WO2019203661A1 - C-fish – fish welfare control - Google Patents
C-fish – fish welfare control Download PDFInfo
- Publication number
- WO2019203661A1 WO2019203661A1 PCT/NO2019/050086 NO2019050086W WO2019203661A1 WO 2019203661 A1 WO2019203661 A1 WO 2019203661A1 NO 2019050086 W NO2019050086 W NO 2019050086W WO 2019203661 A1 WO2019203661 A1 WO 2019203661A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- data
- logger
- decision maker
- interpreter
- fish
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01K—ANIMAL HUSBANDRY; CARE OF BIRDS, FISHES, INSECTS; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
- A01K63/00—Receptacles for live fish, e.g. aquaria; Terraria
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01K—ANIMAL HUSBANDRY; CARE OF BIRDS, FISHES, INSECTS; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
- A01K61/00—Culture of aquatic animals
- A01K61/10—Culture of aquatic animals of fish
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01K—ANIMAL HUSBANDRY; CARE OF BIRDS, FISHES, INSECTS; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
- A01K63/00—Receptacles for live fish, e.g. aquaria; Terraria
- A01K63/02—Receptacles specially adapted for transporting live fish
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01K—ANIMAL HUSBANDRY; CARE OF BIRDS, FISHES, INSECTS; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
- A01K63/00—Receptacles for live fish, e.g. aquaria; Terraria
- A01K63/04—Arrangements for treating water specially adapted to receptacles for live fish
- A01K63/042—Introducing gases into the water, e.g. aerators, air pumps
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/042—Knowledge-based neural networks; Logical representations of neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
Definitions
- the invention relates to management system for continuous surveillance in general and more specifically a system and a method for continuous surveillance and control of the welfare and health of fish in a production or transport system.
- the invention is relevant in the field of aquaculture and will be of use for all personnel caring for the fish, at all stages of fish aquaculture.
- CN 106292802 relating to an intelligent prediction and control system and method for the fish-vegetable coexisting system.
- the acquisition module in the system acquires the environment data of the fish- vegetable coexisting system;
- the prediction control module predicts the content of dissolved oxygen in a fish pound of a period in the future and generates a control command based on the prediction result;
- the circulation module controls the fish- vegetable coexisting system in real time and increases oxygen according to the control command.
- a main objective of the present invention is to provide a system and means for:
- the present invention attains the above-described objective by an Al decision maker for receiving interpreted data from the interpreter, wherein the Al decision maker comprises means for establishing a value for animal welfare, means for optimising said value for animal welfare, and means for outputting actions to a control system for initiating manipulation of said environment.
- the system will comprise a number of devices that can read data automatically (video cameras, sensors for flow, oxygen, temperature etc.) and a user interface. Data will be automatically interpreted and stored for future use. The data will be processed by learning methods, partially programmed manually, partially self-learning with neural network methods. Data will be gathered and compared over different production sites. Using these methods, the system will over time be able to identify room for improvement in the production setup. The results will then be relayed to the user for manual regulation, or improved/regulated automatically.
- a management a system for animal welfare comprising a receiver for receiving readings from a plurality of sensors sensing physical states in an environment relating to animal welfare, an interpreter for receiving sensor data from the receiver, wherein the system further comprises an Al decision maker for receiving interpreted data from the interpreter, wherein the Al decision maker comprises means for establishing an estimate for animal welfare, means for optimising said estimate for animal welfare, and means for outputting actions to a control system for initiating manipulation of said environment.
- the management system further comprises a logger for logging receiver data sent to logger, wherein the interpreter is arranged to read logger data from the logger so that the Al decision maker can read logged received data from the plurality of sensors.
- the management system further comprises at least one from the group comprising rule set database comprising rule sets and learning algorithm database comprising learning algorithms, operatively connected to the Al decision maker.
- a method for operating a management system comprising the steps receiving readings from a plurality of sensors sensing physical parameters in an environment relating to animal welfare using a receiver, receiving sensor data using an interpreter from the receiver, further receiving interpreted data from the interpreter using an Al decision maker for receiving interpreted data from the interpreter, wherein the Al decision maker further establishes a value for animal welfare, optimises said value for animal welfare, and outputs actions to a control system for initiating manipulation of said environment.
- the decision maker enters decisions into a logger.
- the Al decision maker further updates an internal neural network with subsequent interpreted data from the interpreter.
- the Al decision maker further updates an internal neural network with reruns past actions logged in the logger for training.
- the Al reruns past actions logged in the logger for training using at least one from the group comprising different rule sets from a rule set database and different learning algorithm learning algorithm.
- Fig. 1 shows a flow chart that shows how information flows between components within the system.
- Fig. 1 shows a flow chart that shows how information flows between components of the system. Principles forming the basis of the invention
- the underlying principle is that the system collects sensor data and builds up historical data. An Al learns a pattern from the data and will from this training be able to provide an objective current status and/or trends pointing to future events. These can in turn be provided to operators for manual action or used to control the system manually, such as adjusting chemical balance or operating the feeding system.
- past events can be replayed using logged data in combination with different rule sets 1600 and/or learning algorithms 1700 and see what combination will trigger a response. This makes it possible to apply hindsight in analysis by knowing in advance that data points to for instance parasites and that enclosing the fish pens is the correct response.
- Fig. 1 The embodiment of the method according to the invention shown in Fig. 1 comprises sub processes that communicate with each other to form the overall process 100.
- At least one sensor 1100, 1110 feeds readings into a receiver 1200.
- the sensors provide information regarding water quality parameters at strategic points of the water volume, temperature, water flow, biomass density, fish behaviour, fish scales in water and filters, fish appetite, bulk MO2, visible wounds and deformities, degree of emaciation, smoltification status, degree of sexual maturation, visible lice infestation, fish skin colour, water turbidity, vibrations and sounds in the water. Not all parameters have to be made available but the model will improve with more information.
- sensors will have to be local to each fish pen but some such as water temperature can be measured in one or a few locations and extrapolated to the rest of the fish pens in the facility.
- the receiver 1200 receives data from sensors. It may optionally poll or otherwise control the sensors in terms of sampling time and/or rate.
- the receiver outputs data 1202 to an interpreter and outputs data 1204 to a logger.
- Data is typically time stamped and comprises a snapshot of sensor data from a point in time as defined by the time stamp.
- the logger 1300 receives sensor data via the receiver. It stores data and provides the means for the long term use of the collected information. Continuous logging gives means of learning from generation to generation, to improve overall production. The information will be used to improve the system itself, to make the system a tool for experience-based reasoning. Implementation of computer learning methods may be applied.
- the logger emits data 1302 to the interpreter.
- the logger can also emit data 1304 to the User Interface, or Human Interface
- the HI receives logged data 1304 and processes it.
- the main purpose of the HI is to display information to users for information on status and/or trends and also proposed courses of action.
- the Al proposes drastic actions that are best reviewed by humans in order to avoid costly errors. Early in the training process one should expect some errors but over time the Al gains more experiences thus reducing errors.
- the HI processes data to extract relevant data for presentation such as tables and graphs. Typically it may prioritise information to bring emergencies to the front of a priority queue to ensure prompt action.
- Input data to HI comes mainly from the log 1300.
- the interpreter 1400 receives sensor data for interpretation.
- Data can be sensor data received in real time or in near-real time, for actual measurement and analysis of fish health and welfare.
- the interpreter can select stored data 1302 for training purposes.
- the interpreter outputs interpreted data 1402 to an Al for decision making.
- the Al decision maker 1500 receives sensor data via the interpreter as interpreted data 1402. Additionally the Al decision maker receives rule sets 1602 from a rule set database 1600 as well as a learning algorithm 1702 from a database 1700 of at least one learning algorithm.
- Typical outputs of the Al decision maker are decisions 1502. These are typically actions to manipulate steerable parameters of the system to bring the system towards a more optimal situation. While fish health cannot be manipulated directly, it can be improved by adjustments of welfare affecting parameters such as oxygenation oxygen saturation, CC>2-level, water temperature, pH and feeding rate/volume. In many cases the decisions can be safely automated.
- Some actions will require human action, for instance when the decision is severe and require human approval such as light control, and flow speed control, or that manipulations required are not automated so that human intervention is required such as delousing treatments or cleaning of sea cages and removal of bio fouling.
- the Al decision maker emits a report 1504.
- Automated action 1800 is essentially a list of actions to be undertaken that is sent to the control system. This step can comprise interfacing to specific existing control systems.
- Human action 1900 represents reviewing reports issued and approving or rejecting the course of action recommended by the Al decision maker. In some cases this also represents undertaking manipulation of controls not interfaced to the system, or ordering parts or services such as delousing.
- the report can comprise specifics regarding type of treatment or amount of fouling to be removed.
- the control system 2000 represents the means for manipulating parameters according to instructions received automatically by the system or by human action.
- One example is a valve controller controlling oxygen flow into a fish pen.
- Initiate manipulation 2100 represents the actions undertaken by the control system 2000 such as opening or closing a valve controlling oxygen flow into a fish pen.
- the system can be retrained by feeding logged data into the interpreter so that the Al is retraining based on earlier data. This can be performed at high speed.
- the Al can receive pre-interpreted data directly or indirectly from the logger. In a direct way the Al is directly connected to the logger. In an indirect way the interpreter simply passes on stored pre-interpreted data to the Al.
- the logger can also be used to log the outputs 1502 and 1504 from the Al and thus make it simpler to compare different runs of the Al using different rule sets 1602 and learning algorithms 1702, visualising these and optionally the differences on the HI 2200.
- the invention according to the application finds use in fish farming, fish transport and al fish handling for to maintaining fish health and welfare.
Landscapes
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Environmental Sciences (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Biodiversity & Conservation Biology (AREA)
- Animal Husbandry (AREA)
- Marine Sciences & Fisheries (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Zoology (AREA)
- Farming Of Fish And Shellfish (AREA)
- Meat, Egg Or Seafood Products (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A system and a method for continuous surveillance and control of the welfare and health of fish in a production or transport system are provided.The objects of the invention are achieved using an AI trained using logged data that uses the training to provide a more objective measure for status and trend for welfare and health of the fish.
Description
TITLE: C-fish - Fish Welfare Control
Background of the Invention
Field of the Invention
The invention relates to management system for continuous surveillance in general and more specifically a system and a method for continuous surveillance and control of the welfare and health of fish in a production or transport system. The invention is relevant in the field of aquaculture and will be of use for all personnel caring for the fish, at all stages of fish aquaculture.
Background Art
State of the art is reflected in several disparate arts:
Manual regulated systems:
Current art consists of manual decision-making based on visual interpretation (sensor data, eyesight/cameras) and the personal experience of the decision maker.
Automatic regulated systems:
Current art consists of automatic regulation of one parameter, based on one or a few sensors (i.e. regulate the injection of oxygen based on oxygen sensor data). This is done without any consideration of historical data.
Data storage systems:
Current systems store sensor data from different sensors, but are not currently used systematically for learning and improvement purposes.
From prior art one should refer to current practice where one or a small set of parameters are monitored to arrive at a status for fish health and welfare.
From prior art one should also refer to CN 106292802 relating to an intelligent prediction and control system and method for the fish-vegetable coexisting system. The acquisition module in the system acquires the environment data of the fish- vegetable coexisting system; the prediction control module predicts the content of
dissolved oxygen in a fish pound of a period in the future and generates a control command based on the prediction result; the circulation module controls the fish- vegetable coexisting system in real time and increases oxygen according to the control command. There is, however, no functionality disclosed relating to updating using reruns of past events.
There is therefore a need for a method and a system to overcome the above mentioned problems.
Summary of the Invention
Problems to be Solved by the Invention
There is a strong correlation between fish welfare and production efficiency. Fish of high welfare have better health and consequently lower mortality. They show higher growth rate, lower susceptibility to disease, and will contribute to an overall higher meat production. Despite this, there are large variations in the extension to which fish welfare is monitored in the field of aquaculture. In some cases the welfare control depends greatly on the fish farmer and his/her personal perception of the fish welfare. Additionally, production facilities depend on an increasing number of inputs, making it increasingly difficult for a human operator to evaluate the full picture, and thus making a human decision-maker a limiting factor for increased welfare. In taking the human perception out of the equation, the invention gives means for objectively ensuring good fish welfare and a more efficient production.
Therefore, a main objective of the present invention is to provide a system and means for:
providing a comprehensive picture of the fish welfare in the containment system, at any given time,
logging collected data for learning and improvement purposes, and
changing and improving conditions in the containment system when required, either by automated changes or by suggesting actions to be taken by the user.
Means for Solving the Problems
The objective is achieved according to the invention by an management system for animal welfare as defined in the preamble of claim 1 , having the features of the characterising portion of claim 1.
A number of non-exhaustive embodiments, variants or alternatives of the invention are defined by the dependent claims.
The present invention attains the above-described objective by an Al decision maker for receiving interpreted data from the interpreter, wherein the Al decision maker comprises means for establishing a value for animal welfare, means for optimising said value for animal welfare, and means for outputting actions to a control system for initiating manipulation of said environment.
On the production site, the system will comprise a number of devices that can read data automatically (video cameras, sensors for flow, oxygen, temperature etc.) and a user interface. Data will be automatically interpreted and stored for future use. The data will be processed by learning methods, partially programmed manually, partially self-learning with neural network methods. Data will be gathered and compared over different production sites. Using these methods, the system will over time be able to identify room for improvement in the production setup. The results will then be relayed to the user for manual regulation, or improved/regulated automatically.
In a first aspect of the invention a management a system for animal welfare is provided comprising a receiver for receiving readings from a plurality of sensors sensing physical states in an environment relating to animal welfare, an interpreter for receiving sensor data from the receiver, wherein the system further comprises an Al decision maker for receiving interpreted data from the interpreter, wherein the Al decision maker comprises means for establishing an estimate for animal welfare, means for optimising said estimate for animal welfare, and means for outputting actions to a control system for initiating manipulation of said environment.
Preferably the management system further comprises a logger for logging receiver data sent to logger, wherein the interpreter is arranged to read logger data from the logger so that the Al decision maker can read logged received data from the plurality of sensors.
Preferably the management system further comprises at least one from the group comprising rule set database comprising rule sets and learning algorithm database comprising learning algorithms, operatively connected to the Al decision maker.
In a second aspect of the invention a method for operating a management system comprising the steps receiving readings from a plurality of sensors sensing physical parameters in an environment relating to animal welfare using a receiver, receiving sensor data using an interpreter from the receiver, further receiving interpreted data from the interpreter using an Al decision maker for receiving interpreted data from the interpreter, wherein the Al decision maker further establishes a value for animal welfare, optimises said value for animal welfare, and outputs actions to a control system for initiating manipulation of said environment.
Preferably the decision maker enters decisions into a logger.
Preferably the Al decision maker further updates an internal neural network with subsequent interpreted data from the interpreter.
Preferably the Al decision maker further updates an internal neural network with reruns past actions logged in the logger for training.
Preferably the Al reruns past actions logged in the logger for training using at least one from the group comprising different rule sets from a rule set database and different learning algorithm learning algorithm.
Effects of the Invention
The technical differences over prior art is that the Al system can learn and also relearn from past events what actions to take on receiving sensor input. This provides a better basis for decisions.
These effects provide in turn several further advantageous effects: it makes it possible to automate more of the processes involved in aquaculture, and limit human bias. Continuous surveillance and semi-automation can facilitate remote control of a production system, decreasing the demand for personnel at remote or dangerous locations, such as ocean farms,
it makes it possible to predict future event by recognising patterns preceding said events, and
it thus makes it possible to prepare for future events before the problems are detected by traditional means and thus improve animal welfare.
As an example the Al will over time learn the events that can lead to algae problems and warn the users of this even before an official warning is issued, thus providing more time to prepare and reducing losses.
Brief Description of the Drawings
The above and further features of the invention are set forth with particularity in the appended claims and together with advantages thereof will become clearer from consideration of the following detailed description of an [exemplary] embodiment of the invention given with reference to the accompanying drawings.
The invention will be further described below in connection with exemplary embodiments which are schematically shown in the drawings, wherein:
Fig. 1 shows a flow chart that shows how information flows between components within the system.
Description of the Reference Signs
The following reference numbers and signs refer to the drawings:
Detailed Description of the Invention
Various aspects of the disclosure are described more fully hereinafter with reference to the accompanying drawings. This disclosure may, however, be embodied in many different forms and should not be construed as limited to any specific structure or function presented throughout this disclosure. Rather, these aspects are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art. Based on the teachings herein one skilled in the art should appreciate that the scope of the disclosure is intended to cover any aspect of the disclosure disclosed herein, whether implemented independently of or combined with any other aspect of the disclosure. For example, an apparatus may be implemented or a method may be practiced using any number of the aspects set forth herein. In addition, the scope of the disclosure is intended to cover such an apparatus or method which is practiced using other structure, functionality, or structure and functionality in addition to or other than the various aspects of the disclosure set forth herein. It should be understood that any aspect of the disclosure disclosed herein may be embodied by one or more elements of a claim. The invention will be further described in connection with exemplary embodiments which are schematically shown in the drawings, wherein:
Fig. 1 shows a flow chart that shows how information flows between components of the system. Principles forming the basis of the invention
The underlying principle is that the system collects sensor data and builds up historical data. An Al learns a pattern from the data and will from this training be able to provide an objective current status and/or trends pointing to future events. These
can in turn be provided to operators for manual action or used to control the system manually, such as adjusting chemical balance or operating the feeding system.
Also past events can be replayed using logged data in combination with different rule sets 1600 and/or learning algorithms 1700 and see what combination will trigger a response. This makes it possible to apply hindsight in analysis by knowing in advance that data points to for instance parasites and that enclosing the fish pens is the correct response.
Best Modes of Carrying Out the Invention
The embodiment of the method according to the invention shown in Fig. 1 comprises sub processes that communicate with each other to form the overall process 100.
At least one sensor 1100, 1110 feeds readings into a receiver 1200. The sensors provide information regarding water quality parameters at strategic points of the water volume, temperature, water flow, biomass density, fish behaviour, fish scales in water and filters, fish appetite, bulk MO2, visible wounds and deformities, degree of emaciation, smoltification status, degree of sexual maturation, visible lice infestation, fish skin colour, water turbidity, vibrations and sounds in the water. Not all parameters have to be made available but the model will improve with more information.
In many cases sensors will have to be local to each fish pen but some such as water temperature can be measured in one or a few locations and extrapolated to the rest of the fish pens in the facility.
The receiver 1200 receives data from sensors. It may optionally poll or otherwise control the sensors in terms of sampling time and/or rate.
The receiver outputs data 1202 to an interpreter and outputs data 1204 to a logger. Data is typically time stamped and comprises a snapshot of sensor data from a point in time as defined by the time stamp.
The logger 1300 receives sensor data via the receiver. It stores data and provides the means for the long term use of the collected information. Continuous logging gives means of learning from generation to generation, to improve overall production. The information will be used to improve the system itself, to make the
system a tool for experience-based reasoning. Implementation of computer learning methods may be applied.
By using logged data one can review how the system will react based on parameter adjustments and training with large datasets in order to optimise and quickly arrive at a well trained system without being limited to real time constraints. In these cases the logger emits data 1302 to the interpreter.
The logger can also emit data 1304 to the User Interface, or Human Interface
- HI.
The HI receives logged data 1304 and processes it. The main purpose of the HI is to display information to users for information on status and/or trends and also proposed courses of action. In some cases the Al proposes drastic actions that are best reviewed by humans in order to avoid costly errors. Early in the training process one should expect some errors but over time the Al gains more experiences thus reducing errors.
The HI processes data to extract relevant data for presentation such as tables and graphs. Typically it may prioritise information to bring emergencies to the front of a priority queue to ensure prompt action.
Input data to HI comes mainly from the log 1300.
The interpreter 1400 receives sensor data for interpretation. Data can be sensor data received in real time or in near-real time, for actual measurement and analysis of fish health and welfare. Alternatively the interpreter can select stored data 1302 for training purposes.
The interpreter outputs interpreted data 1402 to an Al for decision making.
The Al decision maker 1500 receives sensor data via the interpreter as interpreted data 1402. Additionally the Al decision maker receives rule sets 1602 from a rule set database 1600 as well as a learning algorithm 1702 from a database 1700 of at least one learning algorithm.
Typical outputs of the Al decision maker are decisions 1502. These are typically actions to manipulate steerable parameters of the system to bring the system towards a more optimal situation. While fish health cannot be manipulated directly, it can be improved by adjustments of welfare affecting parameters such as
oxygenation oxygen saturation, CC>2-level, water temperature, pH and feeding rate/volume. In many cases the decisions can be safely automated.
Some actions will require human action, for instance when the decision is severe and require human approval such as light control, and flow speed control, or that manipulations required are not automated so that human intervention is required such as delousing treatments or cleaning of sea cages and removal of bio fouling. In these cases the Al decision maker emits a report 1504.
Automated action 1800 is essentially a list of actions to be undertaken that is sent to the control system. This step can comprise interfacing to specific existing control systems.
Human action 1900 represents reviewing reports issued and approving or rejecting the course of action recommended by the Al decision maker. In some cases this also represents undertaking manipulation of controls not interfaced to the system, or ordering parts or services such as delousing. Optionally the report can comprise specifics regarding type of treatment or amount of fouling to be removed.
In many cases human action will involve manipulation of the control system.
The control system 2000 represents the means for manipulating parameters according to instructions received automatically by the system or by human action. One example is a valve controller controlling oxygen flow into a fish pen.
Initiate manipulation 2100 represents the actions undertaken by the control system 2000 such as opening or closing a valve controlling oxygen flow into a fish pen.
The effects to the manipulation will normally result in new readings from the sensors, thus closing what effectively is a control loop.
Alternative Embodiments
In the above the system can be retrained by feeding logged data into the interpreter so that the Al is retraining based on earlier data. This can be performed at high speed. In an alternative embodiment the Al can receive pre-interpreted data directly or indirectly from the logger. In a direct way the Al is directly connected to the logger. In an indirect way the interpreter simply passes on stored pre-interpreted data to the Al.
The logger can also be used to log the outputs 1502 and 1504 from the Al and thus make it simpler to compare different runs of the Al using different rule sets 1602 and learning algorithms 1702, visualising these and optionally the differences on the HI 2200.
A number of variations on the above can be envisaged. For instance while all examples and embodiments herein refer to fish farming one can imagine a similar system employed to control health for animals such as cows. Furthermore, this system may be adapted to monitor the welfare of live wild caught fish during transport to slaughterhouse, as low stress levels at time of slaughter has been shown to greatly improve filet quality.
Industrial Applicability
The invention according to the application finds use in fish farming, fish transport and al fish handling for to maintaining fish health and welfare.
Claims
1. A management system (100) for animal welfare comprising:
a receiver (1200) for receiving readings from a plurality of sensors (1100) sensing physical states in an environment relating to animal welfare,
an interpreter (1400) for receiving sensor data from the receiver (1200), characterised in that the system further comprises an Al decision maker (1500) for receiving interpreted data from the interpreter (1400),
wherein the Al decision maker comprises means for establishing an estimate for animal welfare,
means for optimising said estimate for animal welfare,
means for outputting actions to a control system (2000) for initiating manipulation (2100) of said environment, and
wherein the Al decision maker further comprises means for updating an internal neural network with reruns past actions logged in the logger for training.
2. The management system (100) according to claim 1 , further comprising a logger (1300) for logging receiver data (1204) sent to logger, wherein the interpreter (1400) is arranged to read logger data (1302) from the logger so that the Al decision maker (1500) can read logged received data from the plurality of sensors (1100).
3. The management system (100) according to claim 1 or 2, further comprising at least one from the group comprising rule set database (1600) comprising rule sets (1602), and learning algorithm database (1700) comprising learning algorithms, operatively connected to the Al decision maker (1500).
4. The management system (100) according to claim 1 or 2, wherein the interpreter (1400) is operatively connected (1302) to the logger (1300) for receiving logged data.
5. A method for operating a management system according to claim 1 , comprising the steps:
receiving readings from a plurality of sensors (1100) sensing physical parameters in an environment relating to animal welfare using a receiver (1200), receiving sensor data using an interpreter (1400) from the receiver (1200), receiving interpreted data from the interpreter (1400) using an Al decision maker (1500) for receiving interpreted data from the interpreter (1400),
wherein the Al decision maker further
establishes a value for animal welfare,
optimises said value for animal welfare, and
outputs actions to a control system (2000) for initiating manipulation (2100) of said environment,
characterised in that the Al decision maker further updates an internal neural network with reruns past actions logged in the logger for training.
6. The method according to claim 5 wherein the decision maker enters decisions into a logger (1300).
7. The method according to claim 5 or 6 wherein the Al decision maker further updates an internal neural network with subsequent interpreted data from the interpreter (1400).
8. The method according to claim 5, wherein the Al reruns past actions logged in the logger for training using at least one from the group comprising different rule sets (1602) from a rule set database (1600) and different learning algorithm (1702) learning algorithm (1700).
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
NO20180513A NO345829B1 (en) | 2018-04-16 | 2018-04-16 | C-fish – fish welfare control |
NO20180513 | 2018-04-16 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2019203661A1 true WO2019203661A1 (en) | 2019-10-24 |
Family
ID=68239740
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/NO2019/050086 WO2019203661A1 (en) | 2018-04-16 | 2019-04-15 | C-fish – fish welfare control |
Country Status (2)
Country | Link |
---|---|
NO (1) | NO345829B1 (en) |
WO (1) | WO2019203661A1 (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113490076A (en) * | 2021-07-07 | 2021-10-08 | 重庆市农业科学院 | Portable fish-vegetable symbiotic water quality rapid detection device |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10599922B2 (en) | 2018-01-25 | 2020-03-24 | X Development Llc | Fish biomass, shape, and size determination |
US11594058B2 (en) | 2019-11-12 | 2023-02-28 | X Development Llc | Entity identification using machine learning |
US11475689B2 (en) | 2020-01-06 | 2022-10-18 | X Development Llc | Fish biomass, shape, size, or health determination |
US11089227B1 (en) | 2020-02-07 | 2021-08-10 | X Development Llc | Camera winch control for dynamic monitoring |
US11659820B2 (en) | 2020-03-20 | 2023-05-30 | X Development Llc | Sea lice mitigation based on historical observations |
US11688154B2 (en) | 2020-05-28 | 2023-06-27 | X Development Llc | Analysis and sorting in aquaculture |
US11615638B2 (en) | 2020-11-10 | 2023-03-28 | X Development Llc | Image processing-based weight estimation for aquaculture |
US11297247B1 (en) | 2021-05-03 | 2022-04-05 | X Development Llc | Automated camera positioning for feeding behavior monitoring |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070257806A1 (en) * | 2004-09-27 | 2007-11-08 | Source Sentinel, Llc | System for monitoring quality of water system |
US20130206078A1 (en) * | 2010-05-18 | 2013-08-15 | Havforskiningsinstituttet | System and method for controlling feeding of farmed fish |
WO2015105523A1 (en) * | 2014-01-12 | 2015-07-16 | Kevin Friesth | Automated hybrid aquaponics and bioreactor system including product processing and storage facilities with integrated robotics, control system, and renewable energy system cross-reference to related applications |
GB2530300A (en) * | 2014-09-18 | 2016-03-23 | Trollhetta As | Monitoring an environmental condition |
CN106292802A (en) * | 2016-09-18 | 2017-01-04 | 中国农业大学 | A kind of Intelligent Prediction Control System for fish and vegetable symbiotic system and method |
CN107743913A (en) * | 2017-10-13 | 2018-03-02 | 南京师范大学 | A kind of new Pelteobagrus fulvidraco transportation resources based on intelligent control |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO1997049279A1 (en) * | 1996-06-24 | 1997-12-31 | Board Of Regents, The University Of Texas System | Automated closed recirculating aquaculture filtration system |
US20170295727A1 (en) * | 2016-04-19 | 2017-10-19 | Suntracker Technologies Ltd. | Temporally modulated lighting system and method |
-
2018
- 2018-04-16 NO NO20180513A patent/NO345829B1/en active IP Right Review Request
-
2019
- 2019-04-15 WO PCT/NO2019/050086 patent/WO2019203661A1/en active Application Filing
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070257806A1 (en) * | 2004-09-27 | 2007-11-08 | Source Sentinel, Llc | System for monitoring quality of water system |
US20130206078A1 (en) * | 2010-05-18 | 2013-08-15 | Havforskiningsinstituttet | System and method for controlling feeding of farmed fish |
WO2015105523A1 (en) * | 2014-01-12 | 2015-07-16 | Kevin Friesth | Automated hybrid aquaponics and bioreactor system including product processing and storage facilities with integrated robotics, control system, and renewable energy system cross-reference to related applications |
GB2530300A (en) * | 2014-09-18 | 2016-03-23 | Trollhetta As | Monitoring an environmental condition |
CN106292802A (en) * | 2016-09-18 | 2017-01-04 | 中国农业大学 | A kind of Intelligent Prediction Control System for fish and vegetable symbiotic system and method |
CN107743913A (en) * | 2017-10-13 | 2018-03-02 | 南京师范大学 | A kind of new Pelteobagrus fulvidraco transportation resources based on intelligent control |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113490076A (en) * | 2021-07-07 | 2021-10-08 | 重庆市农业科学院 | Portable fish-vegetable symbiotic water quality rapid detection device |
CN113490076B (en) * | 2021-07-07 | 2023-04-18 | 重庆市农业科学院 | Portable fish-vegetable symbiotic water quality rapid detection device |
Also Published As
Publication number | Publication date |
---|---|
NO345829B1 (en) | 2021-08-30 |
NO20180513A1 (en) | 2019-10-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2019203661A1 (en) | C-fish – fish welfare control | |
Tripathy et al. | MyGreen: An IoT-enabled smart greenhouse for sustainable agriculture | |
US20200170227A1 (en) | Decision making system and method of feeding aquatic animals | |
Berckmans | Automatic on-line monitoring of animals by precision livestock farming | |
EP3939418B1 (en) | Automated livestock management system | |
TWI661770B (en) | Intelligent deep learning agricultural and fishery training system | |
WO2018184014A1 (en) | Communication and control systems and methods for monitoring information about a plurality of beehives | |
US20170039425A1 (en) | System and method for optimizing chemigation of crops | |
Srinidhi et al. | Smart Hydroponics system integrating with IoT and Machine learning algorithm | |
CN110915734A (en) | Aquaculture system based on Internet of things | |
Stefanec et al. | Governing the swarm: Controlling a bio-hybrid society of bees & robots with computational feedback loops | |
DE102010005563B4 (en) | A management system for operating a rearing apparatus and method for operating a rearing apparatus | |
Abdullah et al. | Development of aquaculture water quality real-time monitoring using multi-sensory system and internet of things | |
Mariano et al. | Evolving robot controllers for a bio-hybrid system | |
CN109359624A (en) | The prediction technique and prediction meanss of laying hen foraging behaviour neural network based | |
CN113408334B (en) | Crayfish full-chain data acquisition and intelligent detection method and device | |
Deepika et al. | Monitoring of Hydroponics Plant and Prediction of Leaf Disease using IOT | |
WO2020109348A1 (en) | Method of controlling a livestock farm | |
Nadaf et al. | Scope and potential applications of artificial intelligence in tropical tasar silkworm Antheraea mylitta D. seed production | |
Hinton | Started from the egg, now we’re here: Examining the development of consistent individual variation in growth and behavior of captive-reared, precocial wood ducks (Aix sponsa) | |
US20230301247A1 (en) | Machine learning systems for autonomous and semiautonomous plant growth | |
CN111047456A (en) | 5G-based industrial circulating water efficient culture remote control and diagnosis system | |
CN116548342B (en) | Intelligent feeding method, system, medium and electronic equipment for aquaculture | |
Parameswari | Artificial Intelligence and IOT Enabled Whiteflies Monitoring and Controlling System | |
Wintergerst | Modulators of foraging behavior by nectar-feeding bats (Glossophaginae) |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 19789376 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 19789376 Country of ref document: EP Kind code of ref document: A1 |