Method, system, and computer program product for determining the condition of a farm animal

Abstract

The invention relates to a method for determining the condition of a farm animal (2), wherein at least one probe device (1) having a multi-axis accelerometer (51) is arranged in the gastrointestinal tract (3) of the farm animal (2), and the accelerometer (51) measures (20) acceleration values at temporally successive measurement times, wherein at least one evaluation unit (12) is located outside the gastrointestinal tract (3) of the farm animal (2) and data can be transmitted between the probe device (1) and the evaluation unit (12), the method being characterised by the following steps: A) aggregating (30) a plurality of acceleration values measured at the same measurement time to form a first aggregation value in a probe control unit (6) of the probe device (1) (30); B) determining (40) at least one first characteristic value from a first evaluation set, consisting of a plurality of first aggregation values obtained from the repeated execution of step A), in the probe control unit (6); C) aggregating (50) a second evaluation set, consisting of first characteristic values obtained from the repeated execution of step B), to form a second aggregation value in the probe control unit (6); D) transmitting (60) at least one or more second aggregation values, determined by repeated application of step C), to the evaluation unit (12) by means of a communication device (8) of the probe device (1); E) determining (70) the condition of the farm animal from second aggregation values transmitted in step D) in the evaluation unit (12). The invention also relates to a system (100) and to a computer program product.

Description

[0001] Method, system and computer program product for determining a condition of a farm animal

[0002] The invention relates to a method for determining a condition of a farm animal, wherein at least one probe device with at least one multi-axis acceleration sensor is arranged within a gastrointestinal tract of the farm animal and the acceleration sensor determines acceleration values ​​at successive measuring times, wherein at least one evaluation unit is located outside the gastrointestinal tract of the farm animal and data can be transmitted between the probe device and the evaluation unit.

[0003] The invention also relates to a system for determining a condition of a farm animal, comprising at least one probe device which can be arranged in the gastrointestinal tract of the farm animal and has at least the following components arranged in a housing: at least one multi-axis acceleration sensor, at least one probe control unit, and a communication device for wirelessly transmitting and receiving data, the system further comprising at least one evaluation unit which is arranged outside the gastrointestinal tract of the farm animal.

[0004] The invention further relates to a computer program product for determining the condition of a farm animal.

[0005] Worldwide, livestock farming is trending toward large-scale operations with hundreds, sometimes thousands, of animals. In such situations—but also with smaller herd sizes—there is a need for optimized management, not least due to increasing cost pressures, for example, in terms of feeding and knowledge of the physical condition of individual animals. For this reason, livestock farming is increasingly turning to electronic management aids to utilize the advantages of information technology. This approach, known in the context of livestock farming under the umbrella term "smart farming," is referred to as "livestock management."

[0006] The applicant's AT 509255 B1 describes a bolus-shaped probe unit that can be inserted into the gastrointestinal tract of a cow and measures physical parameters such as temperature and pH. Measurement data are transmitted wirelessly to an external evaluation unit, with the measurement sensors being at least partially enclosed by a cylindrical protective device within an acid-resistant housing to protect them from mechanical damage. Such a probe device is used in the applicant's AT 521 597 B1 to carry out a method in which acceleration values ​​are measured in the gastrointestinal tract of a farm animal, converted into motility, clocked down, and sent to an evaluation unit located outside the farm animal's gastrointestinal tract for further processing. The motility can then be used to draw further conclusions about the farm animal's state of health.In this way, the amount of data to be transmitted wirelessly can be reduced and the service life of the probe device can be extended. AT 521 597 B1 thus focuses in particular on the state variables of the farm animal's organism.

[0007] In addition to such physiological parameters of the farm animal, other conditions also play a role in livestock management. For example, in cattle, disease can be suspected if the time spent standing has decreased compared to the average over a longer period. Along with mastitis (mammary inflammation) and fertility problems, lameness represents one of the most significant health problems in cattle from an economic perspective. If detected too late, it can cause major problems later on.

[0008] The time spent standing or lying down, and in particular the frequency of changing between these positions, can also be used to determine estrus, the timely detection of which is also of great economic importance for agricultural holdings.

[0009] It is therefore known in the art to use detection systems to determine the condition or body position of cattle. For example, pressure plates are used to determine the force distribution between the legs, or video analysis tools that detect unusual positioning patterns. In one variant, described in JP 2011045284 A, temperature sensors are installed in the floor of a livestock barn, and their measured values ​​are used to determine the standing and lying behavior of the livestock. This device is easy to implement and relatively inexpensive, but it is difficult to distinguish between different animals. Furthermore, it assumes that the animals regularly visit the measuring area, which is not always guaranteed.

[0010] Therefore, solutions are also known in which sensors are attached directly to the farm animal, such as external leg cuffs that can distinguish between standing and lying states using acceleration sensors with a known orientation. Relevant, for example, is JP 2016043121 A. JP 2018007613 A shows a system for managing farm animals in which a measuring device with a three-axis acceleration sensor is preferably attached to a collar near the neck of a farm animal. The measuring device sends data to a "management terminal," where it is processed together with data from external sensors for temperature, humidity, wind speed, etc. Various states of the farm animal, such as feed intake, drinking, walking, lying down, estrus, or rumination, are then determined externally.

[0011] A particular disadvantage of such solutions is the attachment of the measuring devices to the livestock. While these provide valuable insights into the condition of the livestock, they entail increased costs and effort for farms, e.g., in terms of maintenance and calibration. Ankle cuffs and collars can cause discomfort or injury to the livestock, and are therefore often chewed or pulled off, or damaged during such attempts, rendering the data unreliable and creating additional work. Furthermore, the sensors generate large amounts of data that must first be evaluated or interpreted to initiate the necessary steps. For this purpose, existing solutions suggest sending the data to evaluation devices, which consumes a lot of energy, blocks bandwidth, and thus limits user comfort.

[0012] It is therefore an object of the invention to eliminate the above-mentioned disadvantages and to provide a method or a system and a computer program product with which the condition of a farm animal can be determined in a simple and reliable manner and with the lowest possible maintenance and energy expenditure.

[0013] This object is achieved by the method mentioned above by the following steps:

[0014] A) combining several acceleration values ​​determined at the same measurement time to form a first aggregation value in a probe control unit of the probe device;

[0015] B) determining at least one first characteristic value from a first evaluation set, consisting of a plurality of first aggregation values ​​resulting from repeated execution of step A), in the probe control unit of the probe device;

[0016] C) summarizing a second evaluation set, consisting of first characteristic values ​​resulting from repeated execution of step B), to form a second aggregation value in the probe control unit of the probe device;

[0017] D) transmitting at least one or more second aggregation values ​​determined by repeated application of step C) from the probe device to the evaluation unit using a communication device of the probe device;

[0018] E) Determining a condition of the farm animal from second aggregation values ​​transmitted in step D) in the evaluation unit.

[0019] The cascading evaluation of acceleration measurement data in the probe device enables the best possible aggregation of information while eliminating unnecessary or disruptive acceleration influences. The smallest possible amount of data is transmitted to the evaluation unit, which conserves the probe device's energy, computing, and storage resources. The required bandwidth and transmission time are reduced, and the transmission medium is subjected to the lowest possible load, which is particularly important when large numbers of probe devices are actually used on farms. Further evaluation steps are performed on the evaluation unit, which can be designed accordingly. This enables rapid, efficient determination of the condition of a farm animal while ensuring a long service life of the required components.

[0020] In one variant, the condition of the farm animal determined in step E) is output in step F). This output can be done on a display unit, e.g., a computer or a user's mobile device (mobile phone, tablet, etc.), but also as an entry in a database.

[0021] Advantageously, the acceleration sensor of the probe device is a three-axis acceleration sensor. In other words, it is a sensor that measures acceleration in all three spatial directions of a Cartesian coordinate system. The three spatial directions are commonly referred to as the x-axis, y-axis, and z-axis, which are orthogonal to one another. To determine acceleration values ​​at successive measurement points in time, the acceleration sensor performs measurements at a predeterminable or predetermined measurement frequency or rate, sometimes also referred to as the sampling frequency or rate. For example, a measurement frequency of 50 Hz is specified, so that 50 measurements are performed per second.

[0022] In principle, different methods can be used to summarize the acceleration values ​​in step A). ​​For example, the raw measured values ​​or the absolute values ​​or the squares of the accelerations of the x, y and z axes can be summed. However, the probe device is freely movable in the gastrointestinal tract of the farm animal. This means that acceleration forces of various kinds act on the probe device and its acceleration sensor, for example due to a change in position of the farm animal, a change in position of the probe device in the gastrointestinal tract of the farm animal, but also due to rotations of the probe device. These accelerations are measured in the local coordinate system of the acceleration sensor, whose orientation in a global coordinate system, however, changes over time. In order to make measurements independent of the orientation orMovement of the local coordinate system, in one variant of the invention, a rotation-invariant method is used when combining the acceleration values ​​in step A). ​​In other words, the combination of the multiple acceleration values ​​determined at the same measurement time is carried out by a rotation-invariant method that is carried out in the probe control unit of the probe device.

[0023] Rotation-invariant aggregation can utilize functions of magnitude, amount, or length of acceleration, such as the root mean square (RMS) or cylindrical coordinates. In an advantageous variant of the invention, the acceleration values ​​are summarized in step A) by determining the length of the acceleration vector at a measurement time. The length or amount of the acceleration vector is referred to here as r1, resulting in the following formula: r1 = 7x2 +y 2 +z 2 , where “x”, “y” and “z” represent the measured acceleration value of the respective axis.

[0024] To save computing time and program memory in the probe control unit, another variant can be used to omit the square root or to square the resulting value r1. Alternatively or additionally, the value range of the length of the acceleration vector r1 can be reduced by multiplying it by a suitable prefactor. This keeps the number of digits of r1 small or allows the number of digits to be reduced.

[0025] In order to enable a statistical evaluation of the acceleration measurements in step B), different statistical measures can be used, for example moments of the signal distribution, kurtosis (kurtosis) or skewness, or any quantiles that subsequently allow conclusions to be drawn about the condition of the farm animal. Advantageously, the mean and / or the variance are determined from the first evaluation set as the first characteristic values. In other words, the mean and / or the variance are determined from the first aggregation values ​​forming the first evaluation set. In particular, the first evaluation set is formed from several temporally successive first aggregation values. Several first aggregation values, which are based on acceleration values ​​determined at successive measurement times, are combined to form the first evaluation set, to which step B) is applied.By applying step B) to the first evaluation set, the mean and / or variance is determined for the first evaluation set.

[0026] Advantageously, step B) is carried out several times, wherein each time step B) is carried out, the mean value and / or the variance are determined as first characteristic values ​​for a first evaluation set, wherein each time step B) is carried out, the first evaluation set differs from the first evaluation set used in the previous execution of step B) by at least one first aggregation value.

[0027] In other words, step B) is performed and the mean and / or variance is determined for a first evaluation set. Then, at least one first aggregation value of the first evaluation set is replaced by a different first aggregation value, or a new first aggregation value is added, whereby this new first aggregation value was not part of the first evaluation set on which step B) was performed. Then, step B) is performed again on this changed first evaluation set. After that, the first evaluation set is changed again, and so on, so that step B) is repeated several times. Advantageously, first aggregation values ​​that were collected at successive measurement times are used in each case.

[0028] In particular, the multiple execution of step B) in the form described here is carried out using a "sliding window" approach. The mean and / or variance as first key figures are determined iteratively ("sliding") over a section or a window ("window") of temporally successive first aggregation values, the first evaluation set. The section used is shifted in an overlapping manner, i.e. the earliest first aggregation value is repeatedly deleted from the section under consideration, the first aggregation value immediately following the latest value of the section under consideration (which was not yet part of the section under consideration) is added, and the mean and / or variance are again determined as the first key figures for this new section. In one variant, weighting of the first aggregation values ​​contained in a section under consideration would also be possible.

[0029] It is therefore advantageous if the first evaluation set is formed from values ​​that are based on acceleration values ​​determined during a specific measurement period. The measurement period is selected so that a clear recognition of the condition of the farm animal to be determined is possible. The first evaluation set is advantageously formed from initial aggregation values ​​that are based on acceleration values ​​determined during a measurement period of 1 second. In other words, the first evaluation set corresponds to the acceleration values ​​determined during a measurement period of 1 second. In particular, the measurement times at which the acceleration values ​​are determined lie within a measurement period of 1 second. By selecting the measurement period under consideration, the focus can be directed to specific conditions of the farm animal without other acceleration influences on the probe device, e.g.through motility, organ noises or other processes in the gastrointestinal tract of the farm animal.

[0030] A particularly resource-efficient implementation in the probe control unit can be achieved if the number of first aggregation values ​​in a first evaluation set corresponds to a power of two. This has a positive impact on both memory usage and the number of computational operations to be performed, leading to energy savings. With a measurement frequency of 50 Hz, i.e., if acceleration values ​​are determined at 50 measurement times per second, a favorable number of 64 first aggregation values ​​would result as the size of the first evaluation set. With a measurement frequency of 64 Hz, the number of first aggregation values ​​would correspond to a power of two and thus also cover a measurement duration of approximately one second.

[0031] In one variant of the invention, in step B), before determining the first characteristic value from the first evaluation set, the first evaluation set is downsampled to further reduce the amount of data to be processed. For example, downsampling can be implemented to 20 percent of the original data set; in the case described above, this would reduce the frequency from 50 Hz to 10 Hz. Nevertheless, data points corresponding to a measurement duration of approximately one second still need to be combined into the first evaluation set.

[0032] When transmitting data between the probe device and the evaluation unit, several challenges are faced. On the one hand, the more data that needs to be transmitted, the higher the energy consumption on the part of the probe device. On the other hand, certain restrictions that apply when transmitting on the frequencies used here must also be taken into account. For example, there are so-called "duty cycles" that define how long data may be transmitted on a particular frequency in order to ensure sufficient usability of the frequency for the general public. Therefore, in the method according to the invention, the data volume is advantageously further aggregated before transmission from the probe device to the evaluation unit.

[0033] A particularly good compromise between the amount of data to be transmitted and the information transmitted can be achieved if, in step C), the second evaluation set is formed from initial characteristic values ​​derived from acceleration values ​​determined during a measurement period of 1 to 15 minutes, preferably 10 minutes. This allows high-frequency acceleration values ​​to be transmitted with a high information content and a small data size using the characteristic value determination described for step B). At the same time, there are other state variables measured by the probe device in the gastrointestinal tract of the farm animal, which are also aggregated and transmitted for a period of 1 to 15 minutes, preferably 10 minutes, so that assigned "duty cycles" can be optimally utilized to transmit the most complete state information possible.

[0034] Advantageously, the second evaluation set is summarized in step C) using at least one of the following methods: determining a quantile, determining the median, determining the minimum, determining the maximum, or determining the mean. By applying these methods, a good aggregation of the first characteristic values ​​forming the second evaluation set can be achieved with low demands on the computing power of the probe control unit. In one variant of the invention, the second evaluation set is summarized in step C) using a "median cascade" method, which comprises the repeated execution of the following steps:

[0035] C1) dividing an input set of first characteristic values ​​into a number of value blocks, each consisting of several first characteristic values;

[0036] C2) For each block of values, determine the median of the first characteristic values ​​assigned to the block of values;

[0037] C3) Summarizing the medians determined in step C2) to form a new input set and repeating steps C1) and C2) using this new input set, wherein steps C1) to C3) are repeated until only a single block of values ​​results in step C1) and step C2) results in a single median, which then forms the second aggregation value, and wherein the second evaluation set is used as the input set for the first execution of step C1).

[0038] When using the second evaluation set, first characteristic values ​​are used, which are based on acceleration values ​​determined during a specified measurement period, preferably the 10-minute measurement period described above. However, it is advantageous not to apply an overlap of evaluation sets; instead, first characteristic values ​​from a specified measurement period are used, followed by the values ​​from the subsequent measurement period. The value blocks specified in step C1) can have a predetermined or arbitrary size or number of blocks of first characteristic values.

[0039] It is advantageous if the size of the value block is determined as the cube root of the number of first characteristic values ​​that make up the input set, with a rounding function preferably being applied to the result of the cube root of the number of first characteristic values ​​that make up the input set. In this manner, it can be ensured that the aggregation can be determined quickly and with the least possible energy consumption in the probe control unit, since the input set can be reduced to a single resulting median as the second aggregation value by performing steps C1) to C3) three times.

[0040] In the context of the present disclosure, the size of the value block is understood to mean the number of first characteristic values ​​that make up such a value block.

[0041] The rounding function can be used, in particular, as a rounding function (also floor(x) or LxJ), which assigns the nearest non-larger integer to a real number x, or the rounding function (also ceil(x) or Fxl). Applying the rounding function ensures that the number of the first characteristic values ​​forming the value block is an integer.

[0042] In the manner described above with various embodiments, it can also be achieved in particular that the

[0043] The amount of data transmitted from the communication device of the probe device to the evaluation unit, formed from at least one or more second aggregation values, is as small as possible and at the same time contains the greatest possible information content about the condition of the farm animal to be determined.

[0044] While the method according to the invention can be used to determine a wide variety of conditions of a farm animal, it is particularly advantageous if, in step E), the evaluation unit determines from the second aggregation values ​​whether the farm animal is in the standing or lying state. From this information, conclusions can be drawn about the farm animal (e.g., lameness, other activity characteristics, estrus) and the need for any interventions.

[0045] In one variant of the invention, a binary classification is used to determine the condition of the farm animal in step E), whereby the condition of the farm animal is qualified as standing if the transmitted second aggregation value is greater than a threshold value, and the condition of the farm animal is qualified as lying if the transmitted second aggregation value is less than or equal to a threshold value. In other words, the condition of the farm animal is determined using a threshold method in which the transmitted second aggregation values ​​are compared with a threshold value. Depending on whether the respective aggregation value is greater than, less than, or equal to the threshold value, the condition of the farm animal can be qualified as standing or lying during the measurement period covered by the second aggregation value.Step E) essentially consists of a threshold method to distinguish between the standing and lying position of the farm animal during a measurement period.

[0046] The threshold value used for the second evaluation procedure can be fixed and stored in the evaluation unit or can be selected by a user and entered into the evaluation unit, which is particularly advantageous if, for example, individual characteristics of the farm animal are to be taken into account when determining the condition.

[0047] Conveniently, the threshold value is determined and adjusted dynamically. In a corresponding variant of the invention, the threshold value represents the mean value of second aggregation values ​​during an evaluation period. In other words, an evaluation period is considered within which the measurement times of the acceleration values ​​underlying the second aggregation values ​​evaluated in step E) take place, and the mean value of these second aggregation values ​​is used as the threshold value.

[0048] The evaluation period can, for example, be 12 hours.

[0049] In another variant, a moving average can be used as the threshold, which is determined iteratively over a section of the second aggregation values ​​during the evaluation period. The section used is shifted in an overlapping manner, i.e., the last value from the examined section is repeatedly deleted, the first value after the section is added, and a new average is determined. This new average is then used for the thresholding procedure in step E). To calculate the average, the second aggregation values ​​occurring in the respective section can also be weighted as desired, e.g., by time of day or year, or individually for the farm animal under consideration.

[0050] Here, too, the sample size and evaluation period can be freely selected. For example, the sample size can be 12 hours with an unlimited evaluation period, but the evaluation period can also be selected as 24 hours, several days, or several weeks. Alternatives to the binary classification strategies described above include algorithms from classical machine learning (artificial neural networks - "kNN"; "support vector machine" - "SVM"; "decision tree learning"), as well as multi-layer learning or "deep learning" (convolutional neural network - "CNN"; "recurrent neural network" - "RNN"; "long short-term memory" - "LSTM"). Such algorithms enable the condition of farm animals to be determined using appropriately labeled learning data. Furthermore, methods such as "Gaussian mixture models" or "hidden Markov models" can also be used.

[0051] The above-described object is also achieved by a system mentioned above, with which a method as described above can be carried out. In particular, the probe control unit is configured to carry out steps A) to C) and step D), insofar as it takes place in the probe device, while the evaluation unit is configured to carry out step E) and step D), insofar as it takes place in the evaluation unit.

[0052] The object described above is also achieved by a computer program product mentioned at the outset, comprising a first computer program product part and a second computer program product part, wherein the first computer program product part comprises instructions which, when the program is executed by a probe control unit of a probe device, cause the latter to carry out steps A), B), C) and step D), insofar as it takes place in the probe device, of the method described above, and the second computer program product part comprises instructions which, when the program is executed by an evaluation unit, cause the latter to carry out steps E) and step D), insofar as it takes place in the evaluation unit, of the method described above.

[0053] The invention is explained in more detail below using a non-limiting embodiment shown in the drawings. The drawings are for illustrative purposes only and therefore do not limit the invention in any way. Figure 1 shows a cow as an exemplary farm animal and the arrangement of a probe device in its gastrointestinal tract;

[0054] Fig. 2 is a side view of a probe device in the closed state;

[0055] Fig. 3 is a schematic representation of a probe device and its components;

[0056] Fig. 4 is a schematic representation of a system according to the invention with several probe devices and an evaluation unit;

[0057] Fig. 5 is a flow chart of the method according to the invention; and

[0058] Fig. 6 is a schematic representation of a variant of a second aggregation method.

[0059] For reasons of clarity, identical elements are provided with the same reference numerals in the various figures.

[0060] Fig. 1 schematically shows a cow 2 as a possible example of a farm animal, in particular a ruminant farm animal, into whose gastrointestinal tract 3 a probe device 1 according to the embodiment of the invention described here is inserted. Other suitable farm animals would be, for example, sheep, goats, or even wild ruminants such as red deer.

[0061] Feed ingested and (pre-)chewed by a cow 2 enters its gastrointestinal tract 3, for example, the rumen or reticulum. From the reticulum, ingested feed is transported to the rumen and returned to the mouth of cow 2 for rumination.

[0062] By measuring state variables or physical variables in the gastrointestinal tract 3 of the cow 2, conclusions can be drawn about the physical condition of the animal, on the one hand its health, and on the other hand its body position. For this purpose, the probe device 1 is permanently arranged in the gastrointestinal tract 3 of the animal, where it moves freely, with good results being achieved particularly with an end position in the reticulum. Fig. 2 shows a schematic side view of the probe device 1, while Fig. 3 shows a partially transparent view of an embodiment of the probe device 1. The probe device 1 is designed as a bolus, i.e. it essentially has a cylindrical shape extending along a longitudinal axis 400 of the housing with rounded ends that resemble flattened spherical caps.In the illustrated embodiment, a first 51 and a second sensor element 52 are arranged within a housing 4 with a first 41 and a second closure element 42, which form the bolus.

[0063] The first sensor element 51 is a multi-axis acceleration sensor. In the following, the reference symbol "51" is used for both the first sensor element and the acceleration sensor.

[0064] In the context of the present disclosure, a multi-axis acceleration sensor is understood to mean an acceleration sensor that determines accelerations in multiple directions, in particular in multiple spatial directions. In principle, two- or multi-axis acceleration sensors can be used; gyroscopes can also be employed. In particular, however, the first sensor element 51 in the illustrated embodiment is designed as a three-axis acceleration sensor that determines acceleration in all three spatial directions—x-axis, y-axis, z-axis—of a Cartesian coordinate system, with the three axes being orthogonal to one another. The x-axis of the Cartesian coordinate system is oriented parallel to the longitudinal axis 400 of the housing 4 of the probe device 1. The acceleration sensor is advantageously designed as a MEMS type, where MEMS stands for "Micro-Electro-Mechanical System" in the known manner.

[0065] The acceleration sensor 51 is configured to determine acceleration values ​​in the x-, y-, and z-axes at successive measurement times. In other words, the acceleration sensor 51 determines acceleration values ​​at a predetermined or predeterminable measurement frequency, e.g., 50 Hz—this means that 50 acceleration values ​​are recorded for each of the axes every second. The second sensor element 52 can be designed as a temperature sensor; in addition to or instead of this, other sensors can also be used, e.g., those for measuring pH, density, pressure, conductivity, sound, optical properties, or oxygen, CO2, ammonia, glucose, volatile fatty acids, acetate, propionate, butyrate, and lactate. Furthermore, a clock generator 53, e.g., an RTC (“Real Time Clock”), is provided in the probe device 1.

[0066] The sensor elements 51, 52 and the clock generator 53 are connected to a probe control unit 6, which serves to control the probe device 1. The probe control unit 6 is embodied, for example, as an appropriately programmed microprocessor. The probe control unit 6 monitors and processes the data collected by the sensor elements 51, 52. A memory element 7, for example a memory chip or an SD card, can be provided for storing the data. The data from the acceleration sensor 51 can either be transmitted immediately to the probe control unit 6 at each measurement time or - if the acceleration sensor 51 is provided in one variant with its own buffer memory, in particular according to the FIFO principle ("First In, First Out") - with a corresponding delay and / or in packets.The memory element 7 stores both measured values ​​of the sensor elements 51, 52 and operating parameters of the probe device 1 such as radio frequency (for communication with an evaluation unit 12 shown in Fig. 4 or an intermediate transmitting / receiving unit 13), transmission channel, system time, but also configuration parameters such as the measuring frequency of the acceleration sensor and others.

[0067] Data is transmitted and received via a communication device 8, which has an antenna 9, for example, to or from an evaluation unit 12 and / or to or from a transceiver unit 13, which are arranged in the environment of the farm animal 2. The data can be measurement data from the probe device 1 or operating parameters. Conveniently, the communication device 8 is designed as both a transmitting and receiving device, so that it can transmit and receive data.

[0068] The probe device 1 is powered, for example, by a power supply device 10, which can be designed as a battery, accumulator, or capacitor (advantageously a thin-film or supercapacitor). Recharging by energy harvesting or other methods can also be provided.

[0069] In the illustrated embodiment, the components described according to the initially mentioned AT 509 255 B1 are enclosed within the housing 4 by a hollow protective device 11 which surrounds at least the energy supply device 10, which protects against mechanical impact and is made of any resistant material, for example plastic or metal.

[0070] Fig. 4 shows a system 100 according to the invention with several described probe devices 1—for reasons of clarity, the farm animals in whose gastrointestinal tract the probe devices 1 are arranged are not shown—and an evaluation unit 12, with which the probe devices 1 communicate wirelessly via a transmitting / receiving unit 13 in the illustrated embodiment. As the dashed antenna on the evaluation unit 12 shows, direct communication can also take place between the probe devices 1 and the evaluation unit 12 without the provision of a transmitting / receiving unit 13.

[0071] To increase the range or to reduce the necessary transmission power, several transmitting / receiving units 13 can be provided in a variant not shown, which function as relays.

[0072] In the illustrated embodiment, the transmitting / receiving unit 13 is connected in a known manner to an evaluation unit 12, which can be located close to or far away from the livestock. The evaluation unit 12 can, for example, be a mobile or stationary computer on which the corresponding evaluation routines run, but can also be implemented as a server or cloud server, with which a connection exists via the Internet or can be established as needed.

[0073] The evaluation unit 12 can comprise a display unit of a known type or be connected to such a display unit. Such connected display units are shown in Fig. 4 as a mobile phone 14a, tablet 14b, or laptop 14c of a user of the method according to the invention. Also shown is a database unit 14d, which can be part of the evaluation unit 12 or connected to it and to which data can be output in the form of a database entry. These display units 14a, 14b, 14c and / or the database unit 14d can be components of the system 100, as shown in Fig. 4. The system 100 can therefore comprise at least one display unit and / or one database unit.

[0074] According to the invention, the described system 100 comprising probe device 1 and evaluation unit 12 is used to perform a method for determining the condition of a farm animal 2 from acceleration values. This is not a physiological condition or state of health, i.e., information about the organism of the farm animal 2, but essentially a condition that corresponds to its position or behavior in space, e.g., a posture. This can be standing, lying down, walking, running, feeding, drinking, or the like.

[0075] In particular, it must be taken into account that the probe device 1 moves freely within the gastrointestinal tract 3 of the farm animal 2 and therefore accelerations from various sources act on the probe device 1 or its acceleration sensor 51. The measured acceleration values ​​are, among other things, a superposition of the following sources:

[0076] - Movement, i.e. change of the center of mass of the farm animal 2;

[0077] - Motility, i.e. rumen or stomach activity, essentially intestinal movements of the gastrointestinal tract 3 of the farm animal 2 - in other words, motility is contractions of the gastrointestinal tract 3 which act as acceleration forces on the probe device 1;

[0078] - Interference signals from organs, e.g. from the heart, diaphragm or lungs;

[0079] - changes in the orientation of the probe device, which change the measured direction (not the magnitude) of the acceleration due to gravity in the coordinate system of the acceleration sensor;

[0080] - Positional changes in the gastrointestinal tract 3, which lead to long-term changes in signal statistics.

[0081] In order to determine the condition of the farm animal based on acceleration values, the influences described above must be adequately taken into account. The method according to the invention is explained below using an exemplary embodiment in which it is determined whether the farm animal is in a standing or lying state.

[0082] The individual steps are described using the flow chart in Fig. 5 as an example.

[0083] Shown are a probe device 1, a transmitting / receiving unit 13 and an evaluation unit 12, for example a cloud server, whereby it is again pointed out that direct communication between probe device 1 and evaluation unit 12 can also take place.

[0084] Initially, the probe device 1 is inserted into the gastrointestinal tract 3 of a farm animal 2, in the described embodiment into the reticulum of a cow.

[0085] While the probe device 1 is located in the gastrointestinal tract 3, the steps described below are carried out - partly individually, partly repeated several times in succession.

[0086] 20: Using a three-axis acceleration sensor 51 of the probe device 1, acceleration values ​​within the gastrointestinal tract 3 of the cow 2 are determined for the three axes of a Cartesian coordinate system—x-axis, y-axis, and z-axis—at successive measurement times. Thus, acceleration values ​​are available for the x-, y-, and z-axes.

[0087] The determination can be carried out, for example, with a sampling rate or a measurement frequency of 50 Hz. In other words, the multi-axis, in particular triaxial, acceleration sensor 51 determines acceleration values ​​at successive measurement times with a measurement frequency of 50 Hz.

[0088] The measurement by the acceleration sensor 51 preferably takes place continuously, i.e., at successive measurement times, at least over a predetermined measurement duration. In principle, the measurement can also be event-driven, whereby a threshold value—particularly for the acceleration—can be predefined for activation or the start of the measurement. The determined acceleration data are transmitted to the probe control unit 6, where the next steps take place. The acceleration values ​​are measured in the local coordinate system of the three-axis acceleration sensor 51 of the probe device 1. Since the probe device 1 drifts in the gastrointestinal tract 3 of the farm animal 2, the orientation of the local coordinate system in a global coordinate system changes over time, which must be taken into account accordingly.

[0089] 30: In the probe control unit 6 of the probe device 1, the acceleration values ​​for the x-, y-, and z-axes determined at the same measurement time are summarized to form a first aggregation value. The acceleration values ​​measured simultaneously at a measurement time are thus aggregated. To make subsequent evaluations independent of the orientation of the local coordinate system, the acceleration values ​​are summarized using a rotation-invariant method. In other words, a method is used that is invariant to rotations. This ensures that evaluations are independent of the orientation of the probe device 1 in the gastrointestinal tract 3 of the farm animal 2.

[0090] In the embodiment described here, the length of the acceleration vector r1 is determined at the respective measurement time. Therefore, the acceleration value for the x-, y-, and z-axes at the respective measurement time is used. This results in the following formula: r1 = 7x 2 +y 2 +z 2 , where “x”, “y” and “z” represent the measured acceleration value of the respective axis.

[0091] For each measurement time point considered, the acceleration vector r1 of the acceleration values ​​for the x, y, and z axes is obtained as the first aggregation value. The measurement time point can be characterized by using real-time information from the clock generator 53. The clock generator 53 can, for example, also output relative time information such as time or number of time units since activation of the probe device 1, since the start of a predetermined measurement duration, since the last data transmission to the evaluation unit 12, or the like. In other words, relative time information specified by the clock generator 53 can also be taken into account to characterize the measurement time point. The clock generator 53 can, as shown in Fig. 3, be provided separately in the probe device 1 or as part of the probe control unit 6 (not shown in Fig. 3), which is then designed as a microcontroller with an integrated RTC.In a further variant, absolute time information can be transmitted from the evaluation unit 12 to the probe device 1, which combines it with its own relative time information (number of cycles or “ticks”).

[0092] Using the clock generator 53 or the RTC, the signal from the acceleration sensor 51 is time-coded or compared with real time. UTC ("Universal Coordinated Time") is typically stored on the RTC.

[0093] The first aggregation values ​​of several considered measurement points, which result from repeated execution of step 30, form a first evaluation set.

[0094] 40: In order to identify the condition of the farm animal to be determined here and to filter out the factors influencing the acceleration measurement described above, a first characteristic value is determined from the first evaluation set in the probe device 1 (or the probe control unit 6 of the probe device 1). In other words, a first characteristic value is determined from the first evaluation set, which consists of several first aggregation values ​​resulting from repeated execution of step 30. The characteristic value is thus obtained from acceleration values ​​from several different measurement times, which advantageously follow one another directly.

[0095] In the embodiment described here, the mean and / or variance are determined from the first evaluation set or for the first evaluation set. This allows the quasi-local statistics of the first evaluation set to be determined. The mean and variance have the advantage that they can be used to detect temporal changes in acceleration, which allows conclusions to be drawn about the condition of the farm animal. As initial parameters, the mean and variance can be considered together or individually.

[0096] An iterative approach is used here in which step 40 is carried out repeatedly, whereby in each execution of step 40 the first evaluation set differs from the evaluation set used in the previous execution of step 40 by at least one first aggregation value.

[0097] This procedure is referred to here as the "sliding window" approach. The mean and / or variance as first key values ​​are determined iteratively across a section or window of chronologically successive first aggregation values, the first evaluation set. The section used is shifted in an overlapping manner, i.e. the earliest first aggregation value is repeatedly deleted from the section under consideration, the first aggregation value immediately following the latest value in the section under consideration (which was not yet part of the section under consideration) is added, and the mean and / or variance are again determined as the first key values ​​for this new section. For each iteration, the mean and / or variance are therefore determined for a separate first evaluation set that differs from the previous first evaluation set.

[0098] Essentially, this procedure can be represented by the following equations.

[0099] Mean: where Xi are the individual first aggregation values ​​and x t represents the mean value of the first evaluation set considered for each iteration.

[0100] Variance: where o t 2 represents the variance per iteration.

[0101] Dt denotes the domain for which the mean and variance are determined for each iteration. Accordingly, Nt=| Dt| and Nt denotes the number of first aggregation values ​​that form the respective first evaluation set, i.e., the size of the section used. Essentially, the aim of the described exemplary embodiment is for the first evaluation set to be formed from first aggregation values ​​that are based on acceleration values ​​determined during a measurement period of 1 second. In other words, Nt, i.e., the number of first aggregation values ​​that make up the first evaluation set, is selected such that it corresponds to the acceleration values ​​at measurement times that lie within a measurement period of 1 second. With a sampling or measuring rate of the acceleration sensor 51 of 50 Hz, this means that the section used for each iteration (or the first evaluation set used for each iteration) comprises 50 first aggregation values.

[0102] A particularly time- and energy-saving implementation can be achieved when the value for Nt corresponds to a power of two. To get as close as possible to a measurement duration of 1 second, this would require 64 initial aggregation values, which would constitute an initial evaluation set. Increasing the measurement frequency to 64 Hz would result in 64 initial aggregation values ​​corresponding exactly to a measurement duration of 1 second. This would make the sample size large enough to quickly capture significant changes in the signal statistics of the acceleration values, while at the same time making the sample small enough that other acceleration sources, such as motility, do not adversely affect the evaluation. A typical motility contraction lasts approximately 3 seconds and would therefore not interfere with the described procedure.

[0103] 40a: Before determining the first characteristic values, a downsampling of the first evaluation set can also be carried out in one variant, for example down to one fifth, in this case to 10 Hz. This variant is shown in Fig. 5 as a dashed box and is optional.

[0104] 50: To further reduce the amount of data to be transmitted from the probe device 1 to the evaluation unit 12, the first characteristic values ​​are summarized again. The goal is to down-clock the values ​​collected at a high measurement or sampling frequency without losing essential information about the condition of the farm animal. The multiple measurements per second are thus summarized over longer periods. For this purpose, a second evaluation set is formed in the probe device 1 or the probe control unit 6 of the probe device 1 from first characteristic values ​​resulting from repeated execution of step 40. In other words, the second evaluation set consists of first characteristic values ​​resulting from repeated execution of step 40.

[0105] This second evaluation set is then combined into a second aggregation value. This step 50 also takes place on the probe device 1 or in the probe control unit 6 of the probe device 1.

[0106] For the determination of the standing or lying state of the farm animal described in the present exemplary embodiment, it has proven advantageous if first characteristic values ​​derived from a measurement period of 1 minute to 15 minutes, but in particular 10 minutes, are used for the second evaluation set. This means that the first characteristic values ​​of the second evaluation set are based on acceleration values ​​determined by the acceleration sensor 51 during a measurement period of 1 minute to 15 minutes, in particular 10 minutes.

[0107] To summarize the values ​​of the second evaluation set, different methods can be used that map a larger amount of data to a single value, for example statistical parameters such as quantiles, determining the median, determining the minimum, determining the maximum or determining the mean of the second evaluation set.

[0108] The use of the median is particularly advantageous for distinguishing between the standing and lying states.

[0109] The following explains the implementation example of the procedure referred to here as the “median cascade” method.

[0110] This involves taking an input set of values ​​and dividing them into a number of value blocks. The median is then determined for each value block. The median (also called the central value) is the value in the middle of an input set sorted by size. This means that at least 50 percent of the data are less than or equal to the median, and at least 50 percent of the data are greater than or equal to the median. The medians then form another input set, which is divided into value blocks, for which the median is then determined. This process is repeated until only a single value block remains, resulting in a single median.

[0111] In the present embodiment, the second evaluation set is used as the input set at the beginning of the "median cascade" process. To enable a memory- and computation-time-optimized process, the size of the value block nßiock, i.e., the number of first characteristic values ​​that make up the value block, is selected according to the following formula: n B | 0Ck = measurement rate*60*measurement duration| .

[0112] This is because memory consumption is proportional to the sum of the block sizes, and the product of the block sizes must be proportional to the number of input values. The measurement rate refers to the measurement or sampling rate, also referred to elsewhere in this disclosure as the measurement frequency and specified as 50 Hz by way of example.

[0113] The measurement duration refers to the duration in minutes to be considered. As stated above, 10 minutes is assumed here. Thus, the square root is the expression "50 * 60 * 10."

[0114] In addition, the rounding function is applied, although other rounding functions are also possible as long as it is ensured that the number of the first characteristic values ​​results in an integer.

[0115] This results in nBock=31. However, nBock can also be considered as a quasi-free parameter, e.g., nBock=29, to adjust the results. In this way, an input set of size N, i.e., with N values, can be divided into m value blocks according to the formula

[0116] The rounding function is used here again, but can be replaced by another rounding function. The value riBiock is intended as a guideline. To optimize the required computing time, it is advantageous if the number of elements in the first value blocks of the cascade corresponds to a power of two. These value blocks determine the runtime because they process the full number of the first characteristic values.

[0117] As shown in Fig. 6, the median cascade method for the described embodiment can be reduced to three stages.

[0118] When using the second evaluation set, with a measurement frequency of 50 Hz and a measurement duration of 10 minutes, 30,000 initial characteristic values ​​result as input set 501, which is divided into a number of value blocks 502, which are shown only schematically and not completely in Fig. 6. Only one reference symbol "502" is shown, but each of the value blocks is referenced by this reference symbol.

[0119] For each value block 502, the median is then determined - 503 -, and the determined medians are then combined to form a new input set 501'. This is again subdivided into value blocks 502' (again shown only schematically and not completely; the reference symbol "502'" stands for the entirety of the value blocks and each of the individual value blocks), and the median is determined - 503' - for each value block 502'.

[0120] The new input set 501" now consists of 31 medians and thus forms exactly one value block 502", when determining the median - 503" - a single median results as the second aggregation value.

[0121] In this way, the high-frequency acceleration values ​​are reduced in such a way that the smallest possible amount of data needs to be transmitted from the probe device 1 to the evaluation unit 12. The described "median cascade" method enables this in a memory-efficient manner: Each stage contains a buffer of size nßiock (shaded squares in Fig. 6) to determine the medians. Therefore, the decision to distribute the buffer sizes evenly across the three stages minimizes the total memory required.

[0122] The second aggregation values ​​are preferably determined from consecutive second evaluation sets, so there is no overlap of the second evaluation sets that are used as input sets for the “median cascade” method.

[0123] 60: After the processing of the measured acceleration values ​​on the probe device 1 has been completed, in a further step at least one or more second aggregation values ​​are transmitted with the communication device 8 of the probe device 1 to the evaluation unit 12. Several second aggregation values ​​are determined by repeatedly performing step 50. In particular, a wireless transmission via radio with a correspondingly suitable protocol (e.g. LoRa, ZigBee, RFiD, WLAN, or others) is used here, preferably a suitable frequency range, e.g.

[0124] 300 MHz to 900 MHz, where the permeability to radio waves in animals is particularly high. This transmission can occur continuously, but in reality, transmission is not always possible or desired because the livestock 2 is not within the reception range of the evaluation unit 12 or the transmit / receive unit 13, or because, for energy-saving reasons, the data should only be sent in packets. In addition, so-called "duty cycles" are often defined, which determine how long data may be transmitted on a specific frequency, resulting in a further limitation of the transmission options.

[0125] For this reason, the memory element 7 (see Fig. 2) is designed as a RAM and / or ROM or programmable ROM. For example, the values ​​of the first state variable can be stored in the random access memory ("RAM") during a measurement interval, after which they are transferred from the temporary memory to an EEPROM, i.e., a read-only memory. The intermediate step with the EEPROM is particularly advantageous for storing sufficient amounts of data before the next transfer to the evaluation unit 12 is possible.

[0126] 70: In the evaluation unit 12, for example, implemented as a server or cloud server, the state of the farm animal is determined from the second aggregation values ​​transmitted in step 60. For each measurement period corresponding to a second aggregation value, the state of the farm animal 2 can thus be determined. As described, in the present embodiment, it is determined whether the farm animal 2 is in the standing or lying state.

[0127] For this purpose, a binary classification is applied, in which the second aggregation values ​​are compared with a threshold. The threshold can be fixed or variable, with the following threshold classification being preferred:

[0128] Stand if x(t)>threshold (t)

[0129] State (x(t))= Lying if x(t) <Schwellwert (t)’ der Schwellwert wird also dynamisch ermittelt und ist abhängig von der Zeit, es ergibt sich der Zustand zu einem bestimmten Zeitpunkt t.

[0130] Conveniently, the threshold represents the mean of two aggregation values ​​during an evaluation period. This evaluation period typically consists of several measurement periods and should also take into account the characteristics of the livestock as well as changes in the position of the probe device 1 in the gastrointestinal tract 3, which change the signal statistics in the medium term. In particular, an evaluation period of 12 hours is used, for example. 12 hours allows for the daily seasonality of the livestock to be taken into account. As described above, other, particularly longer, periods are also conceivable, which would be more stable against short-term changes in the underlying signals.

[0131] In one implementation variant, the dynamic threshold is implemented as an exponentially smoothed average, which allows for livestock-specific effects and seasonality, such as different behavior during the day and at night, to be taken into account. This is accounted for by the duration of the evaluation period used. The average is then determined iteratively, with the evaluation period used being shifted in an overlapping manner.

[0132] The result of step 70 is the determination of the state of the farm animal as standing or lying down.

[0133] 80: Optionally, in a further step, the status determined in step 70 can be output, either to a terminal device or to a database, as described in connection with Fig. 4. Since this step is optional, it is shown in dashed form in Fig. 5.

[0134] With reference to Figure 5, it should be noted that steps 20 to 50 are performed on the probe device 1, while step 70 takes place in the evaluation unit 12. Step 60 involves the transmission of data from the probe device 1 to the evaluation unit 12 and therefore takes place on both components of the system 100 according to the invention. Depending on the variant, step 80 takes place in the evaluation unit 12 (e.g., if a display unit or the database 14d is part of the evaluation unit 12) or separately.

[0135] The described method also includes a computer program product for determining a condition of the farm animal 2, comprising a first computer program product part and a second computer program product part, wherein the first computer program product part comprises instructions which, when the program is executed by the probe control unit 6 of the probe device 1, cause the probe control unit 6 to execute, in addition to step 20, steps 30, 40, optionally 40a, 50, and 60—insofar as it occurs in the probe device 1. The second computer program product part comprises instructions which, when the program is executed by an evaluation unit 12, cause the evaluation unit 12 to execute step 60, insofar as it occurs in the evaluation unit 12, step 70, and step 80, insofar as it occurs in the evaluation unit 12.

[0136] The method according to the invention enables the reliable determination of the condition of a farm animal based on acceleration values, whereby the computational effort and the amount of data to be transmitted can be kept as low as possible.

[0137] The method steps according to the invention on the probe device 1 reduce the amount of data to be transmitted, which reduces the transmission time, the associated power consumption on the probe device 1, and the required bandwidth, which is particularly advantageous when specifying "duty cycles." This ensures long-term, proper use of the probe device 1 with the best possible information exchange with the evaluation unit 12.

[0138] In the illustrated embodiment, the method allows the distinction between the standing and lying state of a farm animal; this information can then be used for further checks such as determining lameness, estrus or other farm animal characteristics.

Claims

PATENT CLAIMS 1. A method for determining a condition of a farm animal (2), wherein at least one probe device (1) with at least one multi-axis acceleration sensor (51) is arranged within a gastrointestinal tract (3) of the farm animal (2), and the acceleration sensor (51) determines acceleration values ​​(20) at successive measurement times, wherein at least one evaluation unit (12) is located outside the gastrointestinal tract (3) of the farm animal (2), and data can be transmitted between the probe device (1) and the evaluation unit (12), characterized by the following steps: A) combining (30) a plurality of acceleration values ​​determined at the same measuring time to form a first aggregation value in a probe control unit (6) of the probe device (1); B) determining (40) at least one first characteristic value from a first evaluation set, consisting of a plurality of first aggregation values ​​resulting from repeated execution of step A) on acceleration values ​​of a plurality of different measuring times, in the probe control unit (6) of the probe device (1); C) summarizing (50) a second evaluation set, consisting of first characteristic values ​​resulting from repeated execution of step B), to form a second aggregation value in the probe control unit (6) of the probe device (1), wherein each time step B) is executed, the first evaluation set differs from the evaluation set used in the previous execution of step B) by at least one first aggregation value; D) transmitting (60) at least one or more second aggregation values ​​determined by repeated application of step C) from the probe device (1) to the evaluation unit (12) using a communication device (8) of the probe device (1); E) Determining (70) a condition of the farm animal from second aggregation values ​​transmitted in step D) in the evaluation unit (12).

2. Method according to claim 1, characterized in that the acceleration sensor (51) is a three-axis acceleration sensor.

3. Method according to claim 1 or 2; characterized in that a rotation-invariant method is used when summarizing the acceleration values ​​in step A).

4. Method according to one of claims 1 to 3, characterized in that the acceleration values ​​are summarized in step A) by determining the length of the acceleration vector (r1) at a measuring time.

5. Method according to one of claims 1 to 4, characterized in that in step B) the mean value and / or the variance are determined from the first evaluation set as first characteristic values.

6. Method according to claims 1 to 5, characterized in that step B) is carried out several times, wherein each time step B) is carried out, the mean value and / or the variance are determined as first characteristic values ​​for a first evaluation set, wherein each time step B) is carried out, the first evaluation set differs from the first evaluation set used in the previous implementation of step B) by at least one first aggregation value.

7. Method according to one of claims 1 to 6, characterized in that a first evaluation set is formed from first aggregation values ​​which are based on acceleration values ​​determined during a measurement period of 1 second.

8. Method according to one of claims 1 to 7, characterized in that the number of first aggregation values ​​of a first evaluation set corresponds to a power of two.

9. Method according to one of claims 1 to 8, characterized in that in step B) before determining the first characteristic value from the first evaluation set, a down-clocking of the first evaluation set is carried out.

10. Method according to one of claims 1 to 9, characterized in that in step C) the second evaluation set is formed from first characteristic values ​​which are based on a measurement period of 1 minute to 15 minutes, preferably of 10 min, determined acceleration values ​​decrease.

11. Method according to one of claims 1 to 10, characterized in that the summarization of the second evaluation set in step C) is carried out by at least one of the following methods: determination of a quantile, determination of the median, determination of the minimum, determination of the maximum, determination of the mean.

12. Method according to one of claims 1 to 11, characterized in that the summarization of the second evaluation set in step C) is carried out by a median cascade method, which comprises the repeated execution of the following steps: C1) dividing an input set (501, 501', 501") of first characteristic values ​​into a number of value blocks (502, 502', 502") each consisting of a plurality of first characteristic values; C2) For each value block (502, 502', 502"), determining the median of the first characteristic values ​​assigned to the value block (502, 502', 502"); C3) summarizing the medians determined in step C2) to form a new input set (501, 501', 501") and re-executing steps C1) and C2) using this new input set (501, 501', 501"), wherein steps C1) to C3) are repeated until only a single block of values ​​(502") results in step C1) and step C2) results in a single median, which then forms the second aggregation value, and wherein the second evaluation set is used as the input set (601) for the first execution of step C1).

13. The method according to claim 12, characterized in that the size of the value block (502, 502', 502") in step C1) is determined as the cube root of the number of first characteristic values ​​that make up the input set (501), where preferably a rounding function is applied to the result of the cube root of the number of first characteristic values ​​that make up the input set (501 ).

14. Method according to one of claims 1 to 13, characterized in that in step E) it is determined whether the farm animal (2) is in the standing state or in the lying state.

15. The method according to any one of claims 1 to 14, characterized in that a binary classification is used to determine the state of the farm animal (2) in step E), wherein the state of the farm animal (2) is qualified as standing if the transmitted second aggregation value is greater than a threshold value, and the state of the farm animal (2) is qualified as lying if the transmitted second aggregation value is less than or equal to a threshold value.

16. The method according to claim 15; characterized in that the threshold value represents the mean value, preferably the moving mean value, of second aggregation values ​​during an evaluation period.

17. System for determining a condition of a farm animal (2), comprising: - at least one probe device (1) which can be arranged in the gastrointestinal tract (3) of the farm animal (2) and has at least the following components arranged in a housing (4): - at least one multi-axis acceleration sensor (51 ), - at least one probe control unit (6), and - a communication device (8) for wirelessly transmitting and receiving data, and - at least one evaluation unit (12) arranged outside the gastrointestinal tract of the farm animal, wherein a method according to one of claims 1 to 16 can be carried out with the system.

18. A computer program product for determining a condition of a farm animal (2), comprising a first computer program product part and a second computer program product part, wherein the first computer program product part comprises instructions which, when the program is executed, are executed by a Probe control unit (6) of a probe device (1) causes it to carry out steps A), B), C) and D) of the method according to one of claims 1 to 16, and the second computer program product part comprises instructions which, when the program is executed by an evaluation unit (12), cause it to carry out steps D) and E) of the method according to one of claims 1 to 16.

Images