Holding Tank Monitoring System Based On Wireless Sensor Network And Monitoring Method

Abstract

In the breeding of aquatic products, keeping track of changes in the water quality environment is an urgent aquaculture problem to be solved. At present, single sensors are often used for data detection, and there is no coordination between the sensors or between sensors and actuators, resulting in a lack of a unified monitoring and control mechanism. The current invention discloses a holding tank monitoring system based on a wireless sensor network and a monitoring method provided thereof characterized by an unattended operation and automated execution by obtaining environmental data of a holding tank in real-time, and analyzing the data using a suitably trained machine learning algorithm and causing a control action on the actuator to perform a corrective action on the holding tank for optimal operation of the system.

Claims

1. A holding tank monitoring system based on a wireless sensor network, the system comprising of: a holding tank capable of supporting an aquatic life form, said tank having one or more sensors thereof, wherein said one or more sensors is capable of determining at least one or more conditions in the holding tank, and capable of transmitting detected condition to a remote computer; an actuator comprising a corrective measure for the at least one or more conditions in the holding tank determined by the sensor; a micro-controller configured capable to operate the actuator to make a corrective measure for the at least one or more conditions in the holding tank based on a corrective signal received from a remote computer configured with a suitable machine learning algorithm, where the corrective signal is derived from learning based on a plurality of inputs from a plurality of holding tanks; a communication module comprising of a network receiver and transmitter, the module capable of receiving signals from a remote computer, and transmitting sensor inputs to the remote computer; a network, and; a remote computing device comprising of at least a processor, memory and storage device, said computing device capable of receiving a detected condition from a sensor for at least one or more conditions in the holding tank, and storing received condition in a database in the storage device, wherein the computing device's memory is configured with a suitably trained machine learning algorithm to determine a conditions optimal for the aquatic life in the holding tank and generating a corrective signal such as but not limited to a temperature, turbidity, mineral composition etc, and transmitting the signal to the actuator to take a corrective action and cause the condition in the holding tank optimal.

2. The system as in claim 1, further comprising a remote monitoring and control device capable of transmitting a desired productivity at the holding tank to a remote computing device whose memory is configured with a suitably configured machine learning algorithm, wherein said computing device determines the conditions at the holding tank for attaining the productivity received from the monitoring and control device, and subsequently transmitting an actuation signal for the action of the actuator at the holding tank.

3. A method of training a machine learning algorithm for the monitoring and control of a holding tank, the method comprising of: collecting the productivity data of the aquatic life in a holding tank or a plurality of tanks; collecting of corresponding sensor conditions correlating to the productivity data; splitting of the productivity data and corresponding sensor conditions data into a training data set and validation data set; selection of a suitable algorithm, and thereafter using the training data set of sensor conditions to predict the productivity and obtain a suitable machine learning model; using the validation data set to verify the accuracy of the model; using the validation data set to verify the accuracy of the model, and; selecting the verified machine learning model for subsequent prediction.

4. The method of claim 3, further comprising the use of the suitably trained machine learning algorithm, the method of use comprising of: receiving at a computing device with a memory configured with a suitably trained machine learning algorithm the input from one or more sensor at a holding tank, where the one or more sensor detect at least one condition at the holding tank; determining from the input from one or more sensor at a holding tank, if the condition is optimal for the productivity of an aquatic life in the holding tank; determining the amount of correction required to bring back the at least one condition at the holding tank to an optimal level; generating a corrective signal identifying the correction to be made at the holding tank, and; transmitting a corrective signal for corrective action by an actuator.

5. The method of claim 4, further comprising a micro-controller determining the correct actuator for action on a corrective signal.

6. The method of claim 4, wherein the actuator causes a temperature change.

7. The method of claim 4, wherein the actuator causes a mineral composition alteration.

8. The method of claim 4, wherein the actuator causes a water flow change.

9. The method of claim 4, wherein the actuator causes a change in water turbidity.

10. A method of performing a corrective action at a holding tank by an actuator to restore optimal conditions as determined by a remote computing device whose memory is configured with a suitably trained machine learning algorithm, the method comprising of: receiving a corrective signal from a suitably configured computing device; determining the correct actuator for performing the corrective action by a micro-controller, and; making the actuator to perform a corrective action to bring the at least one condition in the holding tank to optimal level.

11. The method of claim 10, further comprising the transmitting a desired productivity at the holding tank by a monitoring and control device, wherein the remote computing device determines the conditions at the holding tank for attaining the productivity received from the monitoring and control device, and transmits an actuation signal for the action of the actuator at the holding tank.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The novel features believed to be characteristic of the illustrative embodiments are set forth in the appended claims. The illustrative embodiments, however, as well as a preferred mode of use, further objectives and descriptions thereof, will best be understood by reference to the following detailed description of one or more illustrative embodiments of the present disclosure when read in conjunction with the accompanying drawings, wherein:

[0031] FIG. 1 of the diagrams is a schematic architectural view of the system as described according to the current invention.

[0032] FIG. 2 of the diagrams illustrates the method of training a machine learning algorithm according to the current invention.

[0033] FIG. 3 of the diagrams illustrates a method of using the trained machine learning algorithm according to the current invention.

[0034] FIG. 4 of the diagrams illustrates the mechanism of activating a control mechanism performed by the actuator for a corrective mechanism according to the current invention.

[0035] FIG. 5 of the diagrams a control mechanism performed by the monitoring and control device according to the current invention.

DETAILED DESCRIPTION OF THE DRAWINGS

[0036] Hereinafter, the preferred embodiment of the present invention will be described in detail with reference to the accompanying drawings. The terminologies or words used in the description and the claims of the present invention should not be interpreted as being limited merely to their common and dictionary meanings. On the contrary, they should be interpreted based on the meanings and concepts of the invention in keeping with the scope of the invention based on the principle that the inventor(s) can appropriately define the terms in order to describe the invention in the best way.

[0037] It is to be understood that the form of the invention shown and described herein is to be taken as a preferred embodiment of the present invention, so it does not expressly limit the technical spirit and scope of this invention. Accordingly, it should be understood that various changes and modifications may be made to the invention without departing from the spirit and scope thereof.

[0038] In a first embodiment according to FIG. 1 of the diagrams is a schematic architectural view of the system as described according to the current invention. Illustrated in the figure is a holding tank 1, a micro-controller 2, a sensor or plurality of sensors 3, an actuator or plurality of actuators 4, a remote computer or server 5 with a storage device 6, a monitoring and control device 7, a networks 8, a signal receiver and transmitter module 9, an aquatic life 10 as well as a computer-implemented module 50 configured on the remote computer 5, said module comprising a suitably trained machine learning algorithm to determine the conditions that would be optimal for the aquatic life form in the holding tank and generates a corrective signal e.g. temperature, turbidity, mineral composition etc, and transmits the signal to the communication module, which makes the actuator to take a corrective action.

[0039] The holding tank 1 is any such tank capable of supporting an aquatic life form e.g. fish, plants or any such in an aqueous environment. There are numerous instances where fish is the preferred example of aquatic life in this disclosure, however, other life form are also anticipated. As an example, water quality is the most important factor affecting fish health and performance in aquaculture production systems, and the same is true to may aquatic life forms. Good water quality refers to what the fish life thrives best. This means that a farmer must understand the water quality requirements of the fish under culture very well. Fish live in and are totally dependent on the water they live in for all their needs. Additionally, different fish species have different and specific range of water quality aspects, which may include one or more of the temperature, pH, oxygen concentration, salinity, hardness, etc. within which they can survive, thrive, grow and reproduce.

[0040] Within these tolerance limits, each species has its own optimum range, that is, the range within which it performs best. It is therefore very important for fish producers to ensure that the physical and chemical conditions of the water remain, as much as possible, within the optimum range of the fish under culture all the time. Outside these optimum ranges, fish will exhibit poor growth, erratic behaviour, and disease symptoms or parasite infestations. Under extreme cases, or where the poor conditions remain for prolonged periods of time, fish mortality may occur. Holding tank water contains two major groups of substances: (a) suspended particles made of non-living particles and very small plants and animals, the plankton, and (b) dissolved substances made of gases, minerals and organic compounds.

[0041] It should be understood that the composition of holding tank water changes continuously, depending on climatic and seasonal changes, the flow of water, and on how a holding tank is used. It is the aim of good management to control the composition to yield the best conditions for the fish. For producers to be able to maintain ideal holding tank water quality conditions, they must understand the physical and chemical components contributing to good or bad water quality. Sensors can provide a great deal of understanding on these conditions.

[0042] Further, the micro-controller 2 is configured to operate the actuator 4 to make a corrective measure for the at least one or more conditions in the holding tank based on a corrective signal received from the remote computer 5 configured with a suitable machine learning algorithm, where the corrective signal is derived from learning based on a plurality of inputs preferably from a plurality of holding tanks. It should be noted that the corrective mechanism is activated at any such preferred frequency and autonomously without a manual input, but rather based on a signal from the remote computer.

[0043] Now, the sensor or plurality of sensors 3 according to the current disclosure, is capable of determining at least one or more conditions in the holding tank. Preferably located in the holding tank and capable of transmitting detected condition to a remote computer. Some of these sensors could be capable of detecting one or more of the following: Temperature, Turbidity, Water pH and acidity, Alkalinity and hardness, Dissolved gases: oxygen, carbon dioxide, nitrogen, Ammonia content or Toxic materials.

[0044] The actuator or plurality of actuators 4 comprise of a mechanism for performing a corrective measure for the at least one or more conditions in the holding tank determined by the sensor 3. For example, it could be heater to heat the water and as such increase the temperature, a fan to blow the surface of the water to reduce temperature, a filter to remove ammonia, a mechanism to pass air into the tank to increase the amounts of dissolved oxygen, a filter to remove suspended particles or a chemical to alter to pH etc. The actuator receives a corrective signal from the micro-controller 2. it should be noted that the signal is received autonomously and the micro-controller activates the actuator actively, but it is also anticipated that there could be an element of human control for the system.

[0045] Further in the figure is a a remote computer or server 5 with a storage device 6, wherein according to the current invention, the remote computer 5 comprises of at least a processor, memory and storage. The memory of the remote computer is configured with a suitably trained machine learning algorithm, which makes up the computer-implemented module 50. Such module comprises an algorithm that could comprise of any such trained algorithm such as: Linear Regression, Logistic Regression, Decision Tree, SVM, Naive Bayes, kNN, K-Means, Random Forest, Dimensionality Reduction Algorithms, Gradient Boosting algorithms (GBM, XGBoost, LightGBM, CatBoost), among others. On the other hand, the storage device 6 comprises of a database capable of receiving from above sensors detected condition for at least one or more conditions in the holding tank. For each detected condition, there are inputs from multiple holding tanks, probably at different locations. However, the setup could also work for a single holding tank.

[0046] Furthermore, the productivity of all holding tanks (if more than one are included) is measured and recorded, so as to be able to determine the correlation between detected conditions and the productivity. This is preferably performed using the monitoring and control device 7. For purposes of this disclosure, the productivity could be the average fish size, the quantity of produced fish, the rate of reproduction, disease rate, the weight of fish produced or even mortality rate of fish. However, it should be understood that similar metrics for other aquatic life forms could be measured as well, and this only forms a suitable example for a person skilled in the art to perform the invention.

[0047] Further still is a network 8, which may comprise of one or more mechanisms capable of transmitting data between any number of computers or any such devices with network interfaces. In the current invention, such networks could comprise broadband networks, fibre optic, Ethernet, cabling, electromagnetic waves etc.

[0048] Further still is a signal receiver and transmitter module 9, which comprises of a network receiver, the module capable of receiving signals from a remote computer, where received signals are processed by the micro-controller 2 to activate the actuator 4 to perform a corrective measure for the at least one or more conditions in the holding tank 1 as determined by the sensor 3.

[0049] In a second embodiment according to the FIG. 2 of the diagrams, it is illustrated a method of training a machine learning algorithm according to the current invention. The method comprises of a step 20 of collecting the productivity data of the aquatic life in a holding tank or a plurality of tanks. The next step 21 entails the collecting of corresponding sensor conditions correlating to the productivity data. In the step 22 is the splitting of the productivity data and corresponding sensor conditions data into a training data set and validation data set. Further in the step 23 is the selection of a suitable algorithm, and thereafter using the training data set of sensor conditions to predict the productivity and obtain a ML model. Subsequently in step 24, is using the validation data set to verify the accuracy of the model. Finally, in the step 25, the method involves the selection of the best ML model for your data set.

[0050] In a further embodiment according to the FIG. 3 of the diagrams, it is illustrated a method of using the trained machine learning algorithm according to the current invention. The method comprises of a step 30 of receiving at a computing device with a memory configured with a suitably trained machine learning algorithm the input from one or more sensor at a holding tank, where the sensor detect at least one condition at the holding tank. The subsequent step 31 entails determining from the input from one or more sensor at a holding tank, if the condition is optimal, and if not optimal, perform the subsequent step 32 of determining the amount of correction required to bring back the at least one condition at the holding tank to an optimal level. In the subsequent step 33, it is generated a corrective signal identifying the correction to be made at the holding tank. Finally in 34, a corrective signal is transmitted for corrective action by the actuator.

[0051] Specifically, a remote computer receives the conditions from the sensor, stores received condition in a database. The computer uses a suitably trained machine learning algorithm to determine the conditions that would be optimal for the aquatic life form in the holding tank and generates a corrective signal e.g. temperature, turbidity, mineral composition etc, and transmits the signal to the communication module, which makes the actuator to take a corrective action. The corrective action is automatically generated, and the signals provided are improved over time based on a learning model.

[0052] In the subsequent embodiment according to the FIG. 4 of the diagrams is a mechanism of activating a control mechanism performed by the actuator for a corrective mechanism according to the current invention. The method comprises of a step 40 of receiving a corrective signal from a suitably configured computer, preferably a remote computer. In the subsequent step 41 is the determining the correct actuator for performing the corrective action. It is preferable that a micro-controller makes the determination of the correct actuator, and making the actuator perform the corrective action. Finally, in the step 42 is making the actuator to perform a corrective action to bring the at least one condition in the holding tank to optimal level.

[0053] In a final embodiment exemplified by the FIG. 5 of the diagrams is illustrated a control mechanism performed by the monitoring and control device according to the current invention. The first step 50 entails receiving at a monitoring and control device a predicted productivity as determined by a suitably trained machine learning model. Next in the step 51 is the receiving the determined conditions at the holding tank by one or more sensors a said holding tank. Finally in the step 52 is the transmitting a desired productivity at the holding tank by the monitoring and control device, wherein the remote computer 5 with its memory configured with a suitably configured machine learning algorithm determines the conditions at the holding tank for attaining the productivity received from the monitoring and control device, and transmit an actuation signal for the action of the actuator at the holding tank.

[0054] Although a preferred embodiment of the present invention has been described for illustrative purposes, those skilled in the art will appreciate that various modifications, additions and substitutions are possible, without departing from the scope and spirit of the invention as disclosed in the accompanying claims.

INDUSTRIAL APPLICATION

[0055] The current invention relates to the use and manufacture intelligent holding tanks for aquaculture and remote monitoring and control of the environmental conditions at water tanks.