Intelligent scene linkage edge computing gateway

Abstract

The invention relates to the technical field of edge computing gateways, and discloses an intelligent scene linkage edge computing gateway, which comprises a communication module, a state verification module, a fault prediction module and a control module, wherein the communication module comprises a first communication unit and a second communication unit; wherein the first communication unit is used for communicating with a lamp, a sensor and a controller and collecting scene data and lamp state data; the second communication unit is used for communicating with an Internet of Things platform; the state verification module is used for judging whether the lamp is in an abnormal state; the fault prediction module carries out fault prediction on the lamp based on the lamp state data; the control module is used for generating a control signal, and the control signal is used for controlling the state of the lamp. According to the invention, the problem of system instability caused by lamp faults can be reduced, energy can be saved, and unnecessary energy consumption can be reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of edge computing gateways, and in particular to an edge computing gateway linked with intelligent scenarios. Background Art

[0002] Traditional lighting control often relies on fixed switching modes or simple timing functions, which cannot meet users' needs for personalization and intelligence. With the advancement of Internet of Things technology, more and more smart devices are integrated into home networks. How to achieve efficient control of lamps and their intelligent linkage with scenes has become a problem that needs to be solved urgently. The edge computing gateway generally used for lamp management and control cannot determine whether the current lamp status is reasonable, such as keeping the lights on when no one is around or turning on strong lights during the day; it also does not have the function of predictive abnormality judgment. Therefore, the present invention provides an edge computing gateway with intelligent scene linkage, which predicts possible faults of the lighting system through historical data, and provides the function of scene-linked lamp status rationality judgment.

[0003] For example, the patent application with publication number CN112565451A discloses a DALI gateway and DALI lamp control system, including a power module, an MCU control module, a WIFI module and a DALI conversion module; the MCU control module is respectively connected to the DALI conversion module and the WIFI module; the power module is respectively connected to the DALI module, the MCU control unit and the WIFI module. The WIFI module is used to connect to an external control device through a WIFI network, and the external control device sends a control instruction to the DALI gateway to adjust the lamp. The external device is connected through the WIFI network to achieve wireless isolation. Even if a lamp controlled by a certain DALI gateway fails, it will not affect the remaining lines and the configuration is simple and flexible, and is not restricted by the line; secondly, most of the existing smart devices have WIFI functions. By connecting to the DALI gateway, users can control lamps through the WIFI network, enriching the use scenarios of the DALI system. However, this patent still has the problem raised by this background technology: it is impossible to judge whether the current lamp state is reasonable.

[0004] The information disclosed in this background technology section is only intended to enhance the understanding of the overall background of the invention and should not be regarded as an acknowledgement or any form of suggestion that the information constitutes the prior art already known to ordinary technicians in this field. Summary of the invention

[0005] The technical problem to be solved by the present invention is to overcome the defects of the prior art and provide an edge computing gateway with intelligent scene linkage to reduce the system instability problem caused by lamp failure, help save energy and reduce unnecessary energy consumption.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] An edge computing gateway for intelligent scene linkage includes a communication module, a state verification module, a fault prediction module, and a control module; wherein:

[0008] The communication module includes a first communication unit and a second communication unit; wherein the first communication unit is used to communicate with lamps, sensors, and controllers to collect scene data and lamp status data; the second communication unit is used to communicate with the IoT platform;

[0009] The state verification module determines whether the lamp is in an abnormal state based on the scene data and the lamp state data; the abnormal state includes abnormal switch state and abnormal brightness;

[0010] The fault prediction module performs fault prediction of the lamp based on the lamp status data;

[0011] The control module is used to generate a control signal, and the control signal is used to control the state of the lamp.

[0012] As a preferred solution of the edge computing gateway for intelligent scene linkage of the present invention, wherein: the scene data includes space occupancy status, date tag, and time point; wherein the space occupancy status includes whether there is someone or no one, which is determined based on a human body sensor; the date tag includes working day and non-working day;

[0013] The lamp status data includes switch status, light brightness, lamp temperature, real-time power consumption, usage time, and turn-on time; wherein, the switch status includes on and off; the usage time is the cumulative time of the lamp from installation to the current moment; the turn-on time refers to the cumulative time of the lamp being turned on this time.

[0014] As a preferred solution of the edge computing gateway for intelligent scene linkage described in the present invention, wherein: the state verification module includes a first processing unit and a first state unit;

[0015] The first processing unit is used to process the scene data and the lamp status data; including data cleaning and standardization of the lamp status data, and encoding the scene data and the lamp status data after standardization into a feature vector;

[0016] The first state unit is configured with a first state model for determining whether the switch state of the lamp is abnormal; the first state model is a trained support vector machine, the input of which includes space occupancy status, date label, time point, and switch state, and the output is a determination result of whether the switch state of the lamp is abnormal.

[0017] As a preferred solution of the edge computing gateway for intelligent scene linkage described in the present invention, wherein: the state verification module is configured with a state verification strategy, which is as follows: when the switch state of the lamp is abnormal, the first state unit sends a first prediction instruction to the fault prediction module, and the fault prediction module responds to the first prediction instruction and returns the probability of electrical connection failure and control circuit failure; the first state unit is configured with a first fault probability threshold. If the probability of electrical connection failure and control circuit failure is lower than the first fault probability threshold, a prompt of abnormal switch state is sent to the Internet of Things platform through the second communication unit; if the Internet of Things platform returns information for adjusting the switch state, the first state unit sends a control instruction for adjusting the switch state to the control module, and the control module responds to the control instruction and generates a control signal to adjust the switch state; if at least one of the probabilities of electrical connection failure and control circuit failure is not lower than the first fault probability threshold, a corresponding fault prompt is sent to the Internet of Things platform through the second communication unit.

[0018] As a preferred solution for the edge computing gateway for intelligent scene linkage described in the present invention, the state verification module also includes a second state unit, which is configured with a second state model for determining whether the lamp has brightness abnormalities; the second state model is a trained support vector machine, and its input includes space occupancy status, date label, time point, and light brightness, and its output is a determination result of whether the lamp has brightness abnormalities.

[0019] As a preferred solution of the edge computing gateway for intelligent scene linkage described in the present invention, the state verification strategy also includes: when the brightness of the lamp is abnormal, the second state unit sends a second prediction instruction to the fault prediction module, and the fault prediction module responds to the second prediction instruction and returns the probability of lamp aging and light source failure; the second state unit is configured with a second fault probability threshold. If the probability of lamp aging and light source failure is lower than the second fault probability threshold, a brightness abnormality prompt is sent to the Internet of Things platform through the second communication unit; if the Internet of Things platform returns information for adjusting the brightness of the lamp, the second state unit sends a control instruction to adjust the brightness of the lamp to the control module, and the control module responds to the control instruction and generates a control signal to adjust the brightness of the lamp; if at least one of the probabilities of lamp aging and light source failure is not lower than the second fault probability threshold, a corresponding fault prompt is sent to the Internet of Things platform through the second communication unit.

[0020] As a preferred solution of the edge computing gateway for intelligent scene linkage of the present invention, wherein: the fault prediction module includes a second processing unit;

[0021] The second processing unit is used to process the lamp status data; specifically as follows:

[0022] Performing data cleaning on the lamp status data;

[0023] Calculating a brightness attenuation index based on the light brightness, and calculating a power consumption fluctuation index based on the real-time power consumption;

[0024] The lamp temperature, usage time, turn-on time, brightness decay index, and power consumption fluctuation index are encoded into a feature vector.

[0025] As a preferred solution of the edge computing gateway for intelligent scene linkage described in the present invention, the method for calculating the brightness attenuation index is as follows:

[0026] Extract the light brightness after M consecutive data cleaning, and establish a time series of light brightness with a length of M; calculate the brightness attenuation index based on the time series of light brightness, the formula is as follows:

[0027]

[0028] Where D represents the brightness attenuation index; b i represents the i-th element in the time series of the light brightness; b i+1 The i+1th element in the time series representing the brightness of the light;

[0029] The method for calculating the power consumption fluctuation index is as follows:

[0030] Extract the real-time power consumption after N consecutive data cleanings, and establish a time series of real-time power consumption with a length of N; perform n sliding interceptions on the time series of real-time power consumption with a time window of fixed length, and calculate the mean of the real-time power consumption of the n sliding interceptions respectively; calculate the power consumption fluctuation index, the formula is as follows:

[0031]

[0032] Where F represents the power consumption fluctuation index; p j represents the mean value of the real-time power consumption captured by the jth sliding operation; p k Represents the kth element in the time series of the real-time power consumption.

[0033] As a preferred solution of the edge computing gateway for intelligent scene linkage described in the present invention, wherein: the fault prediction module also includes a prediction model unit; the prediction model unit is configured with a fault prediction model for predicting lamp faults; the output of the fault prediction model is the probability of occurrence of each type of fault in the lamp, and the fault types include electrical connection failure, control circuit failure, lamp aging, light source failure, and lamp overheating; the input of the fault prediction model includes the characteristic vectors of the lamp temperature, usage time, turn-on time, brightness attenuation index, and power consumption fluctuation index.

[0034] As a preferred solution of the edge computing gateway for intelligent scene linkage described in the present invention, wherein: the fault prediction module is configured with a prediction auxiliary strategy, which is as follows: the input of the fault prediction model also includes a lamp status code; the lamp status code includes a switch status abnormality code and a brightness abnormality code; the fault prediction module performs lamp fault prediction at a fixed period, and sends the probability of each fault to the Internet of Things platform through the second communication unit, wherein the switch status abnormality code and the brightness abnormality code are both set to 0; if the fault prediction module receives the first prediction instruction, the switch status abnormality code is set to 1; if the fault prediction module receives the second prediction instruction, the brightness abnormality code is set to 1.

[0035] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0036] The present invention does not rely solely on a single factor, such as time, but takes into account a variety of situational factors to make situational awareness more comprehensive. Reasonable situational judgment helps save energy, reduce unnecessary energy consumption, and improve energy efficiency. By performing predictive abnormal judgment, preventive measures can be taken before lamp failure occurs, reducing system instability caused by lamp failure.

[0037] An effective interactive feedback loop is formed between the functional modules, which improves the response speed and accuracy of the system, enables the lighting system to self-regulate and optimize, and reduces the need for human intervention. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0039] Figure 1 A schematic diagram of the structure of an edge computing gateway for intelligent scene linkage provided by the present invention;

[0040] Figure 2 A schematic diagram of the collaborative working mode of each module of an edge computing gateway with intelligent scene linkage provided by the present invention;

[0041] Figure 3 This is a structural schematic diagram of the fault prediction model provided by the present invention. DETAILED DESCRIPTION

[0042] The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. The embodiments of the present invention and the technical features in the embodiments may be combined with each other unless there is a conflict.

[0043] This embodiment introduces an edge computing gateway for intelligent scene linkage. Figure 1 , the edge computing gateway includes a communication module, a state verification module, a fault prediction module, and a control module; wherein:

[0044] The communication module includes a first communication unit and a second communication unit; wherein the first communication unit is used to communicate with lamps, sensors, and controllers to collect scene data and lamp status data; the second communication unit is used to communicate with the IoT platform;

[0045] The scene data includes space occupancy status, date tag, and time point; wherein the space occupancy status includes whether there are people or no people, which is determined based on a human body sensor; the date tag includes working day and non-working day;

[0046] The lamp status data includes switch status, light brightness, lamp temperature, real-time power consumption, usage time, and turn-on time; wherein, the switch status includes on and off; light brightness is collected by light sensors arranged around the lamp; lamp temperature is collected by temperature sensors arranged around the lamp; real-time power consumption is the real-time power consumption of the lamp; usage time is the cumulative time of the lamp from installation to the current moment; turn-on time refers to the cumulative time the lamp is turned on this time.

[0047] The state verification module determines whether the lamp is in an abnormal state based on the scene data and the lamp state data; the abnormal state includes abnormal switch state and abnormal brightness;

[0048] The state verification module includes a first processing unit, a first state unit, and a second state unit;

[0049] The first processing unit is used to process the scene data and the lamp status data; including data cleaning and standardization of the lamp status data, and encoding the scene data and the lamp status data after standardization into a feature vector;

[0050] The first state unit is configured with a first state model for judging whether the switch state of the lamp is abnormal; the first state model is a trained support vector machine, the input includes space occupancy state, date label, time point, switch state (these data are input into the support vector machine in the form of feature vectors), and the output is the judgment result of whether the switch state of the lamp is abnormal; the state verification module is configured with a state verification strategy, which is as follows: when the switch state of the lamp is abnormal, the first state unit sends a first prediction instruction to the fault prediction module, and the fault prediction module responds to the first prediction instruction and returns the probability of electrical connection failure and control circuit failure; the first state unit is configured with a first fault probability threshold, if the probability of electrical connection failure and control circuit failure is lower than the first fault probability threshold, the second communication unit sends a prompt of abnormal switch state to the Internet of Things platform; the Internet of Things platform is responsible for interacting with the user; if the Internet of Things platform returns information for adjusting the switch state, the first state unit sends a control instruction for adjusting the switch state to the control module, and the control module responds to the control instruction and generates a control signal to adjust the switch state; if at least one of the probabilities of electrical connection failure and control circuit failure is not lower than the first fault probability threshold, the corresponding fault prompt is sent to the Internet of Things platform through the second communication unit. The collaboration process between modules is as follows: Figure 2 shown.

[0051] The second state unit is configured with a second state model for determining whether the lamp has abnormal brightness; the second state model is a trained support vector machine, the input of which includes space occupancy status, date label, time point, and light brightness (these data are input into the support vector machine in the form of feature vectors), and the output is a determination result of whether the lamp has abnormal brightness;

[0052] The state verification strategy also includes: when the brightness of the lamp is abnormal, the second state unit sends a second prediction instruction to the fault prediction module, and the fault prediction module responds to the second prediction instruction and returns the probability of lamp aging or light source failure; the second state unit is configured with a second fault probability threshold, and if the probability of lamp aging or light source failure is lower than the second fault probability threshold, a brightness abnormality prompt is sent to the Internet of Things platform through the second communication unit; if the Internet of Things platform returns information for adjusting the brightness of the lamp, the second state unit sends a control instruction for adjusting the brightness of the lamp to the control module, and the control module responds to the control instruction and generates a control signal to adjust the brightness of the lamp; if at least one of the probabilities of lamp aging or light source failure is not lower than the second fault probability threshold, a corresponding fault prompt is sent to the Internet of Things platform through the second communication unit.

[0053] Collect historical data, any piece of historical data includes the scene data and the corresponding lamp status data; train the first state model and the second state model respectively based on the historical data, and save the trained model parameters to a file; add labels of normal switch status and abnormal switch status to each piece of historical data used to train the first state model; add labels of normal brightness and abnormal brightness to each piece of historical data used to train the second state model; when the status verification module determines whether the lamp is in an abnormal state, load the model parameters from the file, and make an abnormal state determination of the lamp based on the model parameters.

[0054] Support vector machine is a machine learning algorithm, especially suitable for high-dimensional data and small sample data sets. In the judgment of abnormal switch state and abnormal brightness, support vector machine can handle multi-dimensional features well and has good generalization ability. When the space is unoccupied but the lamp is turned on, it may indicate that the lamp state is abnormal; when the space is occupied but the lamp is turned off, it may also indicate that the lamp state is abnormal. Usage habits may be different on different days. For example, there will be obvious differences in usage patterns between weekdays and weekends. On weekdays, lamps may need to be used for a long time, but not on weekends. Usage habits and needs are different in different time periods. For example, lamps are usually not used late at night, and if they are turned on, it may indicate an abnormality. At the same time, different space occupancy states, different dates, and different times have different requirements for light brightness. By learning from historical data, support vector machines can accurately determine whether lamps have abnormal states.

[0055] The fault prediction module performs fault prediction of the lamp based on the lamp status data;

[0056] The fault prediction module includes a second processing unit and a prediction model unit;

[0057] The second processing unit is used to process the lamp status data; specifically as follows:

[0058] Performing data cleaning on the lamp status data;

[0059] The brightness attenuation index is calculated based on the brightness of the light as follows:

[0060] Extract the light brightness after M consecutive data cleaning, and establish a time series of light brightness with a length of M; M is a positive integer; calculate the brightness attenuation index based on the time series of light brightness, the formula is as follows:

[0061]

[0062] Where D represents the brightness attenuation index; b i represents the i-th element in the time series of the light brightness, that is, the light brightness at the i-th time point; bi+1 Represents the i+1th element in the time series of the light brightness.

[0063] The brightness decay index reflects the degree of decay of the brightness of a lamp over a period of time. It is a standardized measure of the change in the brightness of a lamp, eliminating the differences between different lamps. The brightness decay index of different lamps can be directly compared, which helps to unify the standards. By tracking the trend of the brightness decay index, it is possible to more accurately determine whether the lamp is aging.

[0064] The power consumption fluctuation index is calculated based on the real-time power consumption, and the method is as follows:

[0065] Extract the real-time power consumption after N consecutive data cleaning, and establish a time series of real-time power consumption with a length of N; N is a positive integer; perform n sliding interceptions on the time series of the real-time power consumption with a time window of fixed length, and calculate the average of the real-time power consumption of the n sliding interceptions respectively; let the length of the time window be m, then any sliding interception obtains m consecutive real-time power consumption; average the m consecutive real-time power consumption to obtain the average of the real-time power consumption of the sliding interception; calculate the power consumption fluctuation index, the formula is as follows:

[0066]

[0067] Where F represents the power consumption fluctuation index; p j represents the mean value of the real-time power consumption captured by the jth sliding operation; p k The kth element in the time series representing the real-time power consumption, that is, the real-time power consumption at the kth time point;

[0068] The power consumption fluctuation index reflects the degree of fluctuation of the power consumption of the lamp relative to the average power consumption. It can more sensitively detect abnormal fluctuations in power consumption, thereby more sensitively discovering electrical connection problems. At the same time, the power consumption fluctuation index is a standardized measure of power consumption changes, which also eliminates the differences between different lamps.

[0069] Encode the lamp temperature, usage time, turn-on time, brightness decay index, and power consumption fluctuation index into a feature vector;

[0070] The prediction model unit is configured with a fault prediction model for predicting lamp faults; the output of the fault prediction model is the probability of occurrence of each type of fault of the lamp, and the fault types include electrical connection fault, control circuit fault, lamp aging, light source fault, and lamp overheating;

[0071] Causes of electrical connection failure include poor wire contact, loose plugs, etc., which may cause abnormal shutdown of the lamp and excessive energy consumption. It is necessary to check and tighten the connection terminals or replace the connection wires. Control circuit failure is caused by failure of relays or switching elements in the control circuit, which may cause the lamp to fail to shut down normally. Users are required to check and replace relays or switching elements. Lamp aging is due to the long cumulative use time of the lamp. As the opening time increases, the brightness of the light will gradually decrease. Light source failure can also cause abnormal light brightness. Due to long opening time, poor heat dissipation, ambient temperature and other reasons, by monitoring the temperature around the lamp, it is possible to predict whether the lamp is at risk of overheating.

[0072] The input of the fault prediction model includes the characteristic vector of the lamp temperature, usage time, opening time, brightness decay index, and power consumption fluctuation index; the fault prediction model is a multi-layer perceptron, referring to Figure 3 , including an input layer, a hidden layer, and an output layer; wherein the input layer is used to receive the characteristic vectors of lamp temperature, usage time, opening time, brightness attenuation index, and power consumption fluctuation index as input variables; the hidden layer is used to extract the characteristics of the input variables; the output layer calculates and outputs the probability of occurrence of each fault type based on the characteristics of the input variables. Multilayer perceptrons can capture nonlinear relationships in data and are suitable for lamp fault prediction, because the occurrence of faults is often not a simple linear relationship. Multilayer perceptrons can process high-dimensional data, are suitable for multiple feature inputs, are suitable for the fault prediction task of the present invention, process multiple input features, and simultaneously output the probabilities of multiple faults.

[0073] The fault prediction module is configured with a prediction auxiliary strategy, which is as follows: the input of the multi-layer perceptron also includes a lamp status code; the lamp status code includes a switch status abnormality code and a brightness abnormality code; the fault prediction module predicts the fault of the lamp at a fixed period, and sends the probability of each fault to the IoT platform through the second communication unit, wherein the switch status abnormality code and the brightness abnormality code are both set to 0; if the fault prediction module receives the first prediction instruction, the switch status abnormality code is set to 1; if the fault prediction module receives the second prediction instruction, the brightness abnormality code is set to 1. Based on the prediction auxiliary strategy, when the lamp is in an abnormal situation that does not match the scene, a more targeted fault prediction can be made based on the specific abnormal situation, thereby improving the accuracy of the fault prediction.

[0074] The control module is used to generate a control signal, and the control signal is used to control the state of the lamp.

[0075] The control module includes an instruction parsing unit and a control signal unit;

[0076] The instruction parsing unit is used to parse the control instruction and obtain the adjustment amount of the lamp state; the control signal unit generates a control signal based on the adjustment amount of the lamp state.

[0077] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the purpose and scope of protection of the present invention, which are all within the protection of the present invention.

Claims

1. An edge computing gateway for intelligent scene linkage, characterized in that: It includes a communication module, a state verification module, a fault prediction module, and a control module; wherein: The communication module includes a first communication unit and a second communication unit; wherein the first communication unit is used to communicate with lamps, sensors, and controllers to collect scene data and lamp status data; the second communication unit is used to communicate with the IoT platform; The state verification module determines whether the lamp is in an abnormal state based on the scene data and the lamp state data; the abnormal state includes abnormal switch state and abnormal brightness; The fault prediction module performs fault prediction of the lamp based on the lamp status data; The control module is used to generate a control signal, and the control signal is used to control the state of the lamp.

2. The edge computing gateway for intelligent scene linkage according to claim 1, characterized in that: The scene data includes space occupancy status, date tag, and time point; wherein the space occupancy status includes whether there are people or no people, which is determined based on a human body sensor; the date tag includes working day and non-working day; The lamp status data includes switch status, light brightness, lamp temperature, real-time power consumption, usage time, and turn-on time; wherein, the switch status includes on and off; the usage time is the cumulative time of the lamp from installation to the current moment; the turn-on time refers to the cumulative time of the lamp being turned on this time.

3. The edge computing gateway for intelligent scene linkage according to claim 2, characterized in that: The state verification module includes a first processing unit and a first state unit; The first processing unit is used to process scene data and lamp status data; The method includes performing data cleaning and standardization processing on the lamp status data, and encoding the scene data and the lamp status data after the standardization processing into a feature vector; The first state unit is configured with a first state model for determining whether the lamp has an abnormal switch state; The first state model is a trained support vector machine, the input of which includes space occupancy status, date label, time point, and switch status, and the output is a judgment result of whether the switch status of the lamp is abnormal.

4. The edge computing gateway for intelligent scene linkage according to claim 3, characterized in that: The state verification module is configured with a state verification strategy, which is as follows: when the switch state of the lamp is abnormal, the first state unit sends a first prediction instruction to the fault prediction module, and the fault prediction module responds to the first prediction instruction and returns the probability of electrical connection failure and control circuit failure; the first state unit is configured with a first fault probability threshold. If the probability of electrical connection failure and control circuit failure is lower than the first fault probability threshold, a prompt of abnormal switch state is sent to the Internet of Things platform through the second communication unit; if the Internet of Things platform returns information for adjusting the switch state, the first state unit sends a control instruction for adjusting the switch state to the control module, and the control module responds to the control instruction and generates a control signal to adjust the switch state; if at least one of the probabilities of electrical connection failure and control circuit failure is not lower than the first fault probability threshold, a corresponding fault prompt is sent to the Internet of Things platform through the second communication unit.

5. The edge computing gateway for intelligent scene linkage according to claim 4, characterized in that: The state verification module also includes a second state unit, which is configured with a second state model for determining whether the lamp has brightness abnormalities; the second state model is a trained support vector machine, whose input includes space occupancy status, date label, time point, and light brightness, and whose output is a determination result of whether the lamp has brightness abnormalities.

6. The edge computing gateway for intelligent scene linkage according to claim 5, characterized in that: The state verification strategy also includes: when the brightness of the lamp is abnormal, the second state unit sends a second prediction instruction to the fault prediction module, and the fault prediction module responds to the second prediction instruction and returns the probability of lamp aging or light source failure; the second state unit is configured with a second fault probability threshold, and if the probability of lamp aging or light source failure is lower than the second fault probability threshold, a brightness abnormality prompt is sent to the Internet of Things platform through the second communication unit; if the Internet of Things platform returns information for adjusting the brightness of the lamp, the second state unit sends a control instruction for adjusting the brightness of the lamp to the control module, and the control module responds to the control instruction and generates a control signal to adjust the brightness of the lamp; if at least one of the probabilities of lamp aging or light source failure is not lower than the second fault probability threshold, a corresponding fault prompt is sent to the Internet of Things platform through the second communication unit.

7. The edge computing gateway for intelligent scene linkage according to claim 6, characterized in that: The fault prediction module includes a second processing unit; The second processing unit is used to process the lamp status data; specifically as follows: Performing data cleaning on the lamp status data; Calculating a brightness attenuation index based on the light brightness, and calculating a power consumption fluctuation index based on the real-time power consumption; The lamp temperature, usage time, turn-on time, brightness decay index, and power consumption fluctuation index are encoded into a feature vector.

8. The edge computing gateway for intelligent scene linkage according to claim 7, characterized in that: The method for calculating the brightness attenuation index is as follows: Extract the light brightness after M consecutive data cleaning, and establish a time series of light brightness with a length of M; calculate the brightness attenuation index based on the time series of light brightness, the formula is as follows: Where D represents the brightness attenuation index; b i represents the i-th element in the time series of the light brightness; b i+1 The i+1th element in the time series representing the brightness of the light; The method for calculating the power consumption fluctuation index is as follows: Extract the real-time power consumption after N consecutive data cleanings, and establish a time series of real-time power consumption with a length of N; perform n sliding interceptions on the time series of real-time power consumption with a time window of fixed length, and calculate the mean of the real-time power consumption of the n sliding interceptions respectively; calculate the power consumption fluctuation index, the formula is as follows: Where F represents the power consumption fluctuation index; p j represents the mean value of the real-time power consumption captured by the jth sliding operation; p k Represents the kth element in the time series of the real-time power consumption.

9. The edge computing gateway for intelligent scene linkage according to claim 8, characterized in that: The fault prediction module also includes a prediction model unit; the prediction model unit is configured with a fault prediction model for predicting lamp faults; the output of the fault prediction model is the probability of occurrence of each type of fault in the lamp, and the fault types include electrical connection failure, control circuit failure, lamp aging, light source failure, and lamp overheating; the input of the fault prediction model includes the characteristic vectors of the lamp temperature, usage time, turn-on time, brightness attenuation index, and power consumption fluctuation index.

10. The edge computing gateway for intelligent scene linkage according to claim 9, characterized in that: The fault prediction module is configured with a prediction auxiliary strategy, which is as follows: the input of the fault prediction model also includes a lamp status code; the lamp status code includes a switch status abnormal code and a brightness abnormal code; the fault prediction module performs lamp fault prediction at a fixed period, and sends the probability of each fault to the Internet of Things platform through the second communication unit, wherein the switch status abnormal code and the brightness abnormal code are both set to 0; if the fault prediction module receives the first prediction instruction, the switch status abnormal code is set to 1; if the fault prediction module receives the second prediction instruction, the brightness abnormal code is set to 1.

Images