A control system, device and medium of an intelligent smoke sensor

Abstract

The application relates to a control system, equipment and medium of an intelligent smoke sensor, and belongs to the field of intelligent smoke sensors; the application controls the intelligent smoke sensor to issue a buzzing alarm by setting a high and fixed second smoke concentration threshold value, a second temperature rising rate threshold value and a low and dynamically adjustable first smoke concentration threshold value and a first temperature rising rate threshold value; when the environment is disturbed by normal influencing factors such as cooking, steam and dust, the first smoke concentration threshold value and the first temperature rising rate threshold value are dynamically updated to avoid false positives; when the current smoke concentration reaches the updated first smoke concentration threshold value or the current temperature rising rate reaches the updated first temperature rising rate threshold value, the alarm is triggered, early response is realized, and response lag is avoided; and when the current smoke concentration reaches the second smoke concentration threshold value or the current temperature rising rate reaches the second temperature rising rate threshold value, the alarm is triggered, and no alarm delay is ensured in an extreme case.

Description

Technical Field

[0001] This invention relates to the field of intelligent smoke detectors, and more specifically to a control system, device, and medium for an intelligent smoke detector. Background Technology

[0002] With the development of IoT technology, smart smoke detectors can not only detect smoke concentration and temperature, but also connect to the control system via wireless network and upload data to the control system. The existing control system of smart smoke detectors often controls the smart smoke detector to issue an alarm by using a preset fixed threshold corresponding to smoke concentration and a preset fixed threshold corresponding to temperature rise rate. That is, it receives smoke concentration and temperature data collected by the smart smoke detector in real time, and obtains the temperature rise rate in real time based on the temperature data continuously collected by the smart smoke detector. When the smoke concentration reaches the preset smoke concentration threshold or the temperature rise rate reaches the preset temperature rise rate threshold, it controls the smart smoke detector to issue an alarm.

[0003] However, when the preset fixed threshold is set too high, an alarm will only be issued after an abnormal situation such as a fire occurs, resulting in a delayed response, reduced emergency response time, and inability to provide early warning. When the preset fixed threshold is set too low, it is easily affected by normal influencing factors such as cooking, steam, and dust, resulting in frequent false alarms and a high false alarm rate. Therefore, controlling the alarm of a smart smoke detector based on a preset fixed threshold is prone to delayed response or a high false alarm rate. Summary of the Invention

[0004] To achieve the objectives of this invention, the technical solution adopted is as follows: a control system for an intelligent smoke detector, comprising a processor and a memory storing a computer program, wherein when the computer program is executed by the processor, the following steps are implemented:

[0005] S1. When A≥a2 or B≥b2, the intelligent smoke sensor is controlled to issue a buzzer alarm; otherwise, proceed to step S2; A is the current smoke concentration; a2 is the second smoke concentration threshold; B is the current temperature rise rate; b2 is the second temperature rise rate threshold; wherein, the control system receives the smoke concentration and temperature data collected by the intelligent smoke sensor in real time, and obtains the temperature rise rate in real time based on the temperature data continuously collected by the intelligent smoke sensor.

[0006] S2. When A≥a1 or B≥b1, if the threshold adjustment state is in the first state, then obtain C; if the threshold adjustment state is in the second state, then control the smart smoke sensor to issue a buzzer alarm; C is the time-series data of smoke concentration and temperature rise rate within the key time period with the current time point as the end time point; a1 is the first smoke concentration threshold; b1 is the first temperature rise rate threshold; the first state indicates that the values ​​corresponding to a1 and b1 are the initial values; the second state indicates that the values ​​corresponding to a1 and b1 are the updated values; b2>b1; a2>a1.

[0007] S3. Obtain the vector similarity between T and each preset feature vector in the preset feature vector set; T is the key feature vector, which is obtained by feature extraction from C.

[0008] S4. If the maximum value among all vector similarities corresponding to T is not less than the preset similarity threshold, then a1 and b1 are updated based on the preset smoke concentration adjustment value and preset temperature rise rate adjustment value corresponding to the preset feature vector corresponding to the maximum value, and the threshold adjustment state is adjusted to the second state.

[0009] A non-transitory computer-readable storage medium storing a computer program, which is loaded and executed by a processor to implement the steps performed by the computer program in the control system of the aforementioned intelligent smoke detector when executed by the processor.

[0010] An electronic device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps performed by the computer program in the control system of the aforementioned intelligent smoke detector when executed by the processor.

[0011] Compared with the prior art, the beneficial effects of the present invention are:

[0012] 1. A smart smoke detector is controlled to issue a buzzer alarm by setting a higher and fixed second smoke concentration threshold and a second temperature rise rate threshold, and a lower and dynamically adjustable first smoke concentration threshold and a first temperature rise rate threshold. Specifically, when the current smoke concentration reaches the second smoke concentration threshold or the current temperature rise rate reaches the second temperature rise rate threshold, the smart smoke detector is controlled to issue a buzzer alarm. Otherwise, when the current smoke concentration reaches the first smoke concentration threshold or the current temperature rise rate reaches the first temperature rise rate threshold, and the threshold adjustment state is adjusted to the first state, the first smoke concentration threshold and the first temperature rise rate threshold are updated by analyzing the time-series data of smoke concentration and temperature rise rate within a key time period, and the threshold adjustment state is adjusted to the second state. When the current smoke concentration reaches the first smoke concentration threshold or the current temperature rise rate reaches the first temperature rise rate threshold, and the threshold adjustment state is adjusted to the second state, the smart smoke detector is controlled to issue a buzzer alarm. When there are normal influencing factors in the environment such as cooking, steam, and dust, the system can dynamically update the first smoke concentration threshold and the first temperature rise rate threshold based on a preset feature vector to avoid false alarms. When the current smoke concentration reaches the updated first smoke concentration threshold or the current temperature rise rate reaches the updated first temperature rise rate threshold, an alarm is triggered to achieve early response and avoid response lag. Furthermore, when the current smoke concentration reaches the second smoke concentration threshold or the current temperature rise rate reaches the second temperature rise rate threshold, an alarm is triggered to ensure no delay in alarms under extreme conditions. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] like Figure 1 As shown, the present invention provides a technical solution: a control system for an intelligent smoke detector, the control system including a processor and a memory storing a computer program, wherein when the computer program is executed by the processor, the following steps are implemented:

[0016] S1. When A≥a2 or B≥b2, the intelligent smoke sensor is controlled to issue a buzzer alarm; otherwise, proceed to step S2; A is the current smoke concentration; a2 is the second smoke concentration threshold; B is the current temperature rise rate; b2 is the second temperature rise rate threshold; wherein, the control system receives the smoke concentration and temperature data collected by the intelligent smoke sensor in real time, and obtains the temperature rise rate in real time based on the temperature data continuously collected by the intelligent smoke sensor.

[0017] Specifically, the smart smoke detector is equipped with a smoke sensor, a temperature sensor, and a buzzer; the smoke sensor is used to obtain smoke concentration, the temperature sensor is used to obtain temperature data, and the buzzer is used to issue a buzzing alarm when a warning is triggered.

[0018] Furthermore, the intelligent smoke detector is connected to the control system.

[0019] Specifically, the second smoke concentration threshold and the second temperature rise rate threshold are both thresholds preset by those skilled in the art according to actual needs. For example, the second smoke concentration threshold is 5%obs / m or 250mg / m. 3 The second temperature rise rate threshold is 10℃ / min, which will not be elaborated here.

[0020] S2. When A≥a1 or B≥b1, if the threshold adjustment state is in the first state, then obtain C; if the threshold adjustment state is in the second state, then control the smart smoke sensor to issue a buzzer alarm; C is the time-series data of smoke concentration and temperature rise rate within the key time period with the current time point as the end time point; a1 is the first smoke concentration threshold; b1 is the first temperature rise rate threshold; the first state indicates that the values ​​corresponding to a1 and b1 are the initial values; the second state indicates that the values ​​corresponding to a1 and b1 are the updated values; b2>b1; a2>a1.

[0021] Specifically, the initial values ​​corresponding to the first smoke concentration threshold and the first temperature rise rate threshold are values ​​preset by those skilled in the art based on actual needs, as well as the second smoke concentration threshold and the second temperature rise rate threshold. For example, the initial value corresponding to the first smoke concentration threshold is 2.5%obs / m or 100mg / m. 3 The initial value corresponding to the first temperature rise rate threshold is 6℃ / min. The units of measurement for the first smoke concentration threshold and the second smoke concentration threshold are the same, and will not be elaborated here.

[0022] Specifically, the values ​​corresponding to a1 and b1 are initial values, which can be understood as: the current value of a1 is the initial value corresponding to a1, and the current value of b1 is the initial value corresponding to b1.

[0023] Furthermore, the values ​​corresponding to a1 and b1 are the updated values, which can be understood as: the current value of a1 is the new value obtained after the update, and the current value of b1 is the new value obtained after the update.

[0024] Specifically, the control system sends a warning trigger command to the smart smoke detector to cause the smart smoke detector to sound an alarm; when the smart smoke detector receives the warning trigger command, it determines to trigger the warning and causes the buzzer to sound an alarm.

[0025] Specifically, the duration of the critical time period is set by those skilled in the art according to actual needs, for example, 5 minutes, which will not be elaborated here.

[0026] Specifically, C includes a smoke concentration sequence and a temperature rise rate sequence; wherein, the smoke concentration sequence consists of multiple smoke concentrations sorted from early to late according to the acquisition time; the temperature rise rate sequence consists of multiple temperature rise rates sorted from early to late according to the acquisition time.

[0027] S3. Obtain the vector similarity between T and each preset feature vector in the preset feature vector set; T is the key feature vector, which is obtained by feature extraction from C.

[0028] Specifically, the vector similarity is not less than 0 and not greater than 1. The closer the vector similarity is to 1, the more similar its corresponding preset feature vector is to T.

[0029] In a specific embodiment, the vector distance between T and a preset feature vector is converted into a value between 0 and 1, and this value is used as the vector similarity between T and the preset feature vector. The larger the vector distance, the closer the converted value between 0 and 1 is to 0; the smaller the vector distance, the closer the converted value between 0 and 1 is to 1. Those skilled in the art will understand that any prior art method for converting vector distance into a value between 0 and 1, such that the larger the vector distance, the closer the converted value between 0 and 1 is to 0, and the smaller the vector distance, the closer the converted value between 0 and 1 is to 1, falls within the protection scope of this invention. Examples include conversion methods based on exponential decay or reciprocal transformation, which will not be elaborated upon here.

[0030] Optionally, the vector distance is the Euclidean distance.

[0031] Specifically, after step S2 and before step S3, the following steps S01-S03 are also included:

[0032] S01. Extract features from the smoke concentration sequence in C to obtain the first feature vector.

[0033] Optionally, the smoke concentration sequence in C can be feature-extracted using the TSFresh tool to generate a fixed-dimensional feature vector, which can then be used as the first feature vector.

[0034] S02. Extract features from the temperature rise rate sequence in C to obtain the second feature vector.

[0035] Optionally, the temperature rise rate sequence in C can be feature-extracted using the TSFresh tool to generate a fixed-dimensional feature vector, which can then be used as a second feature vector.

[0036] S03. Concatenate the first feature vector with the second feature vector to obtain T.

[0037] Specifically, the vector concatenation involves joining two vectors end-to-end to form a new vector whose dimension is the sum of the dimensions of the two vectors. For example, if the dimension of the first feature vector is d1 and the dimension of the second feature vector is d2, then the dimension of T is d3 = d1 + d2.

[0038] Through the above steps, feature extraction is performed on the smoke concentration sequence and the temperature rise rate sequence to obtain the first feature vector and the second feature vector, and the first feature vector and the second feature vector are concatenated into the key feature vector so that the key feature vector can comprehensively reflect the multi-dimensional temporal characteristics of smoke concentration and temperature rise rate in the key time period.

[0039] Specifically, the preset feature vector set includes several preset feature vectors, each of which has a corresponding preset smoke concentration adjustment value and preset temperature rise rate adjustment value; wherein, the vector dimension of T is consistent with the vector dimension of the preset feature vector.

[0040] Specifically, each preset feature vector corresponds to a specific interference environment, which is an environment that is likely to cause false alarms in smart smoke detectors but does not actually have abnormal situations such as fires, such as cooking, steam, or dust environments.

[0041] Specifically, the unit of measurement for the preset smoke concentration adjustment value is consistent with the unit of measurement for the first smoke concentration threshold and the second smoke concentration threshold; the unit of measurement for the preset temperature rise rate adjustment value is consistent with the unit of measurement for the first temperature rise rate threshold and the second temperature rise rate threshold.

[0042] Furthermore, the preset smoke concentration corresponding to each preset feature vector is not less than 0 and not greater than the difference between a2 and a1; the preset temperature rise rate adjustment value corresponding to each preset feature vector is not less than 0 and not greater than the difference between b2 and b1.

[0043] S4. If the maximum value among all vector similarities corresponding to T is not less than the preset similarity threshold, then a1 and b1 are updated based on the preset smoke concentration adjustment value and preset temperature rise rate adjustment value corresponding to the preset feature vector corresponding to the maximum value, and the threshold adjustment state is adjusted to the second state.

[0044] Specifically, the threshold adjustment state is initially set to the first state.

[0045] Specifically, the preset similarity threshold ranges from 0.7 to 0.9; preferably, the preset similarity threshold is 0.8.

[0046] Specifically, step S4 includes the following steps:

[0047] S41. Add a1 to the preset smoke concentration adjustment value corresponding to the preset feature vector corresponding to the maximum value, so as to update a1.

[0048] S42. Add b1 to the preset temperature rise rate adjustment value corresponding to the preset feature vector corresponding to the maximum value, so as to update b1.

[0049] Through the above steps, if the maximum value among all vector similarities corresponding to the key feature vector is not less than a preset similarity threshold, it can be determined that the current environment is highly similar to the specific interference environment corresponding to the preset feature vector corresponding to the maximum value. The specific interference environment is an environment that is prone to causing false alarms of the smart smoke detector but does not actually have abnormal situations such as fires. It contains interference from normal influencing factors such as cooking, steam, and dust. Therefore, based on the preset smoke concentration adjustment value and preset temperature rise rate adjustment value corresponding to the preset feature vector, the first smoke concentration threshold and the first temperature rise rate threshold are increased to update the first smoke concentration threshold and the first temperature rise rate threshold. At the same time, the threshold adjustment state is adjusted to the second state. This can filter the interference from normal influencing factors such as cooking, steam, and dust, and reduce false alarms. At the same time, under the premise that the first smoke concentration threshold and the first temperature rise rate threshold have been updated, if the subsequent smoke concentration reaches the first smoke concentration threshold or the temperature rise rate reaches the first temperature rise rate threshold, it is regarded as an early signal of abnormal situations such as fires, triggering an alarm, thereby achieving early response and avoiding response lag caused by waiting for high threshold triggers.

[0050] Furthermore, by setting a high and fixed second smoke concentration threshold and a second temperature rise rate threshold, and a low and dynamically adjustable first smoke concentration threshold and a first temperature rise rate threshold, the intelligent smoke sensor is controlled to issue a buzzer alarm. When interference from normal influencing factors such as cooking, steam, and dust exists in the environment, the first smoke concentration threshold and the first temperature rise rate threshold are dynamically updated based on a preset feature vector to avoid false alarms. An alarm is triggered when the current smoke concentration reaches the updated first smoke concentration threshold or the current temperature rise rate reaches the updated first temperature rise rate threshold, achieving early response and avoiding response lag. Furthermore, an alarm is triggered when the current smoke concentration reaches the second smoke concentration threshold or the current temperature rise rate reaches the second temperature rise rate threshold. This ensures that during the threshold adjustment state (first state) and the threshold update process, if the smoke concentration reaches the second smoke concentration threshold or the temperature rise rate reaches the second temperature rise rate threshold, the intelligent smoke sensor is directly controlled to issue a buzzer alarm. This ensures that in the event of a sudden high-risk fire or other abnormal situation, the algorithm processing delay does not affect the millisecond-level emergency response, ensuring no delayed alarm in extreme cases.

[0051] Specifically, the following steps are included after step S4:

[0052] When the duration of the second state reaches the preset duration, a1 is restored to its initial value, b1 is restored to its initial value, and the threshold adjustment state is restored to the first state. The preset duration is a duration that is preset by those skilled in the art according to actual needs, such as 3 minutes or 5 minutes, which will not be elaborated here.

[0053] Optionally, when the threshold adjustment state is adjusted to the second state, the control timer starts timing to obtain the duration of the second state; wherein, the control system is connected to the timer.

[0054] Furthermore, when the threshold adjustment state is restored to the first state, the timer is controlled to stop counting and the duration recorded by the timer is reset to 0.

[0055] Through the above steps, when the duration of the second state reaches the preset duration, the first smoke concentration threshold is restored to its corresponding initial value, the first temperature rise rate is restored to its corresponding initial value, and the threshold adjustment state is restored to the first state. Through the timed reset mechanism, while effectively reducing false alarms, the system always maintains a reliable detection capability for abnormal situations such as fire hazards, thereby balancing anti-interference capability and the safety of long-term operation of the control system.

[0056] Specifically, the following steps are included after step S4:

[0057] S10. If the maximum value of the similarity among all vectors corresponding to T is less than the preset similarity threshold, then proceed to step S20.

[0058] S20. Obtain time-series data D collected by other sensors deployed in the target area corresponding to the smart smoke detector during the key time period corresponding to C.

[0059] Specifically, the target area corresponding to the smart smoke detector is the physical area covered by the detection range of the smart smoke detector.

[0060] Furthermore, other sensors deployed within the target area corresponding to the smart smoke sensor are sensors outside the smart smoke sensor within that target area, such as humidity sensors, carbon monoxide sensors, carbon dioxide sensors, infrared imaging sensors, etc., which will not be elaborated here.

[0061] Specifically, D includes multiple data lists, each corresponding to another sensor deployed in the target area of ​​the smart smoke detector, and the data in the list is sorted from earliest to latest according to the collection time.

[0062] S30. Based on C, D and anomaly detection models, determine whether there are any anomalies in the target area corresponding to the smart smoke detector.

[0063] Specifically, the anomaly detection model is used to detect whether there are abnormal situations such as fires within the target area.

[0064] Specifically, step S30 includes the following steps S31-S36:

[0065] S31. Extract features from the smoke concentration sequence in C to obtain the first feature vector.

[0066] S32. Extract features from the temperature rise rate sequence in C to obtain the second feature vector.

[0067] S33. Perform feature extraction on each data list in D to obtain the third feature vector corresponding to each data list in D.

[0068] Optionally, feature extraction can be performed on the data list in D using the TSFresh tool to generate a fixed-dimensional feature vector, which can then be used as the third feature vector.

[0069] S34. Concatenate the first feature vector, the second feature vector, and each third feature vector in D to obtain the fused feature vector.

[0070] S35. Input the fused feature vector into the anomaly detection model to obtain the anomaly judgment label output by the anomaly detection model.

[0071] S36. When the anomaly judgment label is the first label, it is determined that there is no anomaly in the target area corresponding to the smart smoke detector; when the anomaly judgment label is the second label, it is determined that there is an anomaly in the target area corresponding to the smart smoke detector.

[0072] Through the above steps, feature extraction is performed on the time-series data collected by the smart smoke detector and the time-series data collected by other sensors deployed in the target area corresponding to the smart smoke detector within the same time period, obtaining feature vectors of multiple modalities. The feature vectors of the multiple modalities are concatenated to form a fused feature vector, which is then input into the anomaly detection model to determine whether there are any anomalies in the target area corresponding to the smart smoke detector. By fusing multi-source heterogeneous sensing information, the fused feature vector can comprehensively reflect multi-dimensional dynamic features such as smoke, temperature, and gas concentration. Anomaly identification is performed based on the fused feature vector and the anomaly detection model, significantly improving the accuracy of anomaly identification.

[0073] Specifically, the first label indicates that there are no abnormal situations such as fires in the target area; the second label indicates that there are abnormal situations such as fires in the target area.

[0074] Specifically, the anomaly detection model is obtained by supervised training of the initial detection model using a training sample set. The initial detection model is a binary classification model, such as decision tree, random forest, or SVM, which will not be elaborated here.

[0075] Specifically, the training sample set includes several training samples, each of which includes a preset fusion feature vector and manually labeled anomaly judgment labels corresponding to the preset fusion feature vector.

[0076] Furthermore, the preset fusion feature vector is obtained based on real fire experiment data, that is: after deploying smart smoke detectors and all other sensors of the same model as those in the target area in the laboratory or test site; simulate real fire scenarios (such as burning cotton cloth or wood) and scenarios that are likely to cause false alarms of smart smoke detectors but do not actually involve fire (such as cooking, steam, or dust) in the laboratory or test site, and simultaneously acquire time-series data of smoke concentration and temperature rise rate within a fixed time window, as well as time-series data collected by other sensors within the fixed time window; using the same method as in steps S31-S35 to obtain the fusion feature vector based on C and D, the preset fusion feature vector is obtained based on the time-series data of smoke concentration and temperature rise rate acquired within the fixed time window, as well as the time-series data collected by other sensors within the fixed time window.

[0077] Specifically, the length of the fixed time window is consistent with the duration of the key time period.

[0078] S40. When an anomaly is detected in the target area corresponding to the smart smoke detector, control the smart smoke detector to issue a buzzer alarm.

[0079] Through the above steps, when the maximum value of the similarity among all vectors corresponding to the key feature vector is less than the preset similarity threshold, it can be determined that the current environment is not similar to the specific interference environment corresponding to any preset feature vector. At this time, it is necessary to determine whether the current environment is in the early stage of an abnormal situation such as a fire. Therefore, the time series data collected by other sensors deployed in the target area corresponding to the smart smoke detector during the key time period is obtained. Based on the time series data, the time series data corresponding to the key feature vector, and the anomaly detection model, it is determined whether there is an anomaly in the target area corresponding to the smart smoke detector. When it is determined that there is an anomaly in the target area corresponding to the smart smoke detector, it indicates that there is an abnormal situation such as a fire in the target area. At this time, the current environment is in the early stage of an abnormal situation such as a fire. Therefore, the smart smoke detector is controlled to issue a buzzer alarm to achieve early response and avoid response lag.

[0080] Specifically, step S40 also includes: when it is determined that there is no abnormality in the target area corresponding to the smart smoke detector, proceed to step S50.

[0081] S50. When A≥a1 and B<b1, if the threshold adjustment state is not the third state, let a1=a1+△a to update a1 and adjust the threshold adjustment state to the third state; if the threshold adjustment state is the third state, let a1=a1+△a to update a1.

[0082] Specifically, △a is a preset minimum smoke concentration increment value. The specific value of △a is set by those skilled in the art according to actual needs, for example: 1%obs / m³ or 20mg / m³. 3 This will not be elaborated upon here.

[0083] Specifically, the unit of measurement for the preset minimum smoke concentration increment is consistent with the unit of measurement for the first smoke concentration threshold and the second smoke concentration threshold.

[0084] S60. When A < a1 and B ≥ b1, if the threshold adjustment state is not the third state, let b1 = b1 + Δb to update b1 and adjust the threshold adjustment state to the third state; if the threshold adjustment state is the third state, let b1 = b1 + Δb to update b1.

[0085] Specifically, △b is the preset minimum temperature rise rate increment value. The specific value of △b is set by those skilled in the art according to actual needs, for example: 2℃ / min, which will not be elaborated here.

[0086] Specifically, the unit of measurement for the preset minimum temperature rise rate increment is consistent with the units of measurement for the first temperature rise rate threshold and the second temperature rise rate threshold.

[0087] S70. When A≥a1 and B≥b1, if the threshold adjustment state is not the third state, let a1=a1+△a to update a1, let b1=b1+△b to update b1, and adjust the threshold adjustment state to the third state; if the threshold adjustment state is the third state, let a1=a1+△a to update a1, and let b1=b1+△b to update b1.

[0088] Specifically, step S2 also includes:

[0089] If the threshold adjustment state is in the third state, then obtain C and D and proceed to step S30.

[0090] Specifically, after step S2, the following is also included:

[0091] No warning is triggered when A < a1 and B < b1.

[0092] Through the above steps, when it is determined that there are no abnormalities in the target area corresponding to the smart smoke detector, it indicates that there are no abnormal situations such as fires in the target area. At this time, the current environment is affected by other normal interference factors, which leads to an increase in smoke concentration or temperature rise rate. The first smoke concentration threshold and the first temperature rise rate threshold are updated by the preset minimum smoke concentration increment value and the preset minimum temperature rise rate increment value, and the threshold adjustment state is adjusted to the third state, indicating that it is no longer possible to determine whether there are other normal interference factors in the current environment based on the preset feature vector set. Therefore, if the smoke concentration subsequently reaches the first smoke concentration threshold or the temperature rise rate reaches the first temperature rise rate threshold, the time series data collected by the smart smoke detector and other sensors corresponding to the smart smoke detector within the key time period with the current time point as the end time point is obtained. Based on the time series data and the anomaly detection model, it is determined whether there are abnormalities in the target area corresponding to the smart smoke detector. When it is determined that there are abnormalities in the target area corresponding to the smart smoke detector, the smart smoke detector is controlled to issue a buzzer alarm. When it is determined that there are no abnormalities in the target area corresponding to the smart smoke detector, the first smoke concentration threshold and the first temperature rise rate threshold are updated again. The closed-loop mechanism can effectively reduce false alarms caused by interference factors, and can trigger early warnings in the early stages of abnormal situations such as fires, enabling early response and avoiding response delays, thus balancing low false alarm rate and timely response capability.

[0093] Specifically, when the threshold adjustment state is adjusted to the third state, the key feature vector corresponding to the most recently obtained time series data of smoke concentration and temperature rise rate is used as the target feature vector corresponding to the third state.

[0094] Furthermore, when the threshold adjustment state is in the third state and a1 or b1 is updated, the key feature vector corresponding to the most recently obtained time series data of smoke concentration and temperature rise rate is used as the target feature vector corresponding to the third state, so as to update the target feature vector corresponding to the third state.

[0095] Specifically, after step S30, the following is also included:

[0096] When an anomaly is detected in the target area corresponding to the smart smoke sensor, if the threshold adjustment state is the third state, the target feature vector corresponding to the third state is inserted into the preset feature vector set as a new preset feature vector, and the difference between a1 and the initial value corresponding to a1 is used as the preset smoke concentration adjustment value corresponding to the new preset feature vector, and the difference between b1 and the initial value corresponding to b1 is used as the preset temperature rise rate adjustment value corresponding to the new preset feature vector.

[0097] Through the above steps, in the third state, the system caches the key feature vectors corresponding to the current smoke concentration and temperature rise rate time series data as target feature vectors. If an anomaly is subsequently determined to exist within the target area, and the threshold adjustment state is still in the third state, it indicates that the target feature vector actually reflects the early evolution pattern of anomalies such as fires. If no anomaly is subsequently determined to exist within the target area, and the threshold adjustment state is still in the third state, it indicates that the target feature vector actually reflects the influence of other normal interference factors in the environment, leading to an increase in smoke concentration or temperature rise rate. Therefore, when an anomaly is determined to exist within the target area, and the threshold adjustment state is in the third state, the target feature vector is added to the preset feature vector set as a new preset feature vector, and the current first... The difference between the smoke concentration threshold and its corresponding initial value, and the difference between the first temperature rise rate threshold and its corresponding initial value, are stored together as preset smoke concentration adjustment values ​​and preset temperature rise rate adjustment values ​​corresponding to the preset feature vector. This enables data reuse and automatic expansion of the preset feature vector set. Consequently, when a key feature vector highly similar to the preset feature vector is detected again, the first smoke concentration threshold and the first temperature rise rate threshold can be directly adjusted through similarity matching. When the smoke concentration threshold or the temperature rise rate reaches the first temperature rise rate threshold, an alarm is directly triggered without repeating multi-sensor data acquisition and complex model inference, thereby significantly improving processing efficiency and enabling early warning of such anomalies.

[0098] Specifically, the preset feature vector set contains multiple preset feature vectors during initialization.

[0099] Furthermore, for each preset feature vector included in the preset feature vector set during initialization, the preset feature vector is obtained by extracting features from the time-series data of smoke concentration and temperature rise rate collected within a fixed time period under the corresponding specific interference environment; the duration of the fixed time period is consistent with the duration of the key time period.

[0100] Furthermore, the preset smoke concentration adjustment value corresponding to the preset feature vector is used to represent the difference between the smoke concentration corresponding to the early stage of an abnormal situation such as a fire and the smoke concentration corresponding to the specific interference environment corresponding to the preset feature vector; the preset temperature rise rate adjustment value corresponding to the preset feature vector is used to represent the difference between the temperature rise rate corresponding to the early stage of an abnormal situation such as a fire and the temperature rise rate corresponding to the specific interference environment corresponding to the preset feature vector; for each preset feature vector included in the preset feature vector set during initialization, the preset smoke concentration adjustment value and the preset temperature rise rate adjustment value corresponding to the preset feature vector are preset by those skilled in the art.

[0101] Specifically, when the threshold adjustment state is in the third state, if a1 or b1 is updated, the duration of the third state is reset to 0.

[0102] Furthermore, when the duration of the third state reaches the preset duration, a1 is restored to its initial value, b1 is restored to its initial value, and the threshold adjustment state is restored to the first state.

[0103] Optionally, when the threshold adjustment state is adjusted to the third state, the timer is controlled to start timing to obtain the duration of the third state; wherein, when the threshold adjustment state is the third state and a1 or b1 is updated, the duration recorded by the timer is reset to 0.

[0104] Through the above steps, when the threshold adjustment state is in the third state, if the first smoke concentration threshold or the first temperature rise rate threshold is updated, it indicates that there are still interfering factors in the current environment that cause parameter adjustment. In order to accurately reflect the continuous activity of the interfering factors, the duration of the third state is reset to zero. Furthermore, when the duration of the third state reaches the preset duration, the first smoke concentration threshold and the first temperature rise rate threshold are restored to their corresponding initial values, and the threshold adjustment state is reset to the first state. Through the timed reset mechanism, while effectively reducing false alarms, the reliable detection capability for abnormal situations such as fire hazards is always maintained, thereby taking into account both anti-interference and the long-term safety of the control system.

[0105] Embodiments of the present invention also provide a non-transitory computer-readable storage medium storing a computer program, which is loaded and executed by a processor to implement the steps implemented by the computer program in the control system of the intelligent smoke detector provided in the above embodiments when executed by the processor.

[0106] Embodiments of the present invention also provide an electronic device, including: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps implemented by the computer program in the control system of the intelligent smoke detector provided in the above embodiments when executed by the processor.

[0107] The embodiments disclosed herein are preferred embodiments, but are not limited thereto. Those skilled in the art can readily grasp the spirit of the present invention based on the above embodiments and make different extensions and variations, but as long as they do not depart from the spirit of the present invention, they are all within the protection scope of the present invention.

Claims

1. A control system for an intelligent smoke detector, characterized in that, The control system includes a processor and a memory storing a computer program. When the computer program is executed by the processor, the following steps are performed: S1. When A≥a2 or B≥b2, the intelligent smoke sensor is controlled to issue a buzzer alarm; otherwise, proceed to step S2; A is the current smoke concentration; a2 is the second smoke concentration threshold; B is the current temperature rise rate; b2 is the second temperature rise rate threshold; wherein, the control system receives the smoke concentration and temperature data collected by the intelligent smoke sensor in real time, and obtains the temperature rise rate in real time based on the temperature data continuously collected by the intelligent smoke sensor. S2. When A≥a1 or B≥b1, if the threshold adjustment state is in the first state, then obtain C; if the threshold adjustment state is in the second state, then control the smart smoke sensor to issue a buzzer alarm; C is the time-series data of smoke concentration and temperature rise rate within the key time period with the current time point as the end time point; a1 is the first smoke concentration threshold; b1 is the first temperature rise rate threshold; the first state indicates that the values ​​corresponding to a1 and b1 are the initial values; the second state indicates that the values ​​corresponding to a1 and b1 are the updated values; b2>b1; a2>a1; S3. Obtain the vector similarity between T and each preset feature vector in the preset feature vector set; T is the key feature vector, which is obtained by feature extraction from C. S4. If the maximum value among all vector similarities corresponding to T is not less than a preset similarity threshold, then a1 and b1 are updated based on the preset smoke concentration adjustment value and preset temperature rise rate adjustment value corresponding to the preset feature vector corresponding to the maximum value, and the threshold adjustment state is adjusted to the second state; including: S41. Add a1 to the preset smoke concentration adjustment value corresponding to the preset feature vector corresponding to the maximum value, so as to update a1. S42. Add b1 to the preset temperature rise rate adjustment value corresponding to the preset feature vector corresponding to the maximum value, so as to update b1. The following steps are included after step S4: S10. If the maximum value of the similarity among all vectors corresponding to T is less than the preset similarity threshold, then proceed to step S20. S20. Obtain time-series data D collected by other sensors deployed in the target area corresponding to the smart smoke detector during the key time period corresponding to C; S30. Based on C, D and anomaly detection models, determine whether there are any anomalies in the target area corresponding to the smart smoke detector; S40. When an anomaly is detected in the target area corresponding to the smart smoke detector, control the smart smoke detector to issue a buzzer alarm.

2. The control system for the intelligent smoke detector according to claim 1, characterized in that, C includes a smoke concentration sequence and a temperature rise rate sequence; wherein, the smoke concentration sequence consists of multiple smoke concentrations sorted from early to late according to the acquisition time; the temperature rise rate sequence consists of multiple temperature rise rates sorted from early to late according to the acquisition time.

3. The control system for the intelligent smoke detector according to claim 2, characterized in that, After step S2 and before step S3, the following steps are also included: S01. Extract features from the smoke concentration sequence in C to obtain the first feature vector; S02. Extract features from the temperature rise rate sequence in C to obtain the second feature vector; S03. Concatenate the first feature vector with the second feature vector to obtain T.

4. The control system for the intelligent smoke detector according to claim 1, characterized in that, The preset feature vector set includes several preset feature vectors, each of which has a corresponding preset smoke concentration adjustment value and preset temperature rise rate adjustment value; wherein, the vector dimension of T is consistent with the vector dimension of the preset feature vector.

5. The control system for the intelligent smoke detector according to claim 4, characterized in that, Each preset feature vector corresponds to a preset smoke concentration that is not less than 0 and not greater than the difference between a2 and a1. The preset temperature rise rate adjustment value corresponding to each preset feature vector is not less than 0 and not greater than the difference obtained by subtracting b1 from b2.

6. The control system for the intelligent smoke detector according to claim 1, characterized in that, The process after step S4 also includes: When the duration of the second state reaches the preset duration, a1 is restored to its initial value, b1 is restored to its initial value, and the threshold adjustment state is restored to the first state.

7. A non-transitory computer-readable storage medium, characterized in that, The storage medium stores a computer program, which is loaded and executed by a processor to implement the steps performed by the computer program in the control system of the intelligent smoke detector as described in any one of claims 1-6 when executed by the processor.

8. An electronic device, comprising: A processor, a memory, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the computer program, it implements the steps implemented when the computer program in the control system of the intelligent smoke detector as described in any one of claims 1-6 is executed by the processor.

Images