A system and method for tracking animals

Abstract

The present disclosure relates to animal surveillance system (100) for an animal cage, the animal surveillance system (100) comprising: a physical platform divided into a grid of cells (106), wherein each cell (106) comprises at least one load cell (108); an animal identification system; a processing unit configured to: identify at least one animal on the physical platform using the animal identification system; continuously obtain load signals from the load cells (108); based on the load signals from the load cells (108), tracking positions and / or change of positions of the identified at least one animal animals on the physical platform. The present disclosure further relates to a method of surveying an animal in an animal cage.

Description

[0001] A SYSTEM AND METHOD FOR TRACKING ANIMALS

[0002] The present disclosure relates to an animal surveillance system for an animal cage. The disclosure further relates to a method for animal surveillance.

[0003] Background

[0004] Rodents are important as model organisms in medical and pharmaceutical research since their metabolism resembles that of human subjects, and because they are susceptible to many of the conditions and infections that afflict humans. The research areas where rodents are used include genetics, developmental biology, cell biology, oncology, and immunology. Therefore, a big amount of medical and pharmaceutical researchers’ time is being spent on tracking, identifying, handling, and extracting data from their rodents.

[0005] The state of the art in handling rodents involves tracking using, for example, microchips, which is an invasive method, or infrared tracking methods. Research indicates that animals may react to being in proximity to humans, which can introduce experimental errors and influence scientific results. Generally, the lack of reliable and digital devices for tracking and managing rodents during animal testing may slow down scientific progress and affect experimental data quality negatively. A further consequence of suboptimal animal management is that more animals than necessary will be required for a specific experiment.

[0006] Thus, there is a need for an improved system for tracking of rodents.

[0007] Summary

[0008] It is an object of the present disclosure is to provide a better solution for researchers handling rodents or other animals for scientific experiments. Specifically, the disclosure relates to an animal surveillance system for an animal cage where the animal surveillance system comprises a physical platform divided into a grid of cells, the cells being delineated such that each cell comprises at least one load cell. The platform may be adapted to be placed or integrated in an animal cage of EU Type 3 for mice. However, the animal surveillance system may be placed in any type of cage and for any animal. The animal surveillance system may be a system that fits into an existing cage, but may alternatively be integrated into the cage when the cage is built or assembled. The animal surveillance system may further comprise an animal identification system. The animal identification system may be, for example, a radiofrequency identification (RFID) based and / or camera-based identification system. The animal surveillance system further comprises a processing unit configured to: identify at least one animal on the physical platform using the animal identification system, continuously obtain load signals from the load cells, and based on the load signals from the load cells, and tracking and / or changes of positions of the identified at least one animal. By using the load signals from the load cells and an auxiliary system for identification of individual animals, the weightsystem can distinguish between animals and track the animals in the cage. Besides allowing the system to track animals over time, the weight data itself may be used for further data processing.

[0009] The researcher will thereby be able to use this cage equipped with a surveillance system to monitor a plurality of animals. The system may typically initially measure and register the weight of each rodent entering the grid. For example, each animal may be weighted individually on a scale. Alternatively, the animals may be weighted on the surveillance system. That reference value may then be compared continuously with new input values of each load cell, where the animal is moving. The above process is achieved by using a computer software which handles the data on each cage.

[0010] The load cell can measure the rodents’ weight by using various methods such as springs below each load cell, or a voltage fluctuation caused by the strain of the load cell when an animal is placed on it. A load cell shall be construed broadly to comprise any suitable type of weight sensor that can be used in the surveillance system. A load cell, generally, converts a force such as tension, compression, pressure, or torque into an electrical signal that can be measured and standardized. It may comprise a force transducer. As the force applied to the load cell increases, the electrical signal changes proportionally. The load cells may be, but are not limited to, resistance-based load cells, such as piezoelectrical load cells and strain gauge load cells. Piezoelectrical load cells and strain gauge load cells are built on similar principles. When a shape of the strain gauge is altered, a change in its electrical resistance occurs. In a piezoelectrical load cell a voltage output is also proportional to the deformation of the load cell.

[0011] The presently disclosed animal surveillance system may be capable of extracting activity changes of the animals over time by continuously measuring the weight on each load cell and translating the weight change to an animal moving across the grid. In addition, it is possible to estimate the weight distribution of a plurality of animals, by employing a sheet, typically a thin sheet, which covers the grid of cells. The thin sheet allows the weight of an animal that is placed on a first load cell to be distributed to additional load cells, such as the neighboring load cells. Being able to observe load differences not only in the cell that the animal is placed in provides a more detailed picture of where the animal is located. This can be used for, for example, the monitoring of the activity of an animal, such as shaking, moving in circles or standing on two legs while the animal remains on the same load cell. It is also possible to, for example, estimate the distance moved over a given time interval for an animal or the frequency of movement among cells based on the data from the load cells. This information can be further used for extracting and / or classifying expected or unexpected behaviors of the animals.

[0012] By computing the activity of an animal over time, the time an animal is resting can be extracted. A movement pattern may also be used to compute whether an animal is moving material across the grid. It may also provide information of how an animal is behaving in contact to other animals in comparison to being alone. All the above information can be extracted by continuously measuring the weight of each load cell, identifying each animal, and tracking their movement in the cage.

[0013] The surveillance system may be adapted to fit into an animal cage of Ell Type 3 for mice. The surveillance system may, for example, have the dimensions 375 mm to 400 mm length and 215 mm to 240 mm width. The grid of cells may be diagonally oriented with respect to the substantially rectangular shape, and / or can have a general orientation differentiating at least 10° with respect to the substantially rectangular shape.

[0014] The present disclosure further relates to a method for surveying a plurality of animals in an animal cage. The principle of load cells can be extended to any type of animal which may need to be surveilled by scaling up the cage and changing the capacity of the load and sensitivity of the load cells.

[0015] In summary, the present disclosure allows researchers to continuously collect data about the weight and behavior of animals, in particular rodents, in an efficient way. Using the above disclosure it is possible to gather more data accurately while spending less time and using fewer animals. At the same time, no animal tagging is required nor manual weighting of each animal, which can otherwise be an invasive and time consuming task respectively.

[0016] Description of the drawings

[0017] Various embodiments are described hereinafter with reference to the drawings. The drawings are examples of embodiments and are intended to illustrate some of the features of the presently disclosed animal surveillance system, and are not limiting to the presently disclosed method and system.

[0018] Figs. 1 A-B show a schematic top-view and side-view respectively of the surveillance system.

[0019] Figs. 2A-B illustrate the surveillance system with the grid which comprises multiple load cells.

[0020] Figs. 3A-B illustrate an embodiment of the surveillance system with the grid which comprises multiple load cells and RFID interrogation units.

[0021] Fig. 4 illustrates an example of a surveillance system with a diagonally oriented grid pattern.

[0022] Fig. 5 shows an example of a flow chart of a method of surveying an animal.

[0023] Fig. 6 describes an example of extracting the movement of an animal over time by continuously measuring the signals given by the load cells.

[0024] Fig. 7 shows an embodiment of a surveillance system equipped with a software interface.

[0025] Fig. 8 shows an exploded perspective view of the surveillance system.

[0026] Fig. 9 illustrates an example of a schematic view of the layer in which the load cells can be placed.

[0027] Fig. 10 shows an example of weight distribution elements disposed under the top sheet.

[0028] Fig. 11 shows a cross-sectional view of an embodiment of the presently disclosed animal surveillance system.

[0029] Figs. 12A-B shows two examples of animal surveillance using the modelled grid of cells.

[0030] Detailed description

[0031] The present disclosure relates to an animal surveillance system for an animal cage. A conceptual example of such an animal surveillance system 100 is shown in Fig. 1A. An example of a grid of cells 106 can be seen in Fig. 2A, where each cell 106 comprises at least one load cell 108. An example of a side-view of the load cells 108 can be seen in Fig. 2B. In the specific embodiment of Fig. 2, the animal surveillance system 100 comprises a physical platform divided into a grid of cells 106. In the example there are 3x3 cells. A “grid of cells” shall be construed to comprise any suitable division of an area into smaller sub-areas, which are the “cells”. Typically, but not necessarily, the cells will have the shape of a square, rectangle or rhombus. It is, however, possible to have cells with irregular and / or rounded shapes.

[0032] The load cells may be configured to be coupled with load cell amplifiers and analog-to- digital converters. A person skilled in the art will understand how to implement such components, which may be part of the load cells themselves are separate components. The load cells 108 in the example of Fig. 2b have a load cell amplifier 121 and an analog-to-digital converter 122. The load cells may be calibrated such that they respond correctly to the weights on the animal surveillance system. This may take into account, for example, variations in how precisely the load cells are mounted and characteristics in how the top sheet transfers forces to the individual load cells.

[0033] Preferably, the animal surveillance system comprises an animal identification system. The animal identification system may be, for example, a radio-frequency identificationbased and / or camera-based identification system. In a camera-based identification system, one or more cameras can be installed strategically within the cage. The cameras can be configured to continuously, or when, for example, a certain load cell is triggered, to capture images. There are a number of ways of identifying animals. For example, computer vision algorithms can be used to extract features from the images or videos. It is possible to mark the animals, for example, by adding color or other visible signs. It is also possible to use a trained machine learning model, such as a convolutional neural network, to identify the animals in the cage.

[0034] Another option is to use an RFID-based animal identification system. In such a system, each animal may be tagged with a small RFID transponder. The RFID tags contain a unique identification code that can be read by an RFID interrogation unit, which may also be referred to as an RFID reader. The animal identification system may thus comprise one or more radio-frequency identification interrogation units. The RFID interrogation units may be installed strategically within, around or under the cage. As the animals move around the cage, the RFID interrogator units detect the presence of RFID tags within a certain range. It would also be possible to use other conceptually similar technologies, such as near field communication (NFC). In the context of the present disclosure, NFC can be considered to be a subset of RFID.

[0035] As stated, the animal identification system comprises one or more radio-frequency identification interrogation units. Preferably, the animal identification system comprises a plurality of radio-frequency identification interrogation units distributed over the physical platform. The radio-frequency identification interrogation units may be arranged at strategic places in the cage. The RFID readers may be placed under the physical platform. Figs. 3A and B show an example of how the load cells and RFID interrogators can be distributed. In the example there are three RFID interrogators 126. There are also at least one load cell per cell, in the example there are four load cells 108 per cell 106. Fig. 3B, which is a side view of the animal identification system, illustrates that both the load cells 108 and the RFID readers 126 can be disposed under the surface on which the animals are located.

[0036] In one embodiment, an activation of the predefined load cell triggers the identification of the animal using an associated predefined radio-frequency identification interrogation unit. This means that starting from a situation in which the system is unaware of where the one or more animals are located and thus not capable of tracking any animals, the system can obtain information that serves as a starting point for further tracking and observation using the load cells. If, for example, an animal which has an RFID tag, is identified by a given RFID reader with a given position in the grid, the position can be used a starting point. From there the load cells can then keep track of the identified animal.

[0037] Placing and using RFID interrogators to track animals and observe their behaviors generally provides a relatively bad resolution. By first using the identification system to get an initial identification and then using the load cells for the actual tracking and observation, can provide more detailed information. Preferably, the animal surveillance system comprises a greater number of load cells than radio-frequency identification interrogation units.

[0038] Whether the animal identification system is an RFID-based system, a camera-based system or another system, the animal identification system provides an approximate position of the at least one animal on the physical platform. The load signals from the load cells may provide a more precise position of the at least one animal on the physical platform. In the example in fig. 3A, an animal is close to one of the RFID interrogation units 126. When the RFID interrogation unit has identified the animal, the tracing using load cells can start from one or more of the load cells 108a / 108b.

[0039] In one embodiment, the processing unit is configured to perform an identification of an animal using the animal identification system when a predefined load cell is activated by the animal. The animal identification system is not necessarily active. If, for example, an RFID reader has a known predefined position, an activation of a nearby loadcell can be used to trigger a reading of the RFID reader. Once the animal has been identified, the system can then track positions and / or change of positions of the identified at least one animal animals starting from the predefined load cell.

[0040] The physical platform may comprise a top sheet covering the grid of cells. The top sheet may be deformed when weight is placed on it. This will allow to register small changes of weight, such that animals, for example, rodents, can be observed. As stated above, the load cells can be implemented in various ways. In addition to the load cells, there may be a number of weight distribution elements, preferably coil springs, disposed under the top sheet. The coil springs, which are preferably distributed over the whole surveillance system, may have several purposes. For example, the top sheet may introduce a, generally, smoother behavior in terms of how the top sheet deforms when weight is applied on it. It may also easy the requirements on how precisely the load cells are mounted. As would be realized by a person skilled in the art, the system needs to be calibrated to take into account that the weight distribution elements may absorb some of the weight on the top sheet. Fig. 10 shows an example of weight distribution elements 125 disposed under the top sheet 111. In the example the weight distribution elements 125 are arranged between the top sheet 111 and the intermediate layer 118 for holding the load cells 108. In the example the animal surveillance system further comprises a number of RFID interrogation units 126.

[0041] The animal surveillance system may further comprise a processing unit configured to obtain individual reference weights of a plurality of animals. This can be done, for example, before placing the animals on the surveillance system. Alternatively it may done as an initial step once the animals are placed on the surveillance system. Based on the load signals from the load cells, the processing unit may track the position of each of the plurality of animals on the physical platform. Continuously obtaining load signals in the context of the present disclosure does not necessarily mean that load signals are obtained all the time. As a person skilled in the art realizes, signals can be sampled at intervals, for example in a period of less than 5 seconds, preferably less than 1 second, more preferably less than 0.5 seconds, while still being considered to be obtained continuously. In order to increase the precision of the surveillance system, multiple signals of a load cell can be measured in a given period of time such as 0.2 seconds and then they can be averaged. This process can provide smaller errors in the measured weight values of the animals. An example of the process of an animal moving across the grid can be seen in Fig. 6, where a mouse 107 moves from a first cell 106a of the grid to a second cell 106b, further on to a third cell 106c. As the weight is changing in the three cells 106a, 106b and 106c, a trajectory 109 can be extracted.

[0042] The term reference weight may refer to the initial weight of each animal that is recorded when entering the platform. A reference load cell can be designed on each surveillance system, where every animal is measured on it before entering the grid. Alternatively, each animal can be weighed individually on a scale and its weight can be recorded, before it may enter the surveillance platform.

[0043] As stated, the presently disclosed animal surveillance system may be adapted to be placed or integrated in an animal cage of Ell Type 3 for mice. More generally, the system may be adapted to track small rodents, such as rodents having a weight of less than 50 gram, including mice. The system may be adapted to track mice with a weight in the range of 10-40 gram. As a person skilled the art would realize, the physical and mechanical properties of the system have to be adapted to the specific circumstances and requirements. The plurality of animals is not limited to any specific number of animals but may be, for example a group of 2-10 animals, or, more specifically, 3-5 animals.

[0044] In one embodiment, the surveillance system can have the top sheet adapted to distribute a part of a weight applied on a first cell of the grid of cells to one or more neighbouring cells. For example, if the sheet is thin enough and / or made of a flexible and / or elastic material, it can allow a distribution of the weight of an animal across the grid. If an animal is located at a position where it causes several load cells to observe a weight on the surveillance system, it may give a more detailed position of the animal compared to only one load cell registering that the animal is on one of the cells of the grid. The more detailed position may serve the purpose of tracking a distance that the animal has moved more accurately. The distribution of the weight to several load cell may also provide information about specific activity of the animal, such as shaking, making circles, or standing on two legs etc. For instance, if an animal is shaking it can trigger vibrations which can be captured as a weight distribution change by the processing unit. Accordingly, by using a machine learning model, it is possible to distinguish and pinpoint the activity of an animal. More details about these activities are provided on the sections below. The processing unit may be configured to extract a weight of the at least one animal, or extract a position or a change of a position of the at least one animal, by combining load signals from the first cell and one or more neighboring cells

[0045] In the example in fig. 3A it can be seen if an animal is located between a number of load cells 108a / 108b, several of these load cells 108a / 108b may provide load signals that can be used to get a more detailed view of where the animal is located or is moving.

[0046] The animal surveillance system can have the processing unit configured to extract the aforementioned weight distribution based on the load signals from the first cell and the one or more neighbouring cells. The measurements of weight distribution can be then used to evaluate various activity and behaviour patterns of animals, as disclosed in more detail in the sections below.

[0047] In one embodiment, the processing unit is further configured to compute a modelled grid of cells having more cells than the grid of cells of the physical platform, based on the extracted weight distribution. This can be achieved by using the extracted weight distribution data. For example, by measuring the load signals of the first load cell where an animal is placed, and the neighbouring load cells, it can be possible by interpolation to calculate additional virtual load signals of intermediate load cells. This process can be repeated for all load cells and a computed grid of cells can be visualized, having higher resolution than the physical grid. The modelled grid of cells provides the functionality of visualizing the positions of animals in the platform with higher precision than using the original grid of cells. Creating the modelled grid of cells may involve a sequence of steps, wherein one or more items, possibly several items with different weights, are placed systematically on different locations on the grid of cells, and the load signals from the load cells are observed. Preferably the one or more items are placed both directly on the load cells and at positions between the load cells where the weight is applied on several load cells. Based on the measurements, the modelled grid of cells having more cells than the grid of cells of the physical platform can be extracted. An example of how the modelled grid of cells can function is illustrated in Figs. 12A and 12B where two snapshots of two rodents 107a and 107b moving in a surveillance platform 100 are shown. As the weight of the rodents is distributed to the platform, four virtual cells 115 and 116 that receive a significant weight are highlighted. The virtual cells 115 and 116 have a higher granularity than the actual 3x3 cells in the grid.

[0048] In one embodiment of the animal surveillance system, the top sheet can be flexible. This feature can enhance the functionality of the system, as it can increase the accuracy of the weight distribution, effectively increasing the resolution of the modelled grid of cells. To achieve higher flexibility, the top sheet can be thinner than 5 mm, preferably thinner than 1 mm. The top sheet may be made of any suitable material, for example a plastic, such as polystyrene or polycarbonate, or a metal, such as aluminium.

[0049] Measuring animal weight and activity

[0050] The animal surveillance system can have the processing unit further configured in order to extract weight changes of specific animals over time, as well as to extract activity changes of specific animals over time. That can be achieved by continuously getting signals from all the load cells of the platform, and thus monitoring the weight at the same time on the whole platform. If an animal is moving along several load cells, that change of weight can be monitored, and the activity of each animal can be extracted.

[0051] Following the above principle, it is also possible to measure the distance covered by an animal over a time internal, and compare it at different points in time, or measure the frequency of movement between cells compared at different points in time. The above can be performed since the size of each load cell is known, therefore as an animal is moving along certain load cells, the time required to reach from point A to point B can be measured accurately. As a result, the distance covered over a time interval can be computed. It may also be possible to collect and store data for an extended period of time such as weeks and configure the processing unit to plot the distance, velocity, weight change over a period of days or weeks for a plurality of animals. The above example can be seen a simplified general illustration of how movement can be calculated and extracted. The load data from the load cells can give a more detailed view of movement.

[0052] The animal surveillance system may also compute the positions of the plurality of animals by associating each animal with an identification number and / or an individual reference weight; obtaining baseline load signals from the load cells; comparing the continuously obtained load signals against the baseline load signals for all cells; and finally identifying a specific animal in a specific cell by finding a difference in load for the specific animal that matches the reference weight of the specific animal. As described, using the capability of continuously monitoring the signals provided by the load cells of the grid, it is possible to identify and monitor each of the animals present on the grid. For example, if two animals are moving towards each other, their initial load cells will identify them by measuring the weight signal. As they approach each other, the in-between load cells will capture the signal and identify their activity. When reaching the same load cell, the total weight can be measured and since it will be an exact summation of their two weights, the system can conclude that these two animals are placed on the load cell. Using the above principle, a plurality of animals can also be identified when moving on the grid. Moreover, the identification of the animals can be more precise when taking into account the weight distribution of each animal. The weight distribution can provide additional information, which can act as a crosscheck to verify the position of each animal.

[0053] When initializing the surveillance of animals in the presently disclosed animal surveillance system, there will typically be a baseline of the load signals from the load cells. The baseline, which may have the purpose of a number of reference levels for the load cells, may be given for a setup in which there are items in the cage. The items may include bedding material, for example wood shavings, or toys, such as wheels or tunnels, for the animals. The processing unit may be configured to take into account and / or make assumptions of a position of an animal positioned, for example, on an item. The system may, for example, use data indicating where items are positioned. If the weight of an item increases with the weight of one or several of the animals, it may be assumed that the animal is positioned on the item. In one embodiment of the presently disclosed surveillance system the processing unit is further configured to compute an amount of material that has been moved from at least one cell to at least one other cell. It is possible to compute whether at least one of the animals is moving material from one cell to another. This can for instance be achieved because an animal will gradually move material, for example wood shavings, from a load cell to another. As the processing unit is capable of continuously monitoring the weight of each load cells, it is possible to compute minor weight changes such as wood shavings being moved by an animal from one load cell to another. These minor weight changes due to material being moved along the unit cells can be stored by the processing unit to recalibrate the baseline weight of each load cell, to prevent inaccurate weight measurements of other animals tracked by the load cells. In addition, it can be feasible to differentiate weight changes between material being moved and animal moving in the platform, as animals would tend to move more dynamically and quickly, while the material will be slowly redistributed in the platform. In one embodiment of the presently disclosed animal surveillance system, the baseline of load signals may be updated. If, for example, an animal in the cage, moves bedding material between cells, it will be possible to update the baseline of load signals. The weight of the animal is still known and can be used to keep track of the animal, observed as differences in load as the animal moves between cells.

[0054] As noted above, the position of several animals positioned in the same cell could also be computed by obtaining the total load of the cell and the weight distribution, and identifying several animals as being positioned in the cell if their combined weight matches the total load of the cell and of the neighbouring cells. This can be further checked, by measuring the surrounding load cells before and after in time, and verifying that the given animals were present on the initial load cell.

[0055] Extracting animal behaviour

[0056] The animal surveillance system can be further configured to extract information regarding the behaviour of the animals. Specifically, the processing unit can characterize the behaviour and / or condition of at least one of the animals based on their activity change and / or weight change over time and also estimate whether several animals are in a state of rest in a group, judging by their movement over time. For example, if an animal is found isolated in a cell and is moving much less than other animals then the processing unit can conclude that its behaviour is not normal, and therefore store information about its condition. On the other hand, an identified animal which is moving a lot among the load cells can be identified as such. The information regarding normal or abnormal behaviour can be used in determining whether certain animals develop signs of neurodegenerative disorders, such as epilepsy or Parkinson’s disease. As discussed in the previous section, it can be possible to extract weight distributions of animals in the grid of cells. Hence, it is possible to differentiate an animal standing on its two or four legs, as the weight distribution will be different, and that can therefore be traced by the load cells and the processing unit. For example, if an animal is standing on its four legs, then the weight will be more evenly distributed among its neighbouring load cells, while if it is standing on two legs, the pressure on the load cell it stands will be larger, and therefore the weight distribution will be focused below the animal. Therefore, the processing unit can conclude whether an animal is standing on two or four legs. Using interpolation between adjacent cells, it is also possible to get a higher spatial resolution than is provided by merely using the distinct position of each cell in the grid.

[0057] The weight distribution can be further used to extract small movements or shakings of animals in the surveillance platform. For instance, if an animal is grooming or fighting with other animals, the small oscillations of the weight distribution can be traced by the processing unit and the specific activity can be distinguished using a model.

[0058] The animal surveillance system may be configured to compute whether an animal is standing on its hind-legs, resting towards a cage wall or hanging in the cage. When an animal rests partially or fully on a different part of the cage apart from the floor (such as cage walls), a difference in the signals from the load cells can be measured. If the top sheet does not experience any bending, the neighbouring load cells shall not experience an increase of weight either, and therefore the processing unit can conclude that the only remaining option is that the animal is hanging in the grid. Alternatively, if there is a decreased signal from the neighbouring cells which is however not zero, it can be concluded that the animal is resting towards a cage wall. In addition, if an animal is standing on two legs it should also be possible to verify that by monitoring its movement, which is typically reduced compared to when it is standing on four legs. By continuously taking inputs from all load cells, that activity can be monitored for a plurality of animals in the cage. A machine learning model can also be implemented, which can be trained with large amounts of data to differentiate with higher accuracy whether an animal is standing on two or four legs.

[0059] The surveillance system may have a processing unit further configured to compare animal activity with and without human and / or animal presence in the room outside of the animal cage. In the case of human or animal activity in the vicinity on the outside of the cage, special infrared sensors or cameras can be coupled to the processing unit to alert it for such an event. Therefore, it is possible to compare the activity of animals in the cage before and after experiencing the outsider presence. Examples of such activities can be frequency of movement, whether they form groups or whether they collect close or far away from the human or animal outside of the cage.

[0060] In addition, it is possible to further configure the processing unit to compute whether one animal leaves a plurality of animals in a state of rest in a group and is instead in a state of rest distanced from the group for a time exceeding a predefined threshold. For example, if the signal from a specific load cell is corresponding to three animals for an extended amount of time without monitoring any movement, it can be concluded that the animals are in a state of rest. At the same time, if a signal from a load cell corresponds to one animal for an extended amount of time, it can be concluded that the animal is isolated while in a state of rest. The extended amount of time can be a parameter threshold which can be configured in the processing unit, and corresponds to the typical rest time of the animal-type studied. Different time thresholds can be included in the processing unit which can for example dictate whether an animal is having less than normal activity or more than normal activity compared to an average animal. Other uses of the thresholds can be applied when an animal is gaining or losing weight, if an animal is running around in circles for an extended amount of time, if an animal is avoiding sleeping huddles, or if an animal is moving along a certain path of the grid back and forth. This process may also be a successfully used in animals which have been administered with a drug and the researchers would want to monitor their activity change over time.

[0061] The animal surveillance system according to the above can have the processing unit further configured to identify quick repetitive movements. This can be computed by continuously taking signals from load cells and comparing the movement of the specific animal to previously stored movement patterns. If the activity of the animal exceeds a given threshold in the processing unit, then an alarm can trigger for the specific animal, which lets the researchers know that it experiences quick repetitive movements. This alarm can be used for any other uses as well, such as weight loss / gain, inactivity / high activity or fighting between animals. The alarm can function as follows: the user receives information about when the activity began, how long it lasted, and if applicable, which animals it concerns. Moreover, the animal surveillance system according to the above can have the processing unit further configured to identify that an animal moves only along cells abutting the cage wall. That can be computed as the processing unit can be configured to comprehend which cells are in the borders of the cage and which are not. Therefore, if an animal is moving only along cells abutting the cage walls, signals from such cells will be sent to the processing unit as the animal is moving, and the processing unit can conclude that the animal is moving along such cells.

[0062] The load signals from the load sensors can also be used to detect specific biomarkers and disease symptoms, e.g. impaired gait resulting from a disease of the motor system or muscular contraction of frequencies typically associated with epileptic fits.

[0063] The load signals from the load sensors can also be used to detect specific behaviours such as aggression, avoidance, exploration, foraging or birth-giving.

[0064] In the case of abnormal behaviour or the occurrence of an anticipated event (e.g. birth), a notification or alarm may be triggered to make the researcher of animal technician aware of the state of an animal in the cage also when they are in another location than the animals.

[0065] Finally, the animal surveillance system can have the processing unit further configured to identify quick circular movement of an animal. As the modelled grid of cells can be stored in rows and columns in a form of a matrix, it is possible to compute when for example four modelled load cells adjacent to a corner are triggered. In that case, if the signal from the load cells is linked to an identical weight, it can be concluded that the same animal is moving through these load cells, and as a result it is performing a circular movement. The time interval between the signals of the modelled load cells can be used to determine whether these circular animal movements are quick or not.

[0066] Platform specifications

[0067] The animal surveillance system may be installed in a physical platform which has a substantially flat top surface, and wherein the load cells are disposed under the substantially flat top surface. The surveillance system can be manufactured in a specific way to include three different layers: the base 119 of the surveillance system, an intermediate layer 118 for holding the load cells 108 and the top sheet 111, as shown in Fig. 8. The top sheet 111 has insertions 120 for accommodating the load cells 108.

[0068] The intermediate layer 118 is also shown in Fig. 9. The intermediate layer 118 has insertions 120 in which load cells 108 are disposed. The view in Fig. 9 is from a bottom side of the intermediate layer 118. In this embodiment each load cell has an elongated shape, wherein a part the observes a weight is arranged in an insertion 120 and a further attachment part extends and is connected to attachment holes 105. The intermediate layer 118 has a number of further general attachment holes 124, which can be used to attach, for example, cables and further components to the intermediate layer 118. Only two of the in total 12 insertions 120 have load cells in the figure - the typical configuration would be that all insertions accommodate load cells 108. When using load cells with an elongated shape, such as the ones in Fig. 9, the distance between the load cells, at least in the longitudinal direction of the load cells, would be relatively long if the load cells were just placed side by side (i.e. short side abutting another short side). If the load cells 108 instead are arranged in an overlapping pattern, preferably also slightly rotated, as in Fig. 9, the load cells 108 can have a similar distance to each other in both the length and width dimensions.

[0069] The physical platform may be dimensioned to fit into an animal cage of Ell Type 3 for mice. The physical platform may, for example, have a length of 300 mm to 1000 mm and a width of 200 mm to 1000 mm. The surveillance system may in principle have any suitable number of cells in any grid configuration MxN cells . A typical configuration may include 3x3 cells, or 4x3 cells, or 4x4 cells, 5x5 cells and so forth. The distance between the load cells may be adapted to the area of use and the animals. For example, the distance between the load cells may be between 30 mm and 200 mm, or between 50 mm and 150 mm.

[0070] Moreover, the surveillance system 100 may have a substantially rectangular shape with rounded corners 103 as shown in, for example, Fig. 1A. The physical platform may be dimensioned to fit into an animal cage of Ell Type 3 for mice. The Type 3 cage can be made by, for example, polycarbonate, polysulphone or polypropylene. In order to fulfil compatibility between the animal surveillance system and the Type 3 cages, tapered sides may designed on the system which can lead to a fine fit between the system and the Type 3 cages. An example of a cross section of the tapered sides 112 of the platform can be seen in Fig. 11. As shown in Figs. 1A and 1 B, the physical platform may have a length 102 of 375 mm to 400 mm, a width 101 of 215 mm to 240 mm, and a height 104 of 20 mm to 60 mm. The grid of cells may be diagonally oriented with respect to the substantially rectangular shape, or rotated in relation to a general orientation of the substantially rectangular shape of the physical platform. It may have a general orientation differentiating at least 10° with respect to the substantially rectangular shape. An example of a grid with diagonally oriented cells 106 is seen in Fig. 4. The animal surveillance system may be designed such that the physical platform has an outer shape comprising a top surface, a bottom surface, and a plurality, such as four, side surfaces, wherein the outer shape is tapered from the top surface to the bottom surface.

[0071] System, software interface, display

[0072] The animal surveillance system can be equipped with a display for showing positions and any other extracted or measured data. An example of such a display 106 can be seen in Fig. 7. Moreover, as shown in Fig. 7, the system can support an input interface 124 for obtaining individual weights and identification numbers of the plurality of animals and a processing unit 123124. The display can either be installed on each cage, from which a researcher can navigate on all the data stored, or it can be a computer software which can be accessed from a central unit that can store and analyse data from various cages. The animal surveillance system may further comprise an animal identification system 127.

[0073] The animal surveillance system may further comprise peripheral components, such as one or more memories, which may be used for storing instructions that can be executed by any of the processors. The animal surveillance system may further comprise internal and external network interfaces, input and / or output ports, a keyboard or mouse etc. As would be understood by a person skilled in the art, a processing unit also may be a single processor in a multi-core / multiprocessor system. Both the computing hardware accelerator and the central processing unit may be connected to a data communication infrastructure.

[0074] The animal surveillance system may include a memory, such as a random access memory (RAM) and / or a read-only memory (ROM), or any suitable type of memory. The animal surveillance system may further comprise a communication interface that allows software and / or data to be transferred between the system and external devices. Software and / or data transferred via the communications interface may be in any suitable form of electric, optical or RF signals. The communications interface may comprise, for example, a cable or a wireless interface.

[0075] Method for animal surveillance

[0076] The present disclosure also relates to a method for surveying a plurality of animals in an animal cage, comprising the steps of: providing a physical platform divided into a grid of cells, wherein each cell comprises at least one load cell; identifying at least one animal on the physical platform using an animal identification system; continuously obtaining load signals from the load cells; and based on the load signals from the load cells, tracking positions and / or change of positions of the identified at least one animal on the physical platform.

[0077] Fig. 5 shows a flow chart of an embodiment of a method 200 of surveying an animal in an animal cage. The method 200 comprises the steps of: providing a physical platform divided into a grid of cells, wherein each cell comprises at least one load cell (201); identifying at least one animal on the physical platform using an animal identification system (202); continuously obtaining load signals from the load cells (203); and based on the load signals from the load cells, tracking positions and / or change of positions of the identified at least one animal on the physical platform (204).

[0078] For example, this method can be applied to any kind of studied animals, where their positions, behaviour, activity and weight is crucial for an experiment. As the load cell can be designed to have different sizes and sensitivities, an animal surveillance platform can be designed accordingly to have larger / smaller size and larger / smaller sensitivity on the load cells, according to the size and weight of the studied animals. Furthermore, the method of surveying a plurality of animals in an animal cage according to the above, can further comprise the step of sending an alarm if an abnormal activity is detected. For instance, if an animal is moving rapidly among load cells, or if an animal is not moving at all for a time longer than the rest time of the animal, the load cells can measure the weight on the corresponding cells and the processing unit can compute the movement or absence of movement of the animal. Therefore, an alarm can be sent to the system if such abnormal activity is detected. The present disclosure further relates to a computer program having instructions which, when executed by a computing device or computing system, cause the computing device or computing system to carry out any embodiment of the presently disclosed method of surveying a plurality of animals in an animal cage. The computer program may be stored on any suitable type of storage media, such as non-transitory storage media.

[0079] In a further embodiment the animal surveillance system can be used in applications outside a cage. The disclosure therefore relates to, according to a further embodiment, an animal surveillance system, the animal surveillance system comprising: a physical platform divided into a grid of cells, wherein each cell comprises at least one load cell; a processing unit configured to: obtain individual reference weights of a plurality of animals; continuously obtain load signals from the load cells; based on the individual reference weights of the plurality of animals and the load signals from the load cells, tracking positions and / or change of positions of each of the plurality of animals on the physical platform.

[0080] In a further embodiment the animal surveillance system comprises: a physical platform divided into a grid of cells, wherein each cell comprises at least one load cell; an animal identification system; a processing unit configured to: identify at least one animal on the physical platform using the animal identification system; based on the load signals from the load cells, extract information about position and movement of the plurality of animals on the physical platform, and extract biomarkers of the plurality of animals on the physical platform.

[0081] Further details

[0082] 1. An animal surveillance system for an animal cage, the animal surveillance system comprising: a physical platform divided into a grid of cells, wherein each cell comprises at least one load cell; a processing unit configured to: obtain individual reference weights of a plurality of animals; continuously obtain load signals from the load cells; based on the individual reference weights of the plurality of animals and the load signals from the load cells, tracking positions and / or change of positions of each of the plurality of animals on the physical platform. An animal surveillance system for an animal cage, the animal surveillance system comprising: a physical platform divided into a grid of cells, wherein each cell comprises at least one load cell; an animal identification system; a processing unit configured to: identify at least one animal on the physical platform using the animal identification system; continuously obtain load signals from the load cells; based on the load signals from the load cells, tracking positions and / or change of positions of the identified at least one animal animals on the physical platform. The animal surveillance system according to item 1 or 2, wherein the physical platform comprises a top sheet covering the grid of cells. The animal surveillance system according to item 3, wherein the top sheet is adapted to distribute a part of a weight applied on a first cell of the grid of cells to one or more neighboring cells. The animal surveillance system according to item 4, wherein the processing unit is configured to extract a weight distribution based on the load signals from the first cell and the one or more neighboring cells. The animal surveillance system according to item 5, wherein the processing unit is configured to compute a modelled grid of cells having more cells than grid of cells of the physical platform based on the extracted weight distribution. 7. The animal surveillance system according to any one of items 3-6, wherein the top sheet is a flexible sheet.

[0083] 8. The animal surveillance system according to any one of items 3-7, wherein the top sheet is thinner than 5 mm, preferably thinner than 1 mm, placed onto the grid of cells.

[0084] 9. The animal surveillance system according to any one of items 3-8, further comprising weight distribution elements, preferably coil springs, disposed under the top sheet.

[0085] 10. The animal surveillance system according to any one of the preceding items, wherein each of the at least one load cell comprises a load cell amplifier and an analog-digital converter.

[0086] 11. The animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to extract weight changes of specific animals over time.

[0087] 12. The animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to extract activity changes of specific animals over time.

[0088] 13. The animal surveillance system according to item 12, wherein the activity changes comprise a measure of distance moved over a time interval, compared at different points in time, or a frequency of movement between cells, compared at different points in time.

[0089] 14. The animal surveillance system according to any one of the preceding items, wherein the positions of the plurality of animals are computed by: associating each animal with an identification number and an individual reference weight; obtaining baseline load signals from the load cells; for all cells, comparing the continuously obtained load signals against the baseline load signals; identifying a specific animal in a specific cell by finding a difference in load for the specific animal that matches the reference weight of the specific animal.

[0090] 15. The animal surveillance system according to item 14, wherein the position of several animals being positioned in the same cell are computed by obtaining a total load of the cell; and identifying several animals as being positioned in the cell if their combined weight matches the total load of the cell.

[0091] 16. The animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to characterize a behavior and / or condition of at least one of the animals based on activity change and / or weight change over time.

[0092] 17. The animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to compute whether several animals are in a state of rest in a group.

[0093] 18. The animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to compute whether at least one of the animals is gradually moving material from one cell to another.

[0094] 19. animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to compute an amount of material that has been moved from at least one cell to at least one other cell.

[0095] 20. The animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to compute whether an animal is standing on its hindlegs, resting towards a cage wall or hanging in the cage.

[0096] 21. The animal surveillance system according to items 5 and 20, wherein computing whether an animal is standing on its hindlegs is based on the weight distribution.

[0097] 22. The animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to compare animal activity with and without other human and / or animal presence in the room outside the animal cage. The animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to compute whether one animal leaves a plurality of animals in a state of rest in a group and is instead in a state of rest distanced from the group for a time exceeding a predefined threshold. The animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to compute whether one animal has significantly less activity than the rest of the animals. The animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to identify quick repetitive movements. The animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to identify that an animal only moves along cells abutting the cage wall. The animal surveillance system according to any one of the preceding items, wherein the processing unit is further configured to identify quick circular movements of an animal. The animal surveillance system according to any one of the preceding items, wherein the physical platform has a substantially rectangular shape with rounded corners. The animal surveillance system according to any one of the preceding items, wherein the physical platform is dimensioned to fit into an animal cage of Ell Type 3 for mice. The animal surveillance system according to any one of the preceding items, wherein the physical platform has a length of 375 mm to 400 mm, and a width of 215 mm to 240 mm. 31. The animal surveillance system according to any one of the preceding items, wherein the grid of cells are diagonally oriented with respect to the substantially rectangular shape, and / or has a general orientation differentiating at least 10° with respect to the substantially rectangular shape.

[0098] 32. The animal surveillance system according to any one of the preceding items, wherein the physical platform has an outer shape comprising a top surface, a bottom surface, and a plurality, such as four, side surfaces, wherein the outer shape is tapered from the top surface to the bottom surface.

[0099] 33. The animal surveillance system according to any one of the preceding items, wherein the load cells have a sensitivity of 0.1 g or better.

[0100] 34. The animal surveillance system according to any one of the preceding items, further comprising a display for showing positions and / or any other extracted or measured data.

[0101] 35. The animal surveillance system according to any one of the preceding items, further comprising an input interface for obtaining individual weights and identification numbers of the plurality of animals.

[0102] 36. A method of surveying a plurality of animals in an animal cage, comprising the steps of: providing a physical platform divided into a grid of cells, wherein each cell comprises at least one load cell; obtaining baseline load signals from the load cells obtaining individual reference weights of the plurality of animals; continuously obtaining load signals from the load cells; based on the individual reference weights of the plurality of animals and the load signals from the load cells, tracking positions and / or change of positions of each of the plurality of animals on the physical platform.

[0103] 37. The method of surveying a plurality of animals in an animal cage according to item 36, further comprising the step of sending an alarm if an abnormal activity is detected.

Claims

Claims1. An animal surveillance system for an animal cage, the animal surveillance system comprising: a physical platform divided into a grid of cells, wherein each cell comprises at least one load cell; an animal identification system; a processing unit configured to: identify at least one animal on the physical platform using the animal identification system; continuously obtain load signals from the load cells; based on the load signals from the load cells, tracking positions and / or change of positions of the identified at least one animal animals on the physical platform.

2. The animal surveillance system according to claim 1, wherein the animal identification system comprises one or more radio-frequency identification interrogation units.

3. The animal surveillance system according to claim 2, the animal identification system comprising a plurality of radio-frequency identification interrogation units distributed over the physical platform.

4. The animal surveillance system according to any one of the preceding claims, wherein an activation of the predefined load cell triggers the identification of the animal using an associated predefined radio-frequency identification interrogation unit.

5. The animal surveillance system according to any one of the preceding claims, comprising a greater number of load cells than radio-frequency identification interrogation units.

6. The animal surveillance system according to claim 1 , wherein the animal identification system is a camera-based animal identification system.

7. The animal surveillance system according to any one of the preceding claims, wherein the identification of the at least one animal on the physical platform using the animal identification system provides an approximate position of the at least one animal on the physical platform.

8. The animal surveillance system according to claim 7, wherein the load signals from the load cells provide a more precise position of the at least one animal on the physical platform.

9. The animal surveillance system according to any one of the preceding claims, wherein the processing unit is configured to perform an identification of an animal using the animal identification system when a predefined load cell is activated by the animal.

10. The animal surveillance system according to claim 9, wherein the processing unit is configured to track positions and / or change of positions of the identified at least one animal animals starting from the predefined load cell.

11. The animal surveillance system according to any one of the preceding claims, wherein the physical platform comprises a top sheet covering the grid of cells, wherein the top sheet is adapted to distribute a part of a weight applied on a first cell of the grid of cells to one or more neighboring cells.

12. The animal surveillance system according to claim 11, wherein the processing unit is configured to extract a weight of the at least one animal, or extract a position or a change of a position of the at least one animal, by combining load signals from the first cell and one or more neighboring cells.

13. The animal surveillance system according to any one of claims 11-12, wherein the processing unit is configured to extract a weight distribution based on the load signals from the first cell and the one or more neighboring cells.

14. The animal surveillance system according to any one of claims 1311-13, wherein the processing unit is configured to compute a modelled grid of cells having more cells than grid of cells of the physical platform based on the extracted weight distribution.

15. The animal surveillance system according to any one of claims 11-14, further comprising weight distribution elements, preferably coil springs, disposed under the top sheet.

16. The animal surveillance system according to any one of the preceding claims, wherein the processing unit is further configured to extract weight changes of specific animals over time.

17. The animal surveillance system according to any one of the preceding claims, wherein the position of several animals being positioned in the same cell are computed by obtaining a total load of the cell.

18. The animal surveillance system according to any one of the preceding claims, wherein the processing unit is further configured to characterize a behavior and / or condition of at least one of the animals based on activity change and / or weight change over time.

19. A method of surveying an animal in an animal cage, comprising the steps of: providing a physical platform divided into a grid of cells, wherein each cell comprises at least one load cell; identifying at least one animal on the physical platform using an animal identification system; continuously obtaining load signals from the load cells; based on the load signals from the load cells, tracking positions and / or change of positions of the identified at least one animal on the physical platform.

20. The method of surveying an animal in an animal cage, using an animal surveillance system according to any one of claims 1-18.

Images