axis Y and to the third axis Z. More in detail, the secondary clean variables Y2, Z2 comprise:

- second secondary clean variables Y2, obtained from corresponding second primary clean variables Yl, according to the relation:

Y2(t) = median (Yl(t), ... , Yl(t+TS2));

- third secondary clean variables Z2, obtained from corresponding third primary clean variables Zl, according to the relation:

Z2(t) = median (Zl(t), ... , Zl(t+TS2)).

Advantageously, the values of the aforesaid first and second smoothing interval TSl, TS2 are defined as a function of the species and the productive category of the animal and in particular they are determined by means of experimental tests. Then, the processing step provides for comparing the primary clean variables Yl, Zl with the corresponding secondary clean variables Y2, Z2, producing corresponding comparison parameters, and for identifying events of ruminal contractions as a function of such comparison parameters.

More in detail, the processing step provides for comparing each second primary clean variable Yl(t) with the corresponding second secondary clean variable Y2(t), and for comparing each third primary clean variable Zl(t) with the corresponding third secondary clean variable Z2(t).

When Yl(t) < Y2(t) or Zl(t) > Z2(t) for a number of consecutive detections t of duration greater than a specific threshold value VS, a corresponding ruminal contraction is identified and the duration thereof is advantageously detected.

In substance, the processing step provides for identifying each ruminal contraction event when the second primary clean variables Yl(t) are lower than the corresponding second secondary clean variables Y2(t) or the third primary clean variables Zl(t) are higher than the corresponding third secondary clean variables Z2(t) for a time interval, in which the corresponding acceleration measurements Ya(t), Za(t) were detected, with a duration greater than the aforesaid threshold value VS.

Advantageously, the aforesaid threshold value VS is comprised in the range between about 1 second and 120 seconds and it is preferably determined, in particular by means of experimental tests, as a function of the species and/or productive category of the animal.

The processing step then provides for calculating, from the identified events of ruminal contractions, the parameters of ruminal motility PM of the animal.

Preferably, the parameters of ruminal motility PM comprise the frequency of the ruminal contractions in the first time intervals Tl (for example expressed as number of contractions per minute), in particular calculated by dividing the number of ruminal contractions by the corresponding first time interval Tl .

Advantageously, the parameters of ruminal motility PM comprise the average duration of the ruminal contractions in the corresponding first time interval Tl in which the animal consecutively assumes the first behavior condition (e.g. decubitus).

The invention thus described therefore attains the pre-established objects.

In particular, the information relative to the ruminal motility, and advantageously the information relative to the behavior categories assumed by the animal, is particularly useful for the early detection of possible dysfunctions of the forestomachs through successive analysis steps, such dysfunctions underlying many pathologies encounterable in ruminant animals.

For example, the daily data relative to the parameters of ruminal motility PM, and advantageously to the behavior categories, can be used for comparing with average physiological values, present for example in the literature, of the species and category of the animal itself, providing indicative information regarding the state of health and/or wellbeing of such animal.

Claims

1. Method for detecting the ruminal motility of farm animals, said method being characterized in that it comprises:

- a step for applying an accelerometer (1) at one side of the trunk of the body of a ruminant animal;

- a step for detecting a time series of acceleration measurements (Xa, Ya, Za), carried out by means of said accelerometer (1);

- a step for the discrimination, from said time series of acceleration measurements (Xa, Ya, Za), of at least one selection group of acceleration measurements (Xa, Ya, Za) detected in at least one first time interval (Tl) and indicative of a first behavior condition of said animal assumed in said first time interval (Tl);

- a step for processing said selection group of said acceleration measurements (Xa, Ya, Za) detected in said first time interval (Tl), such processing step calculating corresponding parameters of ruminal motility (PM) indicative of the detection of ruminal contractions of said animal.

2. Method according to claim 1, characterized in that said accelerometer (1) is positioned at the fossa of the side of said animal.

3. Method according to claim 1 or 2, characterized in that said discrimination step comprises:

- a first step for transforming said time series of acceleration measurements (Xa, Ya, Za) into a corresponding series of transformed variables (VT);

- a second step for transforming said transformed variable series (VT) into a corresponding series of postural parameters (PP) indicative of the motor activity of said animal;

- a step for the selection, from said postural parameters (PP) series, of first postural parameters (PPl) indicative of said first behavior condition, said first postural parameters (PPl) having been derived, through said first and second transformation steps, from said selection group of acceleration measurements (Xa, Ya, Za) detected in said first time interval (Tl).

4. Method according to claim 3, characterized in that said first transformation step provides for:

- calculating a series of differential variables (ΔΧ, ΔΥ, ΔΖ), each of which obtained from the difference between two consecutive measurements of said acceleration measurements (Xa, Ya, Za);

- calculating a series of average variables (Xp, Yp, Zp), each of which obtained from the average of at least two consecutive measurements of said acceleration measurements (Xa, Ya, Za);

- calculating clean differential variables (X.med, Y.med, Z.med), each of which obtained from the median of multiple said differential variables (ΔΧ, ΔΥ, ΔΖ) in a calculation time interval (TC);

- calculating clean average variables (Xp.med, Yp.med, Zp.med), each of which obtained from the average of said average variables (Xp, Υρ,Ζρ) in said calculation time interval (TC);

- calculating overall differential variables (X.sum, Y.sum, Z.sum), each of which obtained from the sum of said differential clean variables (X.med,

Y.med, Z.med) in a sum time interval (TS);

- calculating overall average variables (Xp.sum, Yp.sum, Zp.sum), each of which obtained from the sum of clean average variables (Xp.med, Yp.med, Zp.med) in said sum time interval (TS);

each said transformed variable (VT) comprising corresponding said overall differential variables (X.sum, Y.sum, Z.sum) and said overall average variables (Xp.med, Yp.med, Zp.med).

5. Method according to any one of the preceding claims, characterized in that said first behavior condition corresponds to a decubitus condition of said animal.

6. Method according to any one of the preceding claims, characterized in that, in said detection step, said accelerometer (1) detects said acceleration measurements (Xa, Ya, Za) on three Cartesian axes (X, Y, Z), and such acceleration measurements (Xa, Ya, Za) comprise:

- first axial acceleration measurements (Xa) along a first axis (X) of said Cartesian axes (X, Y, Z) placed substantially orthogonal to the ground when said animal stands in a quadrupedal position;

- second axial acceleration measurements (Ya) along a second axis (Y) of said Cartesian axes (X, Y, Z) placed substantially parallel to the ground when said animal stands in a quadrupedal position and substantially parallel to a longitudinal axis (AL) of said animal;

- third axial acceleration measurements (Za) along a third axis (Z) of said Cartesian axes (X, Y, Z) orthogonal to said first axis (X) and to said second axis (Y) and substantially orthogonal to the longitudinal axis (AL) of said animal.

7. Method according to claim 6, characterized in that said processing step provides for calculating said parameters of ruminal motility (PM) as a function of the values of said second axial acceleration measurements (Ya) and of said third axial acceleration measurements (Za) of said selection group.

8. Method according to any one of the preceding claims, characterized in that said processing step provides for: - calculating primary clean variables (XI, Yl, Zl) obtained as a result of a first smoothing function applied on said selection group of said acceleration measurements (Xa, Ya, Za);

- calculating secondary clean variables (Y2, Z2) obtained as a result of a second smoothing function applied on said primary clean variables (XI, Yl, Zl);

- comparing said primary clean variables (Yl, Zl) with corresponding said secondary clean variables (Y2, Z2), producing corresponding comparison parameters;

- identifying said ruminal contractions events as a function of said comparison parameters;

- calculating, from said ruminal contractions events, said ruminal motility parameters (PM).

9. Method according to claims 7 and 8, characterized in that said primary clean variables (XI, Yl, Zl) comprise:

- second primary clean variables (Yl) obtained, through said first smoothing function, from corresponding said second axial acceleration measurements (Ya) of said selection group;

- third primary clean variables (Zl) obtained, through said first smoothing function, from corresponding said third axial acceleration measurements (Za) of said selection group;

said secondary clean variables (Y2, Z2) comprising:

- second secondary clean variables (Y2) obtained, through said second smoothing function, from corresponding said second primary clean variables

(Yi);

- third secondary clean variables (Z2) obtained, through said second smoothing function, from corresponding said third primary clean variables (Zl);

said second processing step providing for:

- comparing each second primary clean variable (Yl) with the corresponding said second secondary clean variable (Y2);

- comparing each third primary clean variable (Zl) with the corresponding said third secondary clean variable (Z2);

- identifying each said event when said second primary clean variables (Yl) are lower than the corresponding second secondary clean variables (Y2) or said third primary clean variables (Zl) are higher than said third secondary clean variables (Z2) for a time interval, in which said corresponding acceleration measurements (Ya, Za) were detected, with a duration greater than a specific threshold value (VS).

10. Method according to claim 8 or 9, characterized in that said first smoothing function and said second smoothing function are implemented with median functions.

11. Method according to any one of the preceding claims, characterized in that said discrimination step and said processing step are implemented through at least one electronic processing unit.

12. Method according to any one of the preceding claims, characterized in that said parameters of ruminal motility (PM) comprise the frequency of said ruminal contractions in said first time interval (Tl).

13. Method according to any one of the preceding claims, characterized in that said parameters of ruminal motility (PM) comprise the average duration of said ruminal contractions in said first time interval (Tl).