Adaptive Alarm Triggering Method and Device Based on Environment Perception

Abstract

The present invention discloses an adaptive alarm triggering method and device based on environmental perception, belonging to the field of forest fire fighting. The method includes: collecting distributed perception data to obtain multi-source positioning data; establishing an environmental perception model to perform environmental perception on the distributed perception data to determine the basic fire danger index; using an adaptive Kalman filtering algorithm to perform fusion processing on the multi-source positioning data to obtain target positioning information; correcting the basic fire danger index to obtain a fire danger assessment index; when exceeding a preset fire danger index threshold, sending an alarm message through a long-distance redundant communication mechanism. This application solves the technical problems in the prior art that in the forest fire fighting scenario, due to the complex terrain environment, the positioning communication signal is unstable and it is difficult to search and rescue and communicate when lost and trapped, and achieves the technical effects of realizing highly reliable environmental perception, accurate personnel positioning, and stable dual-channel redundant communication in the complex environment of forest fire fighting.

Description

Technical Field

[0001] The present invention relates to the field of forest fire fighting, and specifically to an adaptive alarm triggering method and device based on environmental perception. Background Art

[0002] With the continuous development of forest fire fighting work, the requirements for the reliability of firefighters' positioning and communication are constantly increasing. At present, the forest fire fighting field mainly uses the Beidou positioning system combined with wireless communication for personnel positioning and calling for help. However, in the actual application of forest fire fighting, due to the influence of complex terrain and environmental conditions, the existing technology has the following problems: First, a single positioning technology is difficult to adapt to complex environments. The existing technology mainly relies on the Beidou positioning system and fails to effectively integrate multi-source positioning data such as air pressure, acceleration, gyroscope, magnetometer, and shoe-mounted inertial navigation, resulting in insufficient positioning accuracy and affecting the safety guarantee of firefighters. Second, the environmental perception ability is insufficient. The existing technology lacks a comprehensive analysis of distributed perception data such as temperature, humidity, light intensity, and smoke concentration in the forest environment and cannot timely and accurately evaluate the degree of fire danger, affecting the timeliness and accuracy of alarms. Then, when firefighters are lost or trapped, a single communication method is prone to communication interruption due to insufficient signal coverage, making it difficult to ensure the reliability of communication. Summary of the Invention

[0003] This application provides an adaptive alarm triggering method and device based on environmental perception, aiming to solve the technical problems of unstable positioning and communication signals caused by complex terrain environments in the forest fire fighting scenario and difficult search and rescue communication when lost or trapped.

[0004] In view of the above problems, this application provides an adaptive alarm triggering method and device based on environmental perception.

[0005] In the first aspect disclosed in this application, an adaptive alarm triggering method based on environmental perception is provided. The method includes: collecting distributed perception data of the forest environment, where the distributed perception data includes temperature, humidity, light intensity, and smoke concentration; obtaining multi-source positioning data, where the multi-source positioning data includes Beidou positioning data, air pressure data, acceleration data, gyroscope data, magnetometer data, and shoe-mounted inertial navigation data; establishing an environmental perception model, using the environmental perception model to perform environmental perception on the distributed perception data to determine the basic fire danger index; using an adaptive Kalman filter algorithm to perform fusion processing on the multi-source positioning data to obtain target positioning information; correcting the basic fire danger index according to the target positioning information to obtain a fire danger assessment index; when the fire danger assessment index exceeds a preset fire danger index threshold, sending an alarm message through a long-distance redundant communication mechanism, and the long-distance redundant communication mechanism includes satellite short message communication and narrowband Internet of Things communication.

[0006] Another aspect disclosed in this application provides an adaptive alarm trigger device based on environmental perception. The device includes: an environmental acquisition module for collecting distributed perception data of the forest environment, where the distributed perception data includes temperature, humidity, light intensity, and smoke concentration; a multi-source positioning module for obtaining multi-source positioning data, where the multi-source positioning data includes Beidou positioning data, barometric pressure data, acceleration data, gyroscope data, magnetometer data, and shoe-mounted inertial navigation data; a perception modeling module for establishing an environmental perception model and using the environmental perception model to perform environmental perception on the distributed perception data to determine a basic fire danger index; a positioning fusion module for using an adaptive Kalman filter algorithm to perform fusion processing on the multi-source positioning data to obtain target positioning information; a danger assessment module for correcting the basic fire danger index according to the target positioning information to obtain a fire danger assessment index; and an alarm communication module for sending an alarm message through a long-distance redundant communication mechanism when the fire danger assessment index exceeds a preset fire danger index threshold, where the long-distance redundant communication mechanism includes satellite short message communication and narrowband Internet of Things communication.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] By collecting distributed perception data of the forest environment, including temperature, humidity, light intensity, and smoke concentration, the forest environment conditions can be comprehensively grasped; at the same time, multi-source positioning data, including Beidou positioning data, barometric pressure data, acceleration data, gyroscope data, magnetometer data, and shoe-mounted inertial navigation data, is obtained, laying a foundation for subsequent precise positioning; an environmental perception model is established and used to perform environmental perception on the distributed perception data to determine a basic fire danger index, realizing a preliminary assessment of the environmental danger level; the adaptive Kalman filter algorithm is used to perform fusion processing on the multi-source positioning data to obtain accurate target positioning information, solving the problem of insufficient accuracy of a single positioning method; on this basis, the basic fire danger index is corrected according to the target positioning information to obtain a more accurate fire danger assessment index, improving the reliability of the danger assessment; when the fire danger assessment index exceeds the preset fire danger index threshold, an alarm message is sent through a long-distance redundant communication mechanism including satellite short message communication and narrowband Internet of Things communication, ensuring the reliability of communication. The technical solution solves the technical problems in the prior art of unstable positioning and communication signals and difficult search and rescue communication when lost and trapped due to complex terrain environments in forest fire scenarios, achieving the technical effects of high-reliability environmental perception, precise personnel positioning, and stable dual-channel redundant communication in complex forest fire environments.

[0009] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application, it can be implemented according to the content of the specification. In order to make the above and other purposes, features, and advantages of this application more obvious and understandable, the following specific embodiments of this application are specifically given. Description of the Drawings

[0010] Figure 1 FIG. is a schematic flowchart of an adaptive alarm triggering method based on environmental perception provided by an embodiment of this application;

[0011] Figure 2 FIG. is a schematic structural diagram of an adaptive alarm triggering device based on environmental perception provided by an embodiment of this application.

[0012] Description of the reference numerals in the drawings: environmental acquisition module 11, multi-source positioning module 12, perception modeling module 13, positioning fusion module 14, risk assessment module 15, alarm communication module 16. Specific Embodiments

[0013] The general idea of the technical solution provided by this application is as follows:

[0014] An embodiment of this application provides an adaptive alarm triggering method and device based on environmental perception. First, by collecting distributed perception data including temperature, humidity, light intensity, and smoke concentration, and at the same time obtaining multi-source positioning data such as Beidou positioning data, barometric pressure data, acceleration data, gyroscope data, magnetometer data, and shoe-mounted inertial navigation data, a comprehensive data acquisition system is constructed. On this basis, an environmental perception model is established to process the distributed perception data to obtain a basic fire risk index; an adaptive Kalman filter algorithm is used to fuse and process the multi-source positioning data to obtain accurate target positioning information; further, the target positioning information is used to correct the basic fire risk index to obtain a more reliable fire risk assessment index. When the assessment index exceeds a preset threshold, an alarm message is sent through a long-distance redundant communication mechanism combining satellite short message communication and narrowband Internet of Things communication.

[0015] Through the above technical solution, the intelligentization of environmental perception, the precision of personnel positioning, and the reliability of communication guarantee are realized in the complex environment of forest fire fighting, effectively solving the technical problems in the prior art of unstable positioning and communication signals due to complex terrain environments and difficult search and rescue communication when getting lost and trapped.

[0016] After introducing the basic principle of this application, the following will specifically introduce various non-limiting embodiments of this application in conjunction with the drawings in the specification.

[0017] Embodiment 1, as Figure 1 shown, an embodiment of this application provides an adaptive alarm triggering method based on environmental perception, and the method includes:

[0018] S1: Collect distributed sensing data of the forest environment, where the distributed sensing data includes temperature, humidity, light intensity, and smoke concentration.

[0019] Specifically, an environment sensing module is integrated in the smart terminal carried by forest firefighters. This environment sensing module includes multiple sensor units for real-time monitoring of the forest environment. The environment sensing module integrates a temperature sensor unit, a humidity sensor unit, a light sensor unit, and a smoke sensor unit. Among them, the temperature sensor unit correspondingly collects the temperature of the forest environment, the humidity sensor unit correspondingly collects the humidity of the forest environment, the light sensor unit correspondingly collects the light intensity of the forest environment, and the smoke sensor unit correspondingly collects the smoke concentration of the forest environment. The temperature of the forest environment is collected in real time through the temperature sensor unit to promptly reflect the temperature anomaly caused by forest fires; the humidity of the forest environment is collected in real time through the humidity sensor unit to reflect the possibility of forest fires; the light intensity of the forest environment is collected in real time through the light sensor unit to assist in judging whether a fire has occurred; the smoke concentration of the forest environment is collected in real time through the smoke sensor unit to directly reflect the occurrence of a fire.

[0020] The data collected by each sensor unit is preprocessed by the signal processing circuit of the environment sensing module to form distributed sensing data. By collecting the above-mentioned distributed sensing data, the real-time status of the forest environment can be comprehensively grasped, providing a data basis for subsequent environment sensing and risk assessment.

[0021] S2: Obtain multi-source positioning data, where the multi-source positioning data includes Beidou positioning data, barometric pressure data, acceleration data, gyroscope data, magnetometer data, and shoe-mounted inertial navigation data.

[0022] Specifically, a positioning module is integrated in the smart terminal carried by forest firefighters. This positioning module integrates multiple positioning sensor units for realizing collaborative positioning of multi-source data. Specifically, this positioning module includes a Beidou positioning unit, a barometric pressure sensing unit, an acceleration sensing unit, a gyroscope unit, a magnetometer unit, and an inertial navigation unit integrated in the special shoes for firefighters. Among them, the Beidou positioning unit obtains absolute position coordinates by receiving Beidou satellite signals; the barometric pressure sensing unit obtains altitude information by detecting changes in atmospheric pressure; the acceleration sensing unit is used to detect motion acceleration and obtain movement status information; the gyroscope unit is used to detect angular velocity and obtain steering information; the magnetometer unit is used to detect the geomagnetic field and obtain azimuth information; the shoe-mounted inertial navigation unit obtains relative displacement information by detecting gait characteristics.

[0023] The data obtained by each positioning sensor unit is preprocessed by the signal processing circuit of the positioning module to form multi-source positioning data. By obtaining the above multi-source positioning data, multi-dimensional positioning information can be provided for subsequent data fusion processing, thereby improving the positioning accuracy and reliability.

[0024] S3: Establish an environmental perception model, and use the environmental perception model to perform environmental perception on the distributed perception data to determine the basic fire hazard index.

[0025] Specifically, the intelligent analysis and processing of distributed perception data is realized by establishing an environmental perception model. First, a standard parameter library for the forest environment is established, which stores benchmark parameters such as the standard temperature range, standard humidity range, standard light intensity range, and standard smoke concentration range. Then, historical environmental data within a preset time period is obtained, and the historical environmental data is processed through feature extraction to obtain environmental characteristic parameters. The environmental characteristic parameters are compared with the benchmark parameters in the standard parameter library to establish an environmental parameter deviation mapping relationship. Based on the above environmental parameter deviation mapping relationship, the data is normalized and weighted calculated according to preset weights to establish the corresponding relationship between the sample environmental comprehensive deviation value and the sample basic fire hazard index, thereby training to form an environmental perception model. This model can convert the distributed perception data collected in real time into the basic fire hazard index representing the current environmental hazard level.

[0026] By establishing and applying the environmental perception model, the intelligent assessment of the forest environmental conditions is realized, providing a basic basis for subsequent hazard assessment.

[0027] S4: Use the adaptive Kalman filtering algorithm to perform fusion processing on the multi-source positioning data to obtain the target positioning information.

[0028] Specifically, in order to improve the positioning accuracy, the adaptive Kalman filtering algorithm is used to perform fusion processing on the multi-source positioning data. First, a state equation and an observation equation are established. Among them, the state equation is used to describe the time evolution relationship of the position and speed of the firefighters, and the observation equation is used to describe the measurement values of each positioning sensor. Based on the above equations, the initial state prediction value is calculated. Then, a measurement residual sequence is constructed according to the actual measurement values and the initial state prediction values of the multi-source positioning data. By performing a sliding window process on the measurement residual sequence, the statistical mean and variance of the window measurement residual sequence are calculated, and a measurement error covariance matrix is constructed, which characterizes the credibility of the measurement data of each sensor. Based on the measurement error covariance matrix, the dynamic weight coefficient is adaptively calculated to adaptively adjust the Kalman gain. Then, the initial state prediction value is corrected using the adjusted Kalman gain to obtain accurate target positioning information.

[0029] By adopting an adaptive Kalman filtering algorithm for multi-source data fusion, the positioning accuracy can be effectively improved, the limitations of single positioning methods can be overcome, and reliable position information can be provided for firefighters.

[0030] S5: Correct the basic fire danger index according to the target positioning information to obtain a fire danger assessment index.

[0031] Specifically, combine the target positioning information with geographical information data to correct the basic fire danger index. First, determine the current area where the firefighters are located through the target positioning information. Based on this area information, retrieve the corresponding geographical information data from the geographical information database, including terrain information, vegetation information, and water source distribution information. The terrain information reflects the undulation degree and passage difficulty of the terrain, the vegetation information reflects the distribution of combustibles, and the water source distribution information reflects the availability of rescue resources. Then, calculate the fire fighting difficulty coefficient based on the above geographical information data. The fire fighting difficulty coefficient is a comprehensive evaluation index indicating the difficulty of fire fighting and rescue operations during a forest fire. The fire fighting difficulty coefficient is calculated through four factors: terrain complexity, vegetation density, water source accessibility, and communication intervisibility. Among them, the higher the terrain complexity, the higher the score of the fire fighting difficulty coefficient; the greater the vegetation density, the higher the score of the fire fighting difficulty coefficient; the worse the water source accessibility, the higher the score of the fire fighting difficulty coefficient; the worse the communication intervisibility, the higher the score of the fire fighting difficulty coefficient. The fire fighting difficulty coefficient is obtained through weighted calculation of these four factors. The higher this coefficient is, the greater the difficulty of carrying out fire fighting and rescue operations in this area, and more rescue resources need to be invested and more cautious rescue strategies need to be adopted. After that, perform a weighted operation on the fire fighting difficulty coefficient and the basic fire danger index to obtain a more accurate fire danger assessment index. This assessment index not only reflects the danger degree of the environment itself but also takes into account the impact of the geographical environment on fire fighting and rescue, thus realizing a comprehensive assessment of the danger degree.

[0032] By introducing geographical information data for correction, the accuracy and practicality of danger assessment can be improved, providing a more reliable basis for the safety protection of firefighters.

[0033] S6: When the fire danger assessment index exceeds the preset fire danger index threshold, send an alarm message through a long-distance redundant communication mechanism, and the long-distance redundant communication mechanism includes satellite short message communication and narrowband Internet of Things communication.

[0034] Specifically, long-distance redundant communication is achieved by establishing a dual-channel communication link. This communication mechanism includes two independent communication links: satellite short message communication is used for long-distance data transmission, and narrowband Internet of Things (NB-IoT) communication is used for medium-distance data transmission. When the fire danger assessment index exceeds the preset threshold and triggers an alarm, the alarm information is first processed into data packets according to a preset format to generate redundant data packets. Data packetization can improve the reliability of transmission, ensuring the basic integrity of information even if some data packets fail to be transmitted. Then, the redundant data packets are sent in parallel through two independent communication links. Among them, the satellite short message communication link has the advantages of wide coverage and being unaffected by ground facilities, making it suitable for long-distance communication; the narrowband Internet of Things communication link features fast data transmission rate and low power consumption, making it suitable for medium-distance communication. The narrowband Internet of Things communication link is based on NB-IoT technology and is implemented through the 4G / 5G cellular network base stations of mobile operators, and can be deployed relying on the existing mobile communication network infrastructure. Since the NB-IoT operating frequency band is relatively low, it has strong signal penetration ability and can achieve effective coverage even in complex terrains with poor line-of-sight conditions. In areas with weak signal coverage, temporary base stations can also be deployed for supplementation. After that, at the receiving end, data verification and recombination are performed on the received redundant data packets, and the integrity and accuracy of the alarm information are ensured through cross-verification of the dual-channel data.

[0035] By adopting the dual-channel redundant communication mechanism, the communication reliability in the forest fire fighting scenario can be significantly improved, effectively solving the communication problems in complex terrain environments and providing strong support for the safety guarantee of firefighters.

[0036] Furthermore, the embodiments of the present application further include:

[0037] Establish a standard parameter library for the forest environment, where the standard parameter library includes a standard temperature range, a standard humidity range, a standard light intensity range, and a standard smoke concentration range; obtain historical environmental data within a preset time period; extract features from the historical environmental data to obtain environmental characteristic parameters; compare the environmental characteristic parameters with the standard parameter library to establish an environmental parameter deviation mapping relationship; and establish the environmental perception model based on the environmental parameter deviation mapping relationship.

[0038] In a feasible implementation manner, in order to establish an environmental perception model, first, a standard parameter library for the forest environment is established. The standard parameter library stores various reference values reflecting the normal forest environment state, specifically including: a standard temperature range (such as 15°C - 30°C), a standard humidity range (such as 40% - 70%), a standard light intensity range (such as 1000 - 10000 lux), and a standard smoke concentration range (such as 0 - 0.1mg / m 3)。These standard range values are obtained based on the statistical analysis of a large amount of forest environmental data and can be used as a benchmark for judging environmental anomalies. Preferably, when constructing the standard parameter library, since the differences in the forest geographical environment will significantly affect the distribution characteristics of environmental parameters, it is necessary to consider the influence of factors such as the terrain undulation, vegetation coverage, and tree occlusion of the geographical environment on environmental parameters. Among them, the terrain undulation degree is used to characterize the complexity of the terrain and its influence on the temperature and humidity distribution, the vegetation coverage degree is used to characterize the coverage range of vegetation and its influence on light and air circulation, and the tree density is used to characterize the tree distribution per unit area and its comprehensive occlusion effect on environmental parameters. By correcting the standard range values with these geographical environment parameters, the parameter change rules under different geographical environments can be more accurately reflected, and the judgment accuracy of environmental anomalies can be improved. Then, historical environmental data within a preset time period is obtained. Here, the preset time period can be the past year or the past six months, during which environmental data is collected and stored at a fixed sampling frequency (such as once every 10 minutes). These historical data contain the change rules of the forest environment and can reflect the normal fluctuation range of environmental parameters. Next, feature extraction is performed on the historical environmental data. By calculating statistical features (such as mean, standard deviation, extreme values, etc.), time series features (such as change trends, periodicity, etc.), and correlation features (such as the correlation relationship between parameters), environmental feature parameters that can characterize the environmental characteristics are obtained.

[0039] After that, the extracted environmental feature parameters are compared and analyzed with the benchmark values in the standard parameter library. By calculating the deviation degree between the actual parameters and the standard parameters, an environmental parameter deviation mapping relationship is established. This mapping relationship can be used to quantify the degree of environmental anomalies and provide a basis for subsequent risk assessment. Then, based on the established environmental parameter deviation mapping relationship, an environmental perception model is constructed. This model can map the real-time collected environmental data into corresponding risk degree indicators to realize the intelligent assessment of the forest environmental conditions. Through the established environmental perception model, the timely identification and assessment of forest environmental anomalies can be effectively realized.

[0040] Furthermore, the embodiments of the present application further include:

[0041] Perform data normalization processing on the environmental parameter deviation mapping relationship to obtain a normalized environmental parameter deviation; perform weighted calculation on the normalized environmental parameter deviation according to a preset weight to obtain a sample environmental comprehensive deviation value; establish a correspondence between the sample environmental comprehensive deviation value and the sample fire risk basic index; train the environmental perception model based on the correspondence so that the environmental perception model can convert the real-time collected distributed perception data into a fire risk basic index.

[0042] In a preferred embodiment, first, data normalization is performed on the environmental parameter deviation mapping relationship. Since the dimensions and numerical ranges of parameters such as temperature, humidity, light intensity, and smoke concentration are different, it is necessary to uniformly map the deviation values of each parameter into the interval [0, 1] through normalization processing, so as to obtain comparable normalized environmental parameter deviations. Then, the normalized environmental parameter deviations are weighted and calculated according to preset weights. Among them, the preset weights reflect the influence of different environmental parameters on the degree of fire danger. For example, the weight of smoke concentration is higher than that of light intensity. The sample environmental comprehensive deviation value is obtained through weighted summation calculation, and this value comprehensively reflects the overall degree of environmental anomalies.

[0043] Next, establish the correspondence between the sample environmental comprehensive deviation value and the sample fire danger basic index. By analyzing the fire cases in historical data, determine the corresponding actual danger levels under different environmental comprehensive deviation values, and establish the correspondence between the two. Subsequently, based on the above correspondence, train the environmental perception model. Use machine learning methods (such as support vector machines, neural networks, etc.) to train the model so that it can accurately convert the distributed perception data collected in real time into the fire danger basic index representing the current environmental danger level. During the training process, minimize the error between the prediction result and the actual danger level by adjusting the model parameters. Through the above processing steps, an accurate environmental perception model is established to realize the intelligent assessment of the forest environmental danger level.

[0044] Furthermore, the embodiment of the present application further includes:

[0045] Establish a state equation and an observation equation, where the state equation describes the time evolution relationship of the target position and velocity, and the observation equation describes the measured values of the multi-source positioning data; calculate the initial state prediction value according to the state equation and the observation equation; construct a measurement residual sequence based on the measured values of the multi-source positioning data and the initial state prediction value; use the measurement residual sequence to estimate the measurement error characteristics online, and adaptively adjust the Kalman gain according to the measurement error characteristics; use the Kalman gain to correct the initial state prediction value to obtain the target positioning information.

[0046] In a preferred embodiment, when using the adaptive Kalman filtering algorithm to fuse multi-source positioning data, first, a state equation and an observation equation are established. The state equation uses a uniform motion model to describe the time evolution relationship between the target position and velocity, and the state vector includes three-dimensional spatial position coordinates and their corresponding velocity components; the observation equation describes the relationship between the measured values of multi-source positioning data such as Beidou positioning, barometric pressure, acceleration, gyroscope, magnetometer, and shoe-mounted inertial navigation and the state vector. Then, the initial state prediction value is calculated according to the established state equation and observation equation. Using the state estimate value of the previous moment, the state value of the current moment is predicted through the state equation; at the same time, the predicted measurement values of each sensor at the current moment are calculated based on the observation equation. Next, the actual measurement values of the multi-source positioning data are compared with the predicted measurement values to construct a measurement residual sequence. The measurement residual reflects the deviation between the predicted value and the actual measurement value, and can be used to evaluate the reliability of each sensor data. After that, an online analysis is performed on the measurement residual sequence using a sliding time window to estimate the statistical characteristics of the measurement error. According to the change of the error characteristics, the gain matrix of the Kalman filter is dynamically adjusted so that the filter can adaptively adjust the weights of different sensor data. Subsequently, the initial state prediction value is corrected using the adaptively adjusted Kalman gain. By performing weighted fusion on the predicted value and the actual measurement value, more accurate target positioning information is obtained.

[0047] Through the adaptive Kalman filtering algorithm, multi-source positioning data is effectively fused, improving the positioning accuracy and reliability. At the same time, the fusion weight is automatically adjusted according to the quality of the real-time measurement data, having strong environmental adaptability.

[0048] Furthermore, the embodiment of the present application further includes:

[0049] Extract a preset time window, perform sliding processing on the measurement residual sequence to obtain a window measurement residual sequence, and calculate the statistical mean and variance of the window measurement residual sequence; use the statistical mean and variance of the window measurement residual sequence to construct a measurement error covariance matrix, and the measurement error covariance matrix characterizes the credibility of each sensor's measurement data; adaptively calculate a dynamic weight coefficient according to the measurement error covariance matrix, and adaptively adjust the Kalman gain according to the dynamic weight coefficient.

[0050] In a preferred embodiment, first, a preset time window is extracted and slid. The length of the time window can be set to 10 - 30 sampling periods and slid on the measurement residual sequence at a fixed step size. For each sliding window position, the measurement residual sequence within the window is extracted, and the statistical features of the sequence are calculated. Specifically, the arithmetic mean (statistical mean) and the degree of dispersion (statistical variance) of the measurement residuals of each sensor within the window are calculated. Then, a measurement error covariance matrix is constructed based on the calculated statistical mean and variance. This matrix is a symmetric matrix. The elements on the main diagonal represent the variances of the measurement data of each sensor, reflecting the degree of dispersion of the data; the off-diagonal elements represent the covariance between different sensors, reflecting the correlation between the data. The numerical size in the matrix directly characterizes the credibility of the measurement data of each sensor, and the smaller the value, the higher the credibility. After that, the dynamic weight coefficient is calculated according to the measurement error covariance matrix. The inverse matrix of the covariance matrix is used as the basis for weight calculation, so that the sensor with smaller measurement error obtains a larger weight. The calculated dynamic weight coefficient is substituted into the gain calculation formula of the Kalman filter to achieve adaptive adjustment of the Kalman gain.

[0051] Through the above processing procedure, the dynamic evaluation of the credibility of sensor data is realized, and the data fusion weight is adaptively adjusted accordingly. This method can effectively improve the accuracy and robustness of multi-source positioning data fusion and adapt to the dynamic changes of sensor performance in different environments.

[0052] Furthermore, the embodiments of the present application further include:

[0053] Determine the area where the target is located according to the target positioning information; obtain the geographical information data of the area where the target is located, and the geographical information data includes terrain information, vegetation information, and water source distribution information; calculate the fire fighting difficulty coefficient according to the geographical information data, and the fire fighting difficulty coefficient characterizes the difficulty of fire fighting and rescue in the area where the target is located; correct the basic fire danger index by using the fire fighting difficulty coefficient to obtain the fire danger assessment index.

[0054] In a feasible embodiment, first, determine the area where the target is located according to the target positioning information. By matching the positioning coordinates with the electronic map, the specific area range where the firefighters are currently located is determined. The size of this area range can be set according to actual needs, such as a 500-meter radius range centered on the current position. Then, the geographical information data of this area is retrieved from the geographical information database. Specifically included are: terrain information reflecting the terrain undulation (such as altitude, slope, aspect, etc.), vegetation information reflecting the distribution of combustibles (such as vegetation type, coverage, density, etc.), and water source distribution information reflecting the availability of rescue resources (such as the location and capacity of natural water sources and artificial water sources).

[0055] Next, calculate the fire fighting difficulty coefficient based on the obtained geographical information data. This coefficient is determined by comprehensively evaluating multiple factors: terrain factor (the greater the slope, the higher the difficulty), vegetation factor (the greater the vegetation density, the higher the difficulty), water source factor (the farther the distance from the water source, the higher the difficulty), etc. The weighted summation method is used to synthesize each factor to obtain the final fire fighting difficulty coefficient. The larger this coefficient is, the greater the difficulty of fire fighting and rescue. After that, the fire fighting difficulty coefficient and the basic fire danger index are weighted and fused for calculation to obtain the fire danger assessment index. The specific calculation method is: Fire danger assessment index = Basic fire danger index × (1 + Fire fighting difficulty coefficient). In this way, the influencing factors of the geographical environment on fire fighting and rescue are taken into account, making the danger assessment result more comprehensive and accurate, so as to realize the dynamic assessment and correction of the environmental danger degree and provide a more reliable decision-making basis for the safety protection of firefighters.

[0056] Furthermore, the embodiments of the present application further include:

[0057] Establish a dual-channel communication link, including a first communication link and a second communication link. The first communication link uses satellite short message communication for long-distance data transmission, and the second communication link uses narrowband Internet of Things communication for medium-distance data transmission; packetize the alarm information according to a preset format to generate redundant data packets; send the redundant data packets in parallel through the first communication link and the second communication link; perform data verification and recombination on the received redundant data packets to ensure the integrity of the alarm information.

[0058] In a preferred embodiment, first, a dual-channel communication link is established. The communication link includes two independent communication channels: the first communication link uses satellite short message communication technology, which has the characteristics of wide coverage and is not restricted by geographical environment, and is suitable for long-distance data transmission with a communication distance of up to hundreds of kilometers; the second communication link uses narrowband Internet of Things communication technology, which has the characteristics of low power consumption and strong penetration, and is suitable for medium-distance data transmission with a communication distance of up to dozens of kilometers. Then, the alarm information to be sent is processed by data packet splitting according to a preset format. The preset format includes a packet header, a data segment, and a check segment. The packet header includes information such as a packet sequence number and a timestamp. The data segment includes the content of the alarm information. The check segment is used for data integrity verification. By data packet splitting, data packets with redundancy characteristics can be generated, and even if some data packets fail to be transmitted, it will not affect the restoration of the overall information. Next, the redundant data packets are sent in parallel through the dual channels. The same data packets are sent through the first communication link and the second communication link simultaneously to achieve redundant backup of data transmission. The two communication links use different transmission protocols and frequency bands, which can effectively reduce the risk of communication interruption. After that, the redundant data packets received at the receiving end are processed. First, data verification is performed, and the integrity of the data packets is verified by methods such as checksum and CRC. Then, the verified data packets are reorganized according to the packet sequence number, and the data packets from different communication links are integrated into complete alarm information. If a data packet of a certain communication link is lost or incorrect, the data packet of the other link can be used for supplementation.

[0059] Through the dual-channel redundant communication mechanism, the reliability of alarm information transmission can be significantly improved, and the communication problem in the forest environment can be effectively solved. Even in a complex terrain environment, the timely and reliable transmission of alarm information can be ensured.

[0060] In summary, the adaptive alarm triggering method based on environmental perception provided by the embodiments of the present application has the following technical effects:

[0061] Collect distributed sensing data of the forest environment. The distributed sensing data includes temperature, humidity, light intensity, and smoke concentration, and is used to sense the state of the forest environment. Obtain multi-source positioning data. The multi-source positioning data includes Beidou positioning data, barometric pressure data, acceleration data, gyroscope data, magnetometer data, and shoe-mounted inertial navigation data, and is used for subsequent precise positioning. Establish an environmental perception model, and use the environmental perception model to perform environmental perception on the distributed sensing data to determine the basic fire danger index and obtain a preliminary risk level assessment result. Use the adaptive Kalman filtering algorithm to fuse the multi-source positioning data to obtain target positioning information, improve the positioning accuracy through multi-source data fusion, and overcome the limitations of a single positioning method. Correct the basic fire danger index according to the target positioning information to obtain the fire danger assessment index, and optimize the risk level assessment in combination with the position information to improve the accuracy of the assessment result. When the fire danger assessment index exceeds the preset fire danger index threshold, send an alarm message through a long-distance redundant communication mechanism. The long-distance redundant communication mechanism includes satellite short message communication and narrowband Internet of Things communication, and ensures the reliable transmission of the alarm message through a dual-channel in case of danger.

[0062] Embodiment 2. Based on the same inventive concept as the method for adaptively triggering an alarm based on environmental perception in the foregoing embodiment, as Figure 2 shown, the embodiment of the present application provides an apparatus for adaptively triggering an alarm based on environmental perception. The apparatus includes:

[0063] An environment collection module 11, configured to collect distributed sensing data of the forest environment. The distributed sensing data includes temperature, humidity, light intensity, and smoke concentration.

[0064] A multi-source positioning module 12, configured to obtain multi-source positioning data. The multi-source positioning data includes Beidou positioning data, barometric pressure data, acceleration data, gyroscope data, magnetometer data, and shoe-mounted inertial navigation data.

[0065] A perception modeling module 13, configured to establish an environmental perception model, and use the environmental perception model to perform environmental perception on the distributed sensing data to determine the basic fire danger index.

[0066] A positioning fusion module 14, configured to use the adaptive Kalman filtering algorithm to fuse the multi-source positioning data to obtain target positioning information.

[0067] A risk assessment module 15, configured to correct the basic fire danger index according to the target positioning information to obtain the fire danger assessment index.

[0068] An alarm communication module 16, configured to send alarm information through a long-distance redundant communication mechanism when the fire danger assessment index exceeds a preset fire danger index threshold, where the long-distance redundant communication mechanism includes satellite short message communication and narrowband Internet of Things communication.

[0069] Further, the perception modeling module 13 includes the following execution steps:

[0070] Establish a standard parameter library for the forest environment, where the standard parameter library includes a standard temperature range, a standard humidity range, a standard light intensity range, and a standard smoke concentration range; obtain historical environmental data within a preset time period; perform feature extraction on the historical environmental data to obtain environmental feature parameters; compare the environmental feature parameters with the standard parameter library to establish an environmental parameter deviation mapping relationship; based on the environmental parameter deviation mapping relationship, establish the environmental perception model.

[0071] Further, the perception modeling module 13 further includes the following execution steps:

[0072] Perform data normalization processing on the environmental parameter deviation mapping relationship to obtain a normalized environmental parameter deviation; perform weighted calculation on the normalized environmental parameter deviation according to a preset weight to obtain a sample environmental comprehensive deviation value; establish a correspondence relationship between the sample environmental comprehensive deviation value and the sample fire danger basic index; train the environmental perception model based on the correspondence relationship so that the environmental perception model can convert the distributed perception data collected in real time into a fire danger basic index.

[0073] Further, the positioning and fusion module 14 includes the following execution steps:

[0074] Establish a state equation and an observation equation, where the state equation describes the time evolution relationship of the target position and speed, and the observation equation describes the measured values of the multi-source positioning data; calculate an initial state prediction value according to the state equation and the observation equation; construct a measurement residual sequence based on the measured values of the multi-source positioning data and the initial state prediction value; use the measurement residual sequence to online estimate the measurement error characteristics, and adaptively adjust the Kalman gain according to the measurement error characteristics; use the Kalman gain to correct the initial state prediction value to obtain the target positioning information.

[0075] Further, the positioning and fusion module 14 further includes the following execution steps:

[0076] Extract a preset time window, perform a sliding process on the measurement residual sequence to obtain a window measurement residual sequence, and calculate the statistical mean and variance of the window measurement residual sequence; use the statistical mean and variance of the window measurement residual sequence to construct a measurement error covariance matrix, where the measurement error covariance matrix characterizes the credibility of the measurement data of each sensor; adaptively calculate a dynamic weight coefficient according to the measurement error covariance matrix, and adaptively adjust the Kalman gain according to the dynamic weight coefficient.

[0077] Further, the risk assessment module 15 includes the following execution steps:

[0078] Determine the area where the target is located according to the target positioning information; obtain the geographical information data of the area where the target is located, where the geographical information data includes terrain information, vegetation information, and water source distribution information; calculate a fire fighting difficulty coefficient according to the geographical information data, where the fire fighting difficulty coefficient characterizes the difficulty of fire fighting and rescue in the area where the target is located; correct the basic fire risk index by using the fire fighting difficulty coefficient to obtain the fire risk assessment index.

[0079] Further, the alarm communication module 16 includes the following execution steps:

[0080] Establish a dual-channel communication link, including a first communication link and a second communication link, where the first communication link uses satellite short message communication for long-distance data transmission, and the second communication link uses narrowband Internet of Things communication for medium-distance data transmission; packetize the alarm information in a preset format to generate redundant data packets; send the redundant data packets in parallel through the first communication link and the second communication link; perform data verification and recombination on the received redundant data packets to ensure the integrity of the alarm information.

[0081] Any step of the method described above can be stored as computer instructions or programs in an unrestricted computer memory and can be called and recognized by an unrestricted computer processor to implement any method in the embodiments of the present application, and no redundant restrictions are made here.

[0082] Further, the first or second described above may not only represent an order relationship, but may also represent a certain specific concept, and / or refer to the fact that multiple elements can be selected individually or in whole. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. An adaptive alarm triggering method based on environmental perception, characterized in that Including: Collecting distributed perception data of the forest environment, where the distributed perception data includes temperature, humidity, light intensity, and smoke concentration; Obtaining multi-source positioning data, where the multi-source positioning data includes Beidou positioning data, barometric pressure data, acceleration data, gyroscope data, magnetometer data, and shoe-mounted inertial navigation data; Establishing an environmental perception model, and using the environmental perception model to perform environmental perception on the distributed perception data to determine the basic fire danger index; Using the adaptive Kalman filter algorithm to perform fusion processing on the multi-source positioning data to obtain target positioning information; Correcting the basic fire danger index according to the target positioning information to obtain the fire danger assessment index; When the fire danger assessment index exceeds the preset fire danger index threshold, sending an alarm message through a long-distance redundant communication mechanism, where the long-distance redundant communication mechanism includes satellite short message communication and narrowband Internet of Things communication; Correcting the basic fire danger index according to the target positioning information to obtain the fire danger assessment index, including: Determining the area where the target is located according to the target positioning information; Obtaining geographical information data of the area where the target is located, where the geographical information data includes terrain information, vegetation information, and water source distribution information; Calculating the fire fighting difficulty coefficient according to the geographical information data, where the fire fighting difficulty coefficient characterizes the difficulty of fire fighting and rescue in the area where the target is located; Using the fire fighting difficulty coefficient to correct the basic fire danger index to obtain the fire danger assessment index.

2. The adaptive alarm triggering method based on environmental perception according to claim 1, characterized in that, Establishing an environmental perception model, including: Establishing a standard parameter library for the forest environment, where the standard parameter library includes a standard temperature range, a standard humidity range, a standard light intensity range, and a standard smoke concentration range; Obtaining historical environmental data within a preset time period; Performing feature extraction on the historical environmental data to obtain environmental feature parameters; Comparing the environmental feature parameters with the standard parameter library to establish an environmental parameter deviation mapping relationship; Based on the environmental parameter deviation mapping relationship, establishing the environmental perception model.

3. The adaptive alarm triggering method based on environmental perception according to claim 2, characterized in that Based on the environmental parameter deviation mapping relationship, establishing the environmental perception model, including: Performing data normalization processing on the environmental parameter deviation mapping relationship to obtain a normalized environmental parameter deviation; Performing weighted calculation on the normalized environmental parameter deviation according to a preset weight to obtain a sample environmental comprehensive deviation value; Establishing a corresponding relationship between the sample environmental comprehensive deviation value and the sample basic fire danger index; Training the environmental perception model based on the corresponding relationship so that the environmental perception model can convert the real-time collected distributed perception data into the basic fire danger index.

4. The adaptive alarm triggering method based on environmental perception according to claim 1, wherein Using the adaptive Kalman filter algorithm to perform fusion processing on the multi-source positioning data to obtain target positioning information, including: Establishing a state equation and an observation equation, where the state equation describes the time evolution relationship of the target position and speed, and the observation equation describes the measured values of the multi-source positioning data; Calculating an initial state prediction value according to the state equation and the observation equation; Constructing a measurement residual sequence based on the measured values of the multi-source positioning data and the initial state prediction value; Online estimate the measurement error characteristics using the measurement residual sequence, and adaptively adjust the Kalman gain according to the measurement error characteristics; Correct the initial state prediction value using the Kalman gain to obtain the target positioning information.

5. The adaptive alarm triggering method based on environmental perception according to claim 4, wherein Online estimate the measurement error characteristics using the measurement residual sequence, and adaptively adjust the Kalman gain according to the measurement error characteristics, including: Extract a preset time window, perform a sliding process on the measurement residual sequence to obtain a window measurement residual sequence, and calculate the statistical mean and variance of the window measurement residual sequence; Construct a measurement error covariance matrix using the statistical mean and variance of the window measurement residual sequence, and the measurement error covariance matrix characterizes the credibility of the measurement data of each sensor; Adaptive calculate a dynamic weight coefficient according to the measurement error covariance matrix, and adaptively adjust the Kalman gain according to the dynamic weight coefficient.

6. The adaptive alarm triggering method based on environmental perception according to claim 1, wherein Send an alarm message through a long-distance redundant communication mechanism, including: Establish a dual-channel communication link, including a first communication link and a second communication link, where the first communication link uses satellite short message communication for long-distance data transmission, and the second communication link uses narrowband Internet of Things communication for medium-distance data transmission; Packetize the alarm message according to a preset format to generate redundant data packets; Parallelly send the redundant data packets through the first communication link and the second communication link; Perform data verification and recombination on the received redundant data packets to ensure the integrity of the alarm message.

7. An adaptive alarm triggering device based on environmental perception, characterized in that, Including: An environment acquisition module, which is used to acquire distributed perception data of the forest environment, and the distributed perception data includes temperature, humidity, light intensity, and smoke concentration; A multi-source positioning module, which is used to obtain multi-source positioning data, and the multi-source positioning data includes Beidou positioning data, barometric pressure data, acceleration data, gyroscope data, magnetometer data, and shoe-mounted inertial navigation data; A perception modeling module, which is used to establish an environment perception model, and use the environment perception model to perform environment perception on the distributed perception data to determine the basic fire danger index; A positioning fusion module, which is used to perform fusion processing on the multi-source positioning data using an adaptive Kalman filtering algorithm to obtain target positioning information; A danger assessment module, which is used to correct the basic fire danger index according to the target positioning information to obtain a fire danger assessment index; An alarm communication module, which is used to send an alarm message through a long-distance redundant communication mechanism when the fire danger assessment index exceeds a preset fire danger index threshold, and the long-distance redundant communication mechanism includes satellite short message communication and narrowband Internet of Things communication; The danger assessment module includes the following execution steps: Determine the area where the target is located according to the target positioning information; obtain the geographical information data of the area where the target is located, and the geographical information data includes terrain information, vegetation information, and water source distribution information; calculate the fire fighting difficulty coefficient according to the geographical information data, and the fire fighting difficulty coefficient represents the difficulty level of fire fighting and rescue in the area where the target is located; use the fire fighting difficulty coefficient to correct the basic fire danger index to obtain the fire danger assessment index.

Images