Patent Application
20230270077
The present invention relates to a method for determining in a young animal a phase transition from a first development phase to a second development phase. The invention also relates to a method for training a self-learning data processing model for use in a method as described above. Further, the invention relates to a livestock management system.
In managing livestock in livestock farming, establishing phase transitions in the development is important, for instance to be able to tune the care needs to them. For instance, young animals are typically divided into groups, where the animals within each group have more or less similar care needs, as in respect of feeding or particular health checks. Determining whether a young animal is due to be put in a different group is conventionally done on the basis of age.
Also in quite young animals, the needs are tuned to the development phase, this development phase being estimated on the basis of age. This is for instance done in the weaning of animals. Weaning means: the transition of a wholly or partly milk-fed animal to an animal that gets no milk anymore, but different foodstuffs. In the current situation, the moment of transition from milk feed to completely different feed is mainly determined by the age of the animal and a more or less subjective observation of the farmer who determines whether the animal is ready for weaning.
This method has inherent drawbacks. Weaning an animal too soon or too late may lead to damage. Weaning is a risk moment for the animal's health and resistance. Weaning too soon leads to damage in that the animal is not yet sufficiently able to absorb solid food to compensate for the milk nourishment coming to an end. Weaning too late, however, is not desirable either, the milk nourishment is more expensive than solid food.
Also in the above-mentioned example of transfer to another group, the conventional method leads to unwanted effects. Housing too young animals among older animals leads to competition and stress, and may cause diseases or reduced growth. As yet, however, there are no methods or systems available with which such phase transitions, which are indicative of a transition moment such as weaning or a transfer to another group, can be objectively established.
It is an object of the present invention to remove the above-mentioned disadvantages of the prior art in any case in part and to provide a method which enables objective detection of phase transitions in the development of a young animal.
To this end, the invention according to a first aspect thereof provides a method for determining in a young animal a phase transition from a first development phase to a second development phase, the method comprising: with a measuring instrument, measuring during a period of time one or more body or behavior parameters of the animal and producing one or more measuring values for the or each measured body or behavior parameter; with a controller, receiving the one or more measuring values and detecting the phase transition on the basis of the measuring values; wherein the step of detecting the phase transition comprises a step of, with the controller, processing the or each measuring value and, depending thereon, determining an extent of the animal's attention to solid food, for on the basis thereof detecting the phase transition.
The method according to the present invention makes use of a measuring instrument for measuring body or behavior parameters. This may be a measuring instrument such as a camera, with which the behaviors of an animal can be established, or a measuring instrument which is attached to the animal (for example, an ear tag, leg tag, neck tag, tail tag) or introduced into the animal (for example, a stomach bolus or other introduced sensor). The measuring instrument produces measuring values which are processed by the controller, and which make it possible to objectively establish the extent of the animal's attention to solid food. Thus, on the basis of camera pictures or a proximity sensor, it can be established how often and for how long an animal is in the vicinity of a feeding trough, or how often and for how long an animal suckles its mother. On the basis of these measurements, the extent of attention can be established qualitatively (for instance in categories: ‘much’, ‘average’, ‘little’, or ‘never’) or quantitatively (in a numerical value or an analog measuring signal). This can be used as an objective observation to establish a phase transition. In this way, the method according to the invention enables a livestock farmer to optimally establish a transition moment or transition period (start of weaning, transfer into a different group).
The method according to the invention can be used for identifying phase transitions in animals, and in that regard is not limited to a specific kind of animals. Within livestock farming, the method can be used advantageously, for instance to keep track of the development of piglets, calves, foals, lambs or other young animals. However, the method can also be used in breeding programs, not necessarily within livestock farming. In the following, frequently reference will be made to application examples in calves, but the invention is not limited thereto.
In some embodiments, the step of detecting the phase transition comprises a step of, with the controller, comparing at least one or each measuring value with a limiting value, and establishing the phase transition when the at least one measuring value has exceeded or fallen below the limiting value. The one or more body or behavior parameters may for instance be chosen from a group comprising: eating time, amount of consumed solid food, amount of milk drunk, rumination amount, number of rumination boluses and number of rumination strokes, time spent on drinking milk, ratio between eating time and time spent on drinking milk, ratio between eating time and rumination time, amount of absorbed feed with respect to absorbed amount of solid food, weight, development, height, width, activity, shape, rumen filling, hygiene score, locomotion score, water drinking, time spent in different fields of interest such as feeding fence, concentrate box, water trough, hay rack, and milk issue point. Such parameters can represent absolute or relative values, in the latter case for instance a relative eating time (with respect to all time spent on food), a relative amount of consumed solid feed, or a relative rumination amount with respect to a total amount of feed. A phase transition may for instance be detected by comparing one or more of the above parameters with a limiting value. For instance, if the time spent on drinking milk falls below a limiting value, this can be seen as an indication of a phase transition. In addition, it may be determined whether a particular combination of such conditions is met, such as: the time spent on drinking milk below a limiting value and the relative amount of consumed solid feed above a limiting value, or other combination of conditions.
Time spent in various fields of interest such as spots where solid food (roughage, concentrated feed) is obtainable, or places where water is obtainable, etc., can be measured with a position determination system. Thus, the time can be determined that is spent in these fields or, conversely, the time that is not spent in these fields. Also, specifically an area where milk can be furnished may be qualified as a field of interest, such as an automatic drink dispenser, bottle, the mother's udder, etc. Time that is spent here is also indicative of transition to the weaning phase.
According to some embodiments, the controller is configured for implementing a data processing model, wherein the step of detecting the phase transition comprises a step of, with the controller, inputting at least one of the measuring values into the data processing model and determining with the data processing model whether an amount of solid food that has been consumed by the animal has exceeded a threshold value. On the basis of parameters such as eating time, rumination amount, number of rumination boluses and number of rumination strokes, an indication can be obtained of an amount of solid feed consumed by the animal. This amount may in itself be compared with a limiting value to objectively establish whether there is a phase transition in the animal. Thus, according to some embodiments, the phase transition may for instance be established when the amount of solid food that has been consumed by the animal is at least 0.1 kilogram a day. For different animal species, different limiting values may be applied. For instance, for calves the limiting value for the amount of solid food that has been consumed by the animal can be at least 500 grams (0.5 kilogram) a day, whereas for piglets a limiting value may be used of, for example, 50 grams a day. For other animals, the phase transition may be established when the amount of solid food that has been consumed by the animal is at least 0.5% of a body weight of the animal, a day. The limiting values may for instance be settable for a user, and the above-mentioned values are only indicative.
In an embodiment, the data processing model is for instance a self-learning data processing model, where the self-learning data processing model has been trained in a training method through input of a training set consisting of training measuring values of at least one of the one or more body or behavior parameters and result values associated with the training measuring values, the result values comprising at least one of: a Boolean value indicating whether the amount of solid food that has been consumed by the animal has exceeded a threshold value; an expectation value of the amount of solid food that has been consumed by the animal. By training a self-learning data processing model in this way, such model, once trained, can be used for, on the basis of these measurements, establishing for instance an eaten amount of solid feed or when this has exceeded the limiting value for establishing the phase transition. In an embodiment, the self-learning data processing model is at least one from a group comprising: a neural network, a random forest algorithm, and an arithmetic regression model such as a simple or multiple linear regression model or a simple or multiple nonlinear regression model.
A self-learning data processing model as described above may be trained on the basis of different measurable parameters. For this, for instance, the one or more body or behavior parameters can comprise at least one from a group comprising: eating time, rumination amount, number of rumination boluses and number of rumination strokes, heart rate, heart rate variation, oxygen saturation level, and breathing frequency, weight, shape (condition), time spent in fields of interest. Such parameters, which are related to the activity of the animal (for instance, whether it is eating) or to the metabolism, may in combination be indicative of, for instance, an eaten amount of solid feed or when such amount has exceeded the limiting value.
According to some embodiments, the animal is a calf and the first development phase is a suckling phase and the second development phase is a weaning phase. Being able to objectively determine an optimum moment or optimum period of time when the calf should be weaned is valuable to prevent insufficient food absorption or undernourishment as a result of premature weaning, and also to be able to prevent milk feeding being continued too long.
In some of these embodiments, the method further comprises the step of, depending on the determined extent of the animal's attention to solid food, providing an indication signal for running down the suckling phase. The indication signal has as an advantage for the farmer, especially in keeping a large group of animals, that the signal facilitates the processes in the operation. For instance, the signal may be used for the purpose of, for example, automatic or semiautomatic separation of animals, or it may be incorporated in a daily overview in which it is indicated to the farmer which calves are ready to be weaned.
In some embodiments, the method comprises the step of, depending on the determined extent of the animal's attention to solid food, determining an advised value of an amount of milk per period of time that can be offered to the animal. In this way, on the basis of the determined extent of attention to solid food, the milk consumption of the animal can be gradually run down, so that it is tuned to the animal's individual need.
According to other embodiments, the phase transition is accompanied by a displacement of the animal from a group of animals in a first age category to a group of animals in a second age category, and wherein the method further comprises the steps of, after displacement of the animal, monitoring at least one of the one or more body or behavior parameters, and determining therefrom a state of health of the animal, such as, for example, a stress condition of the animal. Because such a transfer may cause stress in the animal, for instance when transfer has taken place too early after all, monitoring the behavior of the animal after transfer has added value.
According to other embodiments, the method further comprises producing an attention signal when the amount of solid food that has been consumed by the animal has exceeded a threshold value. The amount of solid food that an animal eats is a direct measure of the extent of attention to it and, when determined, forms a good indication for execution of the method.
According to some embodiments, the measuring instrument is at least one from a group comprising: an ear tag, a neck tag, a leg tag, a stomach bolus, a tail sensor, a camera, or a microphone. Such measuring instruments may be provided with units for measuring a variety of behavior and body parameters on the basis of which the method according to the invention can be implemented. Thus, the tags concerned that are worn by the animal (on or in the animal) may be provided with an accelerometer or other movement sensors which can provide insight into specific movements associated with particular behaviors. It may for instance be established whether an animal is eating or is stressed, or is ruminating. According to some embodiments, the sensors comprise at least one sensor from a group comprising: a movement sensor, a heart rate sensor, a breathing sensor, an optical sensor for measuring one or more blood levels, a positioning system for determining a current position, a pressure sensor, or a chromatograph.
According to some embodiments, the one or more body or behavior parameters comprise at least the number of rumination boluses, and the measuring instrument is designed with a movement sensor, in which the number of rumination boluses is measured by at least one of the following steps: with the movement sensor, measuring animal movements, analyzing the measured animal movements for distinguishing movements that are indicative of rumination strokes, recognizing a rumination stroke pattern consisting of one or more series of rumination strokes which are each followed by a pause in which no rumination strokes are observed, and counting at least one of the number of series of rumination strokes or the number of pauses for establishing the number of rumination boluses, wherein each series or each pause represents one rumination bolus; or, with the movement sensor, measuring animal movements, analyzing the measured animal movements for distinguishing movements that are indicative of at least one of regurgitations or swallowing movements, and counting the number of regurgitations or swallowing movements for establishing the number of rumination boluses. Rumination boluses give a specific movement signal that can be properly established with movement sensors (for example on the neck or the ear of the animal, or with a stomach bolus).
According to some embodiments, the measuring instrument is physically and communicatively connected with the controller. The controller may for instance be in the neck tag or the stomach bolus (or other instrument) and for instance process the measurements there directly. For instance, in some embodiments, the measuring instrument and the controller are part of a same device which is attached to or introduced into the animal. When the measuring instrument is a camera, the controller may also be present in or at it. According to some embodiments, the controller is contained in a livestock management server, with the measuring instrument sending the measuring values via a wireless connection to the controller. In this last variant, the measuring values are centrally processed. Also, a mixed form may be used, where measuring values, for instance, are first worked up in the measuring instrument, and then further analyzed centrally in the server. It is also possible that the server is only communicatively in communication (via a wired or wireless connection, for instance via a data communication network) with a controller or with another entity in which the controller is present. Further, the server may or may not be locally present, for example a local livestock management server of a farm, or work may be done with a remote server which receives the measuring values by means of a data communication network, such as a server in the cloud.
As mentioned earlier, the measuring instrument may be attached to or introduced into the animal, and may be provided with one or more sensors. Alternatively or additionally, the measuring instrument may be a camera which is operatively connected with an image recognition system for establishing the body or behavior parameters. A further alternative is the use of a microphone as measuring instrument. A microphone may then be placed in a barn or monitored field, optionally in combination with a camera. Also, merely a camera may be in place. It is also possible that a microphone as a sensor is part of a tag. Similarly to the camera, also the microphone may be used for measuring body or behavior parameters, for instance by recognizing sounds indicating a particular state of mind or activity of an animal.
According to a second aspect, the present invention relates to a method for training a self-learning data processing model for use in a method according to the first aspect. The training method comprises the steps of: inputting into the data processing model a training set consisting of training measuring values, wherein the training measuring values comprise one or more body or behavior parameters and result values, wherein the result values are associated with the training measuring values, and wherein the result values comprise at least one of: a Boolean value indicating whether the amount of solid food that has been consumed by the animal has exceeded a threshold value; an expectation value of an amount of solid food that has been consumed by the animal. Such a trained self-learning data processing model can be used with advantage in a method according to the first aspect.
According to a third aspect, the present invention relates to a livestock management system configured for executing a method according to the first aspect for, in each individual animal of a group of young animals, determining a phase transition from a first development phase to a second development phase, the system comprising a plurality of measuring instruments attachable to or introducible into an animal of the group of animals and wherein the measuring instruments are each provided with one or more sensors for measuring during a period of time one or more body or behavior parameters of the respective animal, and producing one or more measuring values for the or each measured body or behavior parameter; and a controller configured for receiving the one or more measuring values of each measuring instrument and detecting the phase transition on the basis of the measuring values; wherein the step of detecting the phase transition comprises a step of, with the controller, processing the or each measuring value and, depending thereon, determining whether an amount of solid food that has been eaten by the animal has exceeded a threshold value.
The invention will be discussed below on the basis of specific embodiments thereof, not intended as limiting, with reference to the appended figures, in which:
In the example of
From the data gathered with the sensors in each of the measuring instruments 13-17, body or behavior parameters may be determined and be sent as measuring values to the system 85. Thus, for instance, with movement sensors that are in the stomach bolus 16, movements may be measured that can be traced back to rumination activity of the animal 6 concerned. Rumination can also be established, for instance, with movements that are determined by the neck tag 14 or the ear tag 13. In another example, when the cow 6-2 brings down its head to eat the solid feed 12 from the feeding trough 11, such movement is observable with the movement sensors in, for instance, ear tag 13 or neck tag 14. Also, the stomach bolus 16 will be able to recognize stomach movements that are associated with the eating of the solid feed 12. With the tail tag 15, urinating or defecation by the cows 6 may be established, and possibly even an amount of excreted urine or dung. The calf 6-1, in the example of
According the present invention, for the purpose of determining a phase transition, with one or more measuring instruments 13-17 a few body or behavior parameters of the animals 6-1 and 6-2 are measured. The use may depend on the age category of the animal. For instance, when a very young animal is involved, such as a calf, the method according to the present invention may be used for detecting the phase transition of weaning. In weaning, a young animal makes a transition from bottle feeding or suckling its mother, to solid food 12. Initially, the solid food 12 will chiefly consist of concentrated feed (concentrate), but in the course of time the animal 6-2 will be eating roughage more and more. Not only with one of the measuring instruments 13-17 but also with the camera 4 (which also forms a measuring instrument here) can the behaviors of the animals 6-1 and 6-2 be recorded. As already indicated above, when the calf 6-1 gets bottled milk 10, it will keep its head in a particular position. This is detectable by the movement sensors in for instance the neck tag 14 or the ear tag 13. Even when the calf has just one of the measuring instruments 13-17 available (for example, the ear tag 13, the neck tag 14 or the stomach bolus 16), this movement can be established. But also when in a livestock farm no use is made of tags 13-17 attached to the animal, but, for instance, merely the camera 4 is used, it is possible to recognize the behaviors of the animals 6-1 and 6-2. With image recognition, it can be deduced that the calf 6-1 is getting bottle-fed. Also when the animal 6-2 eats from the feeding trough 11, this can be recognized, on the basis of image recognition, in the pictures of camera 4. In a livestock farm where exclusively a camera 4 is employed, there may for instance be several cameras present to be able to properly monitor all remote corners of the barn. Additionally, and also in the use of a camera 4, it is necessary that, either from the gathered video material or in any other way (for instance in an embodiment that does involve additional use of tags 13-17), the animals 6-1 and 6-2 be identified, for them to be registered in the livestock management system 85.
The livestock management system 85 may include a controller 91 and optionally an internal or external memory 88. As has already been noted above, the controller 91 does not necessarily need to be present in the livestock management system 85, as is the case in
According to the invention, the self-learning data processing model is trained in that, at the input 43, the desired body and/or behavior parameters that can have been obtained with the measuring instruments 4, 13-17 which are for instance shown in
In the training phase, the output 48 is fed back via 51 to the controller 91. Via arrow 52, the controller 91 also receives all input values 43 which have been presented to the self-learning data processing model 40. These are the body and/or behavior parameters which have been measured with the measuring instruments 4, 13, 14, 15, 16 and/or 17. Because the controller 91, during the training phase, also has available the actual verification data regarding the occurrence of a phase transition in the respective animal 6-1, the controller 91 can adapt the decision model 45. This has been visualized in
In
The body or behavior parameters that may be gathered by any measuring instruments 13-17 are sent as measuring values 80 via the transceivers 81 to the server 85, and are presented as input to the self-learning data processing model 40, or to another arithmetic algorithm which processes the data. The established measuring values may for instance be compared with limiting values to enable establishing whether a phase transition has occurred. In a specific embodiment, it may for instance be deduced from the body or behavior parameters what the amount of solid food that has been eaten by each of the animals 6 is, and from this amount the extent of attention to solid food can be established for each of the animals 6 individually. This may be converted to a numerical value. The amount of solid food eaten by the animal 6 may be compared with a limiting value to establish whether a phase transition has taken place. Also, it is possible to compare different body or behavior parameters with limiting values to determine a phase transition on the basis of a combination of conditions. The server 85 thereupon determines the occurrence of a phase transition, and can record this in its internal memory 90, for instance in the livestock management records. Also, it is possible that livestock management system 85 generates an attention signal, which is for instance wirelessly transmitted to the farmer's cell phone or is brought to the farmer's attention in a different manner.
In yet another embodiment, shown in
In step 65 the processed measuring values are compared with a limiting value 66. For instance, a measured time duration of the time spent by an animal on milk consumption 10 may be compared with a limiting value 66, to establish whether the measured time duration falls below the limiting value 66. When the measured time duration is lower than the limiting value 66, this may be an indication that the animal pays more attention to solid food. Also, a combination of measuring values may be compared; for instance, the time duration spent on eating may be compared with a limiting value, while also the rumination time that the animal 6 spends on ruminating is measured and compared with the limiting value 66. For unweaned animals, which are still being raised entirely on milk consumption 10, it holds that neither eating time nor rumination time is measured, since the milk is not chewed, hence the animal spends no time doing so. However, when the animal slowly starts to eat solid feed, the eating time will increase (since the animal 6 now eats solid feed). In ruminants, also rumination time now starts to increase. The eating time and the rumination time may therefore both be compared with a limiting value again to establish whether the animal already eats solid feed and attention to it increases. When the animal 6 at some point is going to eat roughage alone, the eating time increases, and also the rumination time increases. By comparing this eating time and rumination time once again with limiting value 66, it can therefore be established in what measure the animal 6 is eating roughage already. On the basis thereof it can be reasonably accurately predicted whether a phase transition is taking place. In step 65, such comparisons with limiting values take place. Thereupon, in step 67, it is determined whether the data from step 65 show that a phase transition is taking place. When this is not so, the method proceeds with step 60 in which the body and behavior parameters are measured. If a phase transition is taking place, the method proceeds with step 69. In step 69, optionally an attention signal is produced by the livestock management system 85. The livestock management system 85 can draw the farmer's attention to the circumstance that a respective animal 6 is experiencing a phase transition, and that appropriate measures should therefore be taken. What these measures are, depends on the phase transition to be considered. Thus, the system may be used to see whether calves should be weaned, but it may also be used for migrating animals 6 to a new (older) group. The associated measures are slightly different in the two cases, as will be further explained in the following.
In step 70, the system, depending on the phase transition to be considered, can give advice regarding the measures to be taken. Thus, it is possible that the system 85 in step 70 has a run-down program for running down the amount of milk an animal 6-1 gets offered per day. In this way, weaning of the animal 6-1 can take place in a gradual manner. For instance in the case of migration to an older group, the system 85 can indicate a period of time within which such a migration should take place. In this manner, the system may for instance make a prediction of the different group sizes at different points in time.
The generic part of the method ends after step 70, as is indicated in terminal 71. The method can continue in different manners, depending on the phase transition to be detected. It is also possible that the system 85 implements the method to cause different phase transitions for different animals to be recorded. Thus, all young animals within the farmer's business can be monitored with the same system, keeping track, where the older animals are concerned, of when they are up for relocation to a different group and, where the younger animals are concerned, of when they should be weaned. Depending on the phase transition to be detected, the method continues via arrow 72 or arrow 73. When the phase transition concerns the weaning of animals, the method continues with step 74. In step 74 the system 85 provides, according to the method, advice with respect to the running down of the milk program for the respective animal. On the basis thereof, in step 75, the animal may for instance be transferred to a group of animals that are being weaned. Thereafter, optionally, in step 76, the animals concerned that are being weaned may be monitored with the aid of the measuring instruments 4 and 13-17.
When the phase transition concerns the relocation of the young animal 6 to an older age-group, in step 77 a period of time may be indicated within which the migration can take place optimally. In step 78, the relocation actually takes place, for example by automatic separation method, or so that the farmer himself puts the animal in the older group. Next, in step 79, a stress monitoring program is implemented, whereby the body or behavior parameters such as heart rate and/or unexpected movements are recorded to establish whether the animal 6 is suffering from stress. When the animal after transfer suffers from stress, it may be that the animal, despite the advice, is not yet ready for relocation, after all.
The above-described specific embodiments of the invention are intended for illustration of the principle of the invention. It is believed that the implementation and the operation of the invention are readily apparent from the foregoing description and the appended illustrations. The invention is not limited to any embodiment described or shown herein. For the sake of clarity and conciseness of the description, features have been described herein as part of the same or of separate embodiments; it will be clear to a person skilled in the art that embodiments comprising combinations of any or all of the features described also fall within the scope of protection of the invention. Within the ability of those skilled in the art, alterations are possible which are to be understood to be within the scope of protection. Also, all kinematic inversions are understood to be within the scope of protection of the present invention. Expressions such as “consisting of”, when used in this description or the appended claims, should be construed not as an exhaustive enumeration but rather in an inclusive sense of “at least consisting of”. Indications such as “a” or “one” may not be construed as a limitation to just a single specimen, but have the meaning of “at least a single specimen” and do not preclude plurality. Expressions such as: “means for . . . ” should be read as: “component configured for . . . ” or “member constructed to . . . ” and should be construed to cover all equivalents of the structures described. The use of expressions such as: “critical”, “advantageous”, “preferably”, “desired”, et cetera, is not intended to limit the invention. Moreover, also features that are not specifically or expressly described or claimed in the construction according to the invention but do lie within reach of the skilled person, are understood to be encompassed without departing from the scope of protection as defined by the appended claims.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2025991 | Jul 2020 | NL | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/NL2021/050417 | 7/2/2021 | WO |
