Intelligent Interconnected Control System

Abstract

An intelligent interconnected control system includes a central node signal-connected to an external control terminal and multiple edge nodes signal-connected to the central node. The central node receives a control request from the external control terminal relating to one of the edge nodes as a target node, generates and transmits a control command to the target node based on the control request, and the target node further uses machine learning algorithms based on acquired sensing data and the received control command to obtain a corrective operating rule related to its operation control, and updates its operating rule according to the corrective operating rule, thereby achieving self-learning and improving system adaptability.

Description

Technical Field

[0001] This invention relates to a control system for Internet of Things (IoT) connectivity, and more particularly to an intelligent interconnected control system with self-learning capabilities. Prior Technology

[0002] As living standards continue to improve, people desire a highly automated and convenient control system to remotely control the equipment they wish to manage. Thus, the Internet of Things (IoT) control system has emerged. This system connects various devices and systems via a network to achieve remote monitoring, data collection, and automated operation. It is widely used in smart factories, smart homes, industrial automation, and energy management.

[0003] However, existing IoT control systems mainly receive a control request from a remote location through their central gateway, and generate and transmit a control command to a device under control based on the control request. The device under control is only passively controlled by the remote location and cannot adaptively learn to optimize its operation. There are still areas for improvement in its use. Summary of the Invention

[0004] Therefore, the purpose of this invention is to provide an intelligent interconnected control system that can learn on its own and thus improve its adaptability.

[0005] Therefore, the intelligent interconnected control system of the present invention is signal-connected to an external control terminal. The intelligent interconnected control system includes a central node that is signal-connected to the external control terminal, and multiple edge nodes that are signal-connected to the central node.

[0006] Each edge node includes at least one of a sensing unit for collecting sensing data and a control unit for controlling according to an operating rule related to its operation.

[0007] The central node is used to receive a control request from one of the external control terminals related to one of the edge nodes to be controlled, and to generate and transmit a control command to the node to be controlled according to the control request. The node to be controlled is also used to obtain a correction operation rule related to its operation control by using a machine learning algorithm based on the obtained sensing data and the received control command, and to update its operation rule according to the correction operation rule.

[0008] The advantage of this invention is that the node to be controlled is also used to obtain the correction operation rules by using the machine learning algorithm based on the obtained sensing data and the received control commands, so as to achieve the effect of the node to be controlled learning itself and thereby improving the system adaptability. Simple Explanation of the Diagram

[0009] Other features and effects of the present invention will be clearly presented in the embodiments with reference to the drawings, wherein: Figure 1 is a block diagram illustrating the signal connection of an embodiment of the intelligent interconnected control system of the present invention to an external control terminal; Figure 2 is a flowchart illustrating a method for configuring a single node in an embodiment of the intelligent interconnected control system of the present invention; Figure 3 is a flowchart illustrating a transfer method in an embodiment of the intelligent interconnected control system of the present invention; Figure 4 is a flowchart illustrating an automatic control method implemented in an embodiment of the intelligent interconnected control system of the present invention; and Figure 5 is a flowchart illustrating a self-learning method of an embodiment of the intelligent interconnected control system of the present invention. Implementation

[0010] Referring to Figure 1, an embodiment of the intelligent interconnected control system of the present invention is connected to an external control terminal 3. The intelligent interconnected control system signal includes a central node 1 connected to the external control terminal 3, and multiple edge nodes 2 connected to the central node 1. The central node 1 can be connected to the external control terminal 3 via a communication network 100, and to the edge nodes 2 via a wireless network such as Bluetooth.

[0011] Each edge node 2 includes at least one of a Bluetooth communication unit (not shown), a sensing unit (not shown) for collecting sensing data, and a control unit (not shown) for controlling according to an operating rule related to its operation. Each edge node 2 is also used to transmit the collected sensing data to the external control terminal 3 via the central node 1 for real-time monitoring and analysis by the external control terminal 3.

[0012] Referring to Figures 1 and 2, the following will use a node configuration method to illustrate how the central node 1 configures the group relationships and publish / subscribe relationships of the edge nodes 2. The node configuration method includes the following steps.

[0013] In step 51, the central node 1 configures a unique identifier and a key corresponding to each edge node 2 according to a preset configuration rule. This allows data transmission to be encrypted, thereby improving the security and reliability of data transmission.

[0014] In step 52, the central node 1 configures the group to which each edge node 2 belongs according to the preset configuration rules, and transmits the configured unique identifier and the group code to the corresponding edge node 2. Furthermore, it adds the affiliation of each edge node 2 and its assigned group to a group relationship table. This group relationship table contains the unique identifier of each edge node 2 and the group code to which it belongs. Each edge node 2 can be configured with groups according to application requirements, thereby improving the system's flexibility.

[0015] In step 53, the central node 1 sets at least one of the edge nodes 2 as an information publishing end according to the preset configuration rules.

[0016] In step 54, for each information publishing end, the central node 1 sets at least one of the edge nodes 2 as the information subscription end corresponding to the information publishing end according to the preset configuration rules, transmits the unique identifier of the information subscription end corresponding to the information publishing end to the information publishing end, and adds the correspondence between the information publishing end and its corresponding information subscription end to a subscription relationship table. The subscription relationship table contains the unique identifier of each information publishing end and the unique identifier of its corresponding information subscription end.

[0017] Referring to Figures 1 and 3, the following will illustrate how another central node (not shown) configures the group relationships and publish-subscribe relationships of other edge nodes (not shown) using a transfer method. This transfer method includes the following steps. For example, this intelligent interconnected control system can be applied to a campus with multiple smart classrooms. The central node 1 and the edge nodes 2 are located in one of the smart classrooms. The control unit of the edge nodes 2 includes, for example, at least one light fixture, one air conditioner, one dehumidifier, and one air purifier in the smart classroom. Since each smart classroom has the same hardware, the control unit of another edge node in another smart classroom also includes at least another light fixture, another air conditioner, another dehumidifier, and another air purifier in that other smart classroom. Once the central node 1 and the edge nodes 2 in one smart classroom are configured, the same configuration can be easily copied to other smart classrooms through this transfer method.

[0018] In step 61, the other central node configures the unique identification code corresponding to each other edge node according to a preset configuration method that is the same as the rule for configuring the unique identification code corresponding to each edge node 2 in the central node 1. Edge nodes of the same category in different smart classrooms have the same unique identification code; that is, another air conditioner in one smart classroom has the same unique identification code as the air conditioner in the other smart classroom, and another dehumidifier in one smart classroom has the same unique identification code as the dehumidifier in the other smart classroom, and so on.

[0019] In step 62, the other central node loads the group relationship table to configure the group to which each additional edge node belongs.

[0020] In step 63, the other central node loads the subscription relationship table to set at least one of the other edge nodes as an information publisher and the information subscriber corresponding to each information publisher.

[0021] Referring to Figures 1 and 4, the following will use an automatic control method to illustrate how one of the information subscription terminals to be triggered performs automatic control. This automatic control method includes the following steps. It is worth noting that the operating rules of each subscription terminal include a trigger condition and a corresponding action to be executed.

[0022] In step 71, for each information publishing terminal, the information publishing terminal transmits the collected sensing data to its corresponding information subscription terminal.

[0023] In step 72, when the subscriber to be triggered receives sensing data from the corresponding information publishing terminal, it determines whether the received sensing data meets the stored trigger condition. When the subscriber to be triggered determines that the received sensing data meets the stored trigger condition, the process proceeds to step 73; when the subscriber to be triggered determines that the received sensing data does not meet the stored trigger condition, the process returns to step 72 to continue receiving the next piece of sensing data and determining whether the next piece of sensing data meets the stored trigger condition.

[0024] In step 73, the subscription client to be triggered executes the action to be performed.

[0025] Referring to Figures 1 and 5, the following will explain how one of the edge nodes 2, the node to be controlled, performs self-learning using a self-learning method. This self-learning method includes the following steps. It is worth noting that the node to be controlled can be any edge node 2 with a control unit.

[0026] In step 81, after the central node 1 receives a control request from one of the external control terminals 3 related to the node to be controlled, the central node 1 generates and transmits a control command to the node to be controlled according to the control request.

[0027] In step 82, after the node to be controlled receives the control command, the node to be controlled performs corresponding operations according to the control command, and uses the most recently acquired sensing data before receiving the control command and the control command as training data. The most recently acquired sensing data includes at least one of the sensing data most recently collected by the sensing unit of the node to be controlled before receiving the control command, and the sensing data most recently received by the node to be controlled from its corresponding information publishing terminal before receiving the control command.

[0028] In step 83, the node to be controlled determines whether the number of training data points obtained is greater than a preset number. When the node to be controlled determines that the number of training data points obtained is greater than the preset number, the process proceeds to step 84; when the node to be controlled determines that the number of training data points obtained is not greater than the preset number, the process returns to step 82 to continuously receive the next control command generated by the central node 1 and obtain the next training data point.

[0029] In step 84, the node to be controlled uses a machine learning algorithm based on the obtained training data to obtain a corrected operating rule related to its operation control, and updates its operating rule according to the corrected operating rule.

[0030] In step 85, the node to be controlled transmits the obtained correction operation rules to the external control terminal 3 via the central node 1.

[0031] Regarding how the node to be controlled obtains the correction operation rules, at least the following three implementation states are included.

[0032] In the first embodiment, the control unit of the node to be controlled is an air purifier, and the sensing unit of the information publishing terminal corresponding to the node to be controlled is an air quality detector. The operating rules of the node to be controlled include triggering conditions and corresponding actions to be executed. For example, the triggering conditions and corresponding actions to be executed are as follows: when the concentration of suspended particulate matter smaller than 2.5 micrometers (i.e., PM2.5) is greater than a preset concentration threshold, such as 15, the control unit of the node to be controlled, i.e., the air purifier, is activated. When the node to be controlled receives a control command indicating that the air purifier should be activated, the control unit of the node to be controlled is activated according to the control command. The most recently obtained sensing data (i.e., the most recently received sensing data from its corresponding information publishing terminal, i.e., the air quality detector, i.e., the PM2.5 concentration) and the control command are used as training data. After collecting training data multiple times until the number of training data obtained is greater than the preset number, the node to be controlled can use the machine learning algorithm based on the obtained training data to obtain the corrected operating rules. For example, if the air purifier is activated by control command multiple times when the air quality detector detects a PM2.5 concentration of 10, the trigger condition and corresponding action to be executed of the calibration operation rule obtained by the node to be controlled through the training data are, for example, when the PM2.5 concentration is greater than a preset concentration threshold, such as 10, the control unit of the node to be controlled, i.e., the air purifier, will be activated.

[0033] In the second embodiment, the control unit of the node to be controlled is a dehumidifier, the sensing unit of the node to be controlled is a humidity sensor, and the operating rules of the node to be controlled include operating conditions and their corresponding actions to be performed. For example, when the humidity sensor detects that the humidity is greater than a preset humidity threshold, such as 70%, the control unit of the node to be controlled, i.e., the dehumidifier, starts. When the node to be controlled receives a control command indicating that the dehumidifier should be started, the control unit of the node to be controlled is activated according to the control command. It uses the most recently acquired sensing data (i.e., the sensing data collected by the sensing unit most recently, i.e., the sensing data collected by the humidity sensor, i.e., humidity) and the control command as training data. After collecting training data multiple times until the number of training data points exceeds the preset number, the node to be controlled can use the machine learning algorithm based on the acquired training data to obtain the corrected operating rules. For example, if the dehumidifier is repeatedly activated by control commands when the humidity sensor detects a humidity of 65%, the operating conditions and corresponding actions of the corrected operating rules obtained by the node to be controlled through the training data are, for example, when the humidity sensed by the humidity sensor is greater than the preset humidity threshold, such as 65%, the control unit of the node to be controlled will activate the dehumidifier.

[0034] In the third embodiment, the control unit of the node to be controlled is an air conditioner, the sensing unit of the node to be controlled is a temperature sensor, the sensing unit of the information publishing terminal corresponding to the node to be controlled is a humidity sensor, and the operating rules of the node to be controlled include a judgment model for obtaining a judgment result based on the input temperature and humidity, wherein the judgment result of the judgment model indicates the result of starting the air conditioner or not starting the air conditioner. When the node to be controlled receives a control command instructing the air conditioner to start, the control unit of the node to be controlled is activated according to the control command. The control unit uses the most recently acquired sensing data (i.e., the sensing data collected by the most recently sensing unit, i.e., the sensing data collected by the temperature sensor, i.e., the temperature, and the sensing data received most recently from its corresponding information publishing terminal, i.e., the humidity sensor, i.e., the humidity) and the control command as training data. After collecting training data multiple times until the number of training data obtained is greater than the preset number, the node to be controlled can correct the judgment model according to the obtained training data to obtain the correction operation rule.

[0035] In summary, the intelligent interconnected control system of this invention enables the controlled node to obtain corrective operating rules based on the acquired sensing data and received control commands using machine learning algorithms. This achieves the effect of self-learning by the controlled node, thereby improving the system's adaptability. Furthermore, after the other central node configures the unique identification code corresponding to each other edge node according to the preset configuration method, it can quickly complete the group to which the other edge nodes belong and their publish-subscribe relationships based on the group relationship table and the subscription relationship table. This rapidly constructs the topological relationships between nodes, simplifying the node device connection and setup process for efficient management. Moreover, when the external control terminal 3 or the central node 1 does not transmit control commands, each edge node 2 can still perform self-control and response based on the sensing data obtained from the network topology or the sensing data collected by the corresponding sensing unit, thus effectively achieving the purpose of this invention.

[0036] However, the above description is merely an embodiment of the present invention and should not be construed as limiting the scope of the present invention. Any simple equivalent changes and modifications made in accordance with the scope of the patent application and the contents of the patent specification of the present invention shall still fall within the scope of the patent of the present invention.

[0037] 1: Central Node 2: Edge nodes 3: External control terminal 100: Communication Network Steps 51-54 Steps 61-63 71~73: Steps 81~85: Steps

Claims

1. An intelligent interconnected control system, signal-connected to an external control terminal, the intelligent interconnected control system comprising: a central node, signal-connected to the external control terminal; and a plurality of edge nodes, signal-connected to the central node, each edge node including at least one of a sensing unit for collecting sensing data and a control unit for controlling according to an operating rule related to its operation; wherein, The central node receives a control request from an external control terminal relating to one of the edge nodes to be controlled. Based on the control request, it generates and transmits a control command to the edge node. The edge node then performs corresponding operations according to the control command. Furthermore, the edge node uses machine learning algorithms to obtain a corrected operating rule based on acquired sensor data and the received control command, and updates its operating rule accordingly. The central node configures a unique identifier for each edge node according to a preset configuration rule. The central node is configured with a unique identifier and a group to which each edge node belongs. The central node transmits the unique identifier and the group code to the corresponding edge node. The central node is also configured to set at least one of the edge nodes as an information publisher. For each information publisher, the central node is configured to set at least one of the edge nodes as an information subscriber. The central node transmits the unique identifier of the information subscriber to the information publisher. For each information publisher, the central node transmits the collected sensing data to its corresponding information subscriber.

2. The intelligent interconnected control system as described in claim 1, wherein, The node to be controlled is also used to transmit the obtained correction operation rules to the external control terminal via the central node.

3. The intelligent interconnected control system as described in claim 1, wherein, The node to be controlled will use the most recent sensing data obtained before receiving the control command and the control command as training data. After the number of training data obtained is greater than a preset number, the node will use the machine learning algorithm to obtain the correction operation rule based on the training data.

4. The intelligent interconnected control system as described in claim 1, wherein, The operating rules of one of the information subscription terminals to be triggered include a trigger condition and a corresponding action to be executed. When the subscription terminal to be triggered receives sensing data from the corresponding information publishing terminal, it determines whether the received sensing data meets the trigger condition stored therein. When the subscription terminal to be triggered determines that the received sensing data meets the trigger condition stored therein, the subscription terminal to be triggered executes the action to be executed.

5. The intelligent interconnected control system as described in claim 1, wherein, The central node also establishes a group relationship table for each edge node and the group to which it belongs. The group relationship table contains a unique identifier for each edge node and the group code of the group to which it belongs.

6. The intelligent interconnected control system as described in claim 5, wherein, The central node also establishes a subscription relationship table for each information publishing end and its corresponding information subscription end. The subscription relationship table contains a unique identifier for each information publishing end and a unique identifier for its corresponding information subscription end.

7. The intelligent interconnected control system as described in claim 6 further comprises: another central node; and a plurality of additional edge nodes, signal-connected to the other central node and including at least one of another sensing unit for collecting sensing data and another control unit for performing control; wherein, The other central node configures the unique identifier corresponding to each other edge node according to a preset configuration method that is the same as the central node in configuring the unique identifier corresponding to each edge node. It loads the group relationship table to configure the group to which each other edge node belongs, and loads the subscription relationship table to set at least one of the other edge nodes as an information publishing end, and each information publishing end as an information subscription end.

8. The intelligent interconnected control system as described in claim 1, wherein, The central node also configures a corresponding key for each edge node.

Images