Internet of Things alarm audio call scheduling method and system

Abstract

The invention discloses an Internet of Things alarm audio call scheduling method and system, and relates to the technical field of Internet of Things intelligent decision, and the method comprises the steps: receiving an alarm request message from an Internet of Things sensing terminal, and carrying out the alarm audio call scheduling according to a device identifier and a region code in the alarm request message; collecting and generating a multi-dimensional information report of real-time environment noise data, audio playing equipment state information and user state information; according to the multi-dimensional information report, generating a dynamic priority score of the alarm request message by querying a preset strategy library; and sorting the dynamic priority scores of the alarm request messages to be processed, and generating a preliminary scheduling strategy for suggesting playing equipment and broadcasting parameters for the alarm request message with the highest priority. According to the method, high-fidelity simulation is performed through the acoustic digital twin model to generate the intelligibility index report, the optimal strategy is selected based on the simulation result to generate the execution instruction, and the execution instruction is issued to the target audio equipment, so that full-process intelligent scheduling from multi-dimensional dynamic perception to acoustic effect optimization is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet of Things intelligent decision-making, in particular to a method and system for alarm audio call scheduling of Internet of Things. BACKGROUND

[0002] Under the background of rapid development of Internet of Things technology, various types of environment sensing terminals are widely deployed in the fields of intelligent security, industrial Internet of Things and smart home, etc., for real-time monitoring of abnormal events such as smoke, intrusion, leakage, etc. After triggering an alarm, these terminals need to issue an alarm call to the target area through an audio playing device to prompt personnel to respond in time. In the prior art, after the system receives an alarm, it assigns a fixed priority to the alarm according to the inherent level of the alarm type, and then distributes the audio broadcast task to the idle playing devices in the area in order of priority. Usually, only limited dimensions such as alarm type and device availability are considered, and the scheduling is completed through simple loop or preemption strategy, aiming to achieve fast broadcast of alarm information. To some extent, this kind of scheme meets the needs of orderly triggering of multiple alarms, and because of its simple rules and low computational overhead, it is easy to deploy and implement on resource-constrained edge computing nodes.

[0003] However, with the complication of Internet of Things application scenarios and the improvement of users' requirements for alarm effectiveness, the above-mentioned scheduling method based on static rules and device status gradually shows its limitations. In terms of dynamic environment adaptability, the prior art is difficult to make prospective prediction and optimization of the actual propagation effect of alarm broadcasting. Since the real-time environmental noise level at the time of alarm occurrence, the actual performance status of the playing device, the current activity status of the user and other dynamic context information are not fully considered, and there is a lack of quantitative simulation of the propagation characteristics of audio in a specific physical space, the scheduling decision may deviate from the actual acoustic effect. SUMMARY

[0004] In view of the above-mentioned existing problems, the present application is proposed.

[0005] Therefore, the present application provides a method for alarm audio call scheduling of Internet of Things, which solves the problem of deviation between scheduling decision and actual broadcast effect caused by lack of multi-dimensional dynamic information fusion and acoustic propagation effect simulation in the prior art.

[0006] To solve the above technical problems, the present application provides the following technical solutions: In a first aspect, the present application provides a method for alarm audio call scheduling of Internet of Things, which comprises receiving an alarm request message from an Internet of Things sensing terminal, collecting and generating a multi-dimensional information report of real-time environmental noise data, audio playing device state information and user state information according to the device identifier and region code in the alarm request message; According to the multi-dimensional information report, a dynamic priority score of the alarm request message is generated by querying a preset strategy library. The dynamic priority score of the alarm request message to be processed is sorted, and a preliminary scheduling strategy of the recommended playing device and the broadcast parameter is generated for the alarm request message with the highest priority; The preliminary scheduling strategy is input into the acoustic digital twin model of the alarm occurrence area, and through the digital twin, the preliminary scheduling strategy is subjected to high-fidelity acoustic simulation and ambiguity analysis to generate a simulation report of intelligibility indicators by changing from static environment modeling to a dynamic scheduling strategy prediction and optimization engine; According to the simulation report of the intelligibility indicators, the preliminary scheduling strategy is changed from the priority score to a quantifiable expected propagation effect comparison based on physical acoustic simulation, and the strategy parameters are fine-tuned and optimized according to the acoustic confusion risk assessment, and the preliminary scheduling strategy with the optimal intelligibility indicators and the lowest ambiguity risk is set as the final scheduling execution instruction, and the final scheduling execution instruction is issued to the target audio playing device.

[0007] As a preferred scheme of the method for scheduling the alarm audio call of the Internet of Things, wherein: receiving the alarm request message from the Internet of Things sensing terminal, collecting and generating a multi-dimensional information report of real-time environmental noise data, audio playing device state information and user state information according to the device identifier and the alarm occurrence area code in the alarm request message, including the following steps: Parsing the device unique identifier and the alarm occurrence area code in the alarm request message, querying the environmental monitoring service according to the device unique identifier and the alarm occurrence area code to obtain real-time environmental noise data, and querying the device management device according to the device unique identifier and the alarm occurrence area code to obtain the audio playing device state information; Querying the user state service according to the device unique identifier and the alarm occurrence area code to obtain the user state information; Integrating the real-time environmental noise data, the audio playing device state information and the user state information to generate a multi-dimensional information report.

[0008] As a preferred scheme of the method for scheduling the alarm audio call of the Internet of Things, wherein: according to the multi-dimensional information report, a dynamic priority score of the alarm request message is generated by querying a preset strategy library, including the following steps: Reading the environmental noise value in the multi-dimensional information report and directly converting it into an environmental noise coefficient; Extracting the device state data, collecting the device state data through multi-dimensional device health indicators, generating a dynamic health score by using weighted fusion, establishing a device health score library, and directly obtaining a device reliability coefficient; Mapping the user activity information to a user state emergency coefficient, generating a basic priority coefficient from the alarm type extracted from the alarm request message, obtaining the time stamp in the alarm request message and deriving the time difference from the current time, and deriving a time attenuation factor through an exponential decay function; An associated event enhancement factor is generated by encoding the alarm occurrence area into the query log statistics associated event quantity; The basic priority coefficient, the environmental noise coefficient, the user state emergency coefficient, the device reliability coefficient, the time attenuation factor and the associated event enhancement factor are integrated to generate a dynamic priority score.

[0009] As a preferred scheme of the alarm audio call scheduling method of the Internet of Things, wherein: the dynamic priority score of the alarm request message to be processed is sorted, and a preliminary scheduling strategy of the recommended playing device and the broadcast parameter is generated for the alarm request message with the highest priority, including the following steps: The alarm request messages to be processed are sorted from high to low according to the dynamic priority score, an ordered sorted alarm request queue is generated, the alarm request message at the first position in the sorted alarm request queue is selected, and the highest priority alarm to be scheduled is obtained; The online audio playing device registered in the device management platform is queried according to the alarm occurrence area code in the alarm request message, and a candidate audio playing device list and a candidate audio playing device list for alarm broadcast are obtained; For each audio playing device in the candidate audio playing device list, a device-environment matching degree score is screened out; The expected broadcast utility value of each audio playing device is generated by combining the device-environment matching degree score, the dynamic priority score of the alarm request message and the environmental noise data in the multi-dimensional information report, and a preliminary scheduling strategy of the recommended volume, tone and repetition number is generated for each device based on the expected broadcast utility value.

[0010] As a preferred scheme of the alarm audio call scheduling method of the Internet of Things, wherein: the preliminary scheduling strategy is input into the acoustic digital twin model of the alarm occurrence area, the digital twin is converted from static environment modeling to dynamic scheduling strategy prediction and optimization engine, high-fidelity acoustic simulation and ambiguity analysis are performed on the preliminary scheduling strategy, and a simulation report of intelligibility index is generated, including the following steps: Based on the physical space three-dimensional geometric data corresponding to the alarm occurrence area code, the interface material acoustic parameters and the real-time environmental noise data, the acoustic digital twin model is obtained by the finite element method; Each strategy in the preliminary scheduling strategy is converted into an acoustic simulation parameter, a point sound source is set in the acoustic digital twin model according to the position of the target audio playing device in each strategy in the preliminary scheduling strategy, the sound power level is calculated according to the broadcast volume parameter in the strategy, the alarm audio segment to be broadcast is loaded, and the sound source signal is obtained; In the acoustic digital twin model, a virtual microphone array is deployed as a receiving point in the user activity area, the sound wave propagation, reflection and absorption process in the acoustic digital twin model corresponding to each strategy is calculated in parallel by using the ray tracing method, and impulse response data at the receiving point is obtained; Based on the impulse response data and the alarm audio segment, the actual audio signal at the receiving point is simulated by convolution, and the mel frequency cepstral coefficient (MFCC) feature of the simulated actual audio signal is extracted by time-frequency analysis; Based on the impulse response data and the alarm audio segment, the voice transmission index value of the receiving point is calculated; Integrate the voice transmission index value, semantic confusion probability and simulation parameters corresponding to each preliminary scheduling strategy to generate a simulation report of intelligibility index.

[0011] As a preferred scheme of the method for scheduling the alarm audio call of the Internet of Things, wherein: according to the simulation report of the intelligibility index, the preliminary scheduling strategy is converted from the priority score to the quantifiable expected propagation effect comparison based on physical acoustic simulation, and the strategy parameters are fine-tuned and optimized according to the acoustic confusion risk assessment, and the preliminary scheduling strategy with the optimal intelligibility index and the lowest ambiguity risk is selected as the final scheduling execution instruction, including the following steps: Extract the voice transmission index value corresponding to each preliminary scheduling strategy from the simulation report to generate a mapping list; Calculate the device resource consumption coefficient based on the preliminary scheduling strategy parameters in the mapping list; Query the target audio playback device identifier specified in the strategy in the mapping list, obtain the device reliability coefficient and calculate the execution risk coefficient; Substitute the voice transmission index value, the semantic confusion probability, the device resource consumption coefficient and the execution risk coefficient into the comprehensive performance score formula to calculate the comprehensive performance score of each strategy; Select the preliminary scheduling strategy with the highest comprehensive performance score as the final scheduling execution instruction.

[0012] As a preferred scheme of the method for scheduling the alarm audio call of the Internet of Things, wherein: the final scheduling execution instruction is sent to the target audio playback device, including the following steps: Encapsulate the final scheduling execution instruction into a standard control protocol data packet recognizable by the device, and send the control protocol data packet to the network endpoint corresponding to the target audio playback device identifier through a secure transmission layer protocol encrypted channel; The target audio playback device receives and analyzes the control protocol data packet to extract the broadcast parameters; The target audio playback device performs local audio broadcast operation according to the broadcast parameters, and the scheduling engine returns a status code to confirm the response.

[0013] In a second aspect, the present application provides a system for alarm audio call scheduling of Internet of Things, comprising a fusion module configured to receive an alarm request message from an Internet of Things sensing terminal, collect and generate a multi-dimensional information report of real-time environmental noise data, audio playing device state information and user state information according to a device identifier and a region code in the alarm request message; a dynamic priority module configured to generate a dynamic priority score of the alarm request message by querying a preset strategy library according to the multi-dimensional information report; a preliminary scheduling strategy generation module configured to sort the dynamic priority scores of the alarm request messages to be processed, and generate a preliminary scheduling strategy of a suggested playing device and a broadcast parameter for an alarm request message with the highest priority; a simulation report module configured to input the preliminary scheduling strategy into an acoustic digital twin model of an alarm occurrence region, and perform high-fidelity acoustic simulation and ambiguity analysis on the preliminary scheduling strategy by a digital twin from static environment modeling to a dynamic scheduling strategy prediction and optimization engine to generate a simulation report of intelligibility indicators; a decision module configured to convert the preliminary scheduling strategy from a priority score to a quantifiable expected propagation effect comparison based on physical acoustic simulation according to the simulation report of intelligibility indicators, fine-tune and optimize the strategy parameters according to an acoustic confusion risk assessment, select a preliminary scheduling strategy with optimal intelligibility indicators and lowest ambiguity risk as a final scheduling execution instruction, and issue the final scheduling execution instruction to a target audio playing device.

[0014] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, any step of the method for alarm audio call scheduling of Internet of Things according to the first aspect of the present application is implemented.

[0015] In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, any step of the method for alarm audio call scheduling of Internet of Things according to the first aspect of the present application is implemented.

[0016] The present application has the following beneficial effects: the alarm request message from the Internet of Things sensing terminal is received, the multi-dimensional information report is generated based on the device identifier and the region code to collect the environmental noise, device state and user state information, and the dynamic priority score is generated by using weighted calculation; the preliminary scheduling strategy is generated for the highest priority alarm after the alarm request is sorted, the intelligibility indicator report is generated by high-fidelity simulation through the acoustic digital twin model, the optimal strategy is selected based on the simulation result to generate the execution instruction and issue it to the target audio device, and the intelligent scheduling of the whole process from multi-dimensional dynamic sensing to acoustic effect optimization is realized. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0018] Fig. 1 Flow chart of the method for scheduling alarm audio call of the Internet of Things.

[0019] Fig. 2 Schematic diagram of the system for scheduling alarm audio call of the Internet of Things. DETAILED DESCRIPTION

[0020] In order to make the above objectives, features and advantages of the present application more apparent and comprehensible, the specific embodiments of the present application will be described in detail below with reference to the drawings.

[0021] In the following description, a large number of specific details are set forth in order to facilitate a thorough understanding of the present application, but the present application can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the connotation of the present application, therefore the present application is not limited by the specific embodiments disclosed below.

[0022] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the present application. "In one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an embodiment that is separate or alternative to other embodiments.

[0023] Reference Figs. 1-2 For one embodiment of the present application, the embodiment provides a method for scheduling alarm audio call of the Internet of Things, comprising the following steps: S1, receiving an alarm request message from an Internet of Things sensing terminal, collecting and generating a multi-dimensional information report of real-time environmental noise data, audio playback device state information and user state information according to the device identifier and the area code in the alarm request message.

[0024] S1.1, parsing the device unique identifier and the alarm occurrence area code in the alarm request message, querying the environmental monitoring service according to the device unique identifier and the alarm occurrence area code to obtain real-time environmental noise data, and querying the device management device according to the device unique identifier and the alarm occurrence area code to obtain audio playback device state information.

[0025] Further, the receiving Internet of Things sensing terminal sends an alarm request message through a message queue or an interface, parses the protocol load of the alarm request message, extracts the device unique identifier and the alarm occurrence area code; takes the device unique identifier and the alarm occurrence area code as joint query conditions to call the application program interface of the environmental monitoring service, the environmental monitoring service returns the current environmental noise decibel value collected by the audio sensor in the physical space corresponding to the alarm occurrence area code as real-time environmental noise data, takes the device unique identifier and the alarm occurrence area code as index query application program interface of the device management platform, and the device management platform returns a detailed state set of all registered online audio playback devices in the alarm occurrence area code as audio playback device state information. The detailed state set includes device network round-trip delay, central processor current load, battery remaining power percentage, and current volume setting gear.

[0026] S1.2, query the user state service according to the device unique identifier and the alarm occurrence area code, and obtain user state information.

[0027] Further, the application program interface of the user state service is called with the device unique identifier and the alarm occurrence area code as parameters, the user state service returns the target user's current activity state inference result as user state information based on the fusion of intelligent wearable device sensor data and calendar schedule information, and the user state information includes sleep, meeting, and idle state classification.

[0028] S1.3, integrate real-time environmental noise data, audio playback device state information, and user state information to generate a multi-dimensional information report.

[0029] Further, the real-time environmental noise data, audio playback device state information, and user state information are converted into a standardized format, the data fields are aligned and time stamped, and a structured multi-dimensional information report containing all context information is generated. The multi-dimensional information report provides a complete data basis for subsequent scheduling decisions.

[0030] S2, according to the multi-dimensional information report, generate a dynamic priority partition of the alarm request message by querying a preset strategy library.

[0031] S2.1, read the environmental noise value in the multi-dimensional information report and directly convert it into an environmental noise coefficient.

[0032] Further, reading the environmental noise value in the multi-dimensional information report and directly converting it into an environmental noise coefficient is achieved through a pre-defined noise level mapping table. The mapping table divides continuous environmental noise values into discrete level intervals and assigns a standardized environmental noise coefficient to each interval. The environmental noise value is input immediately after matching query in the mapping table and the corresponding environmental noise coefficient is output to complete the direct conversion from physical quantity to coefficient.

[0033] S2.2, extract equipment state data, collect equipment state data through multi-dimensional equipment health indicators, generate dynamic health score by weighted fusion, establish equipment health score library, and directly obtain equipment reliability coefficient.

[0034] Further, the extraction of equipment state data is through the continuous collection of multi-dimensional equipment health indicators during equipment operation, including current remaining percentage signal strength coefficient, historical online rate and recent failure record frequency. The dynamic health scores of all equipment are persistently stored and maintained to form an equipment health score library. When the equipment reliability coefficient is needed, the latest score corresponding to the equipment identifier is directly retrieved from the equipment health score library as the equipment reliability coefficient.

[0035] S2.3, analyze user activity information and map it to user state emergency coefficient, extract alarm type from alarm request message to generate basic priority coefficient, obtain timestamp in alarm request message and derive time difference from current time, derive time attenuation factor through exponential decay function.

[0036] Further, the analysis of user activity information and the mapping to user state emergency coefficient is based on user activity state classification rules to classify user activity information into predefined emergency state levels, each level corresponding to a fixed user state emergency coefficient value. The extraction of alarm type from alarm request message to generate basic priority coefficient is through querying the alarm type priority table, which sets a fixed basic priority coefficient for each alarm type. The time difference between the time represented by the timestamp and the current time is calculated in seconds. The time attenuation factor is derived by substituting the time difference into the exponential decay function to obtain a time attenuation factor that asymptotically approaches zero as the time difference increases.

[0037] S2.4, generate associated event enhancement factor by encoding alarm occurrence area and querying log to count associated event number.

[0038] Further, the generation of associated event enhancement factor by encoding alarm occurrence area and querying log to count associated event number is to use alarm occurrence area encoding as a query key to retrieve the number of similar alarm event records recently occurred in the area in the event log database. The number of events is input into a linear normalization function, which maps the number to an associated event enhancement factor value within a predetermined range. The more the number of events, the larger the value of the associated event enhancement factor.

[0039] S2.5, integrate the basic priority coefficient, environmental noise coefficient, user state emergency coefficient, equipment reliability coefficient, time attenuation factor and associated event enhancement factor to generate dynamic priority score.

[0040] Further, the base priority coefficient, the environmental noise coefficient, the user state coefficient, the emergency coefficient, the device reliability coefficient, the time decay factor and the correlation event enhancement factor are integrated to generate a dynamic priority score.

[0041] S3. The dynamic priority scores of the alarm request messages to be processed are sorted, and a preliminary scheduling strategy of the recommended playing device and the broadcast parameter is generated for the alarm request message with the highest priority.

[0042] S3.1. The alarm request messages to be processed are sorted from high to low according to the dynamic priority scores, an ordered sorted alarm request queue is generated, the alarm request message at the first position in the sorted alarm request queue is selected, and the highest priority alarm to be scheduled at present is obtained.

[0043] Further, the dynamic priority scores of all alarm request messages to be processed are sorted in descending order, a sorted alarm request queue arranged from high to low according to the dynamic priority scores is generated, the alarm request message at the first position in the sorted alarm request queue is selected, and the highest priority alarm to be scheduled at present is obtained.

[0044] S3.2. The online audio playing devices registered in the device management platform are queried according to the alarm occurrence area code in the alarm request message, and the candidate audio playing device list and the candidate audio playing device list for alarm broadcast are obtained.

[0045] Further, according to the alarm occurrence area code contained in the alarm request message of the highest priority alarm, the online audio playing device list registered in the device management platform is queried, the identifiers and detailed state parameters of all audio playing devices in the area code with the online state are obtained, and the candidate audio playing device list is formed.

[0046] S3.3. For each audio playing device in the candidate audio playing device list, a device-environment matching degree score is screened out.

[0047] Further, for each audio playing device in the candidate audio playing device list, the device-environment matching degree score is screened out by the following steps: first, the real-time network delay coefficient and signal strength coefficient of the audio playing device are obtained, and the current environmental noise data is extracted from the multi-dimensional information report; then, the network delay coefficient, the signal strength coefficient and the environmental noise data are taken as inputs to query the pre-defined device-environment matching degree mapping table, and the obtained device-environment matching degree score is directly output as the result.

[0048] S3.4, combine the device-environment matching score, the dynamic priority score of the alarm request message, and the environmental noise data in the multi-dimensional information report to generate an expected broadcast utility value for each audio playing device, and generate a preliminary scheduling strategy of the recommended volume, tone, and repetition number for each device based on the expected broadcast utility value.

[0049] Furthermore, by means of multi-factor weighted fusion and acoustic utility optimization technology, the independent evaluation bias of the device-environment matching score, the dynamic priority score, and the environmental noise data is linearly weighted and nonlinearly normalized to integrate the discrete evaluation indicators into a unified expected broadcast utility value. Furthermore, by means of parameter mapping and rule reasoning technology, the strategy generation ambiguity of the expected broadcast utility value is segmented functionally mapped and empirically rule-matched to generate a preliminary scheduling strategy including a quantitative recommended volume, an adaptive tone adjustment, and a dynamic repetition number.

[0050] S4, input the preliminary scheduling strategy into the acoustic digital twin model of the alarm occurrence area, and through digital twinning, convert the static environment modeling into a dynamic scheduling strategy prediction and optimization engine, perform high-fidelity acoustic simulation and ambiguity analysis on the preliminary scheduling strategy, and generate a simulation report of intelligibility indicators.

[0051] S4.1, based on the alarm occurrence area encoding corresponding physical space three-dimensional geometric data, interface material acoustic parameters, and real-time environmental noise data, obtain the acoustic digital twin model by the finite element method.

[0052] Furthermore, by means of finite element mesh partitioning and wave equation discretization technology, the complex boundary representation deficiency of the alarm occurrence area encoding corresponding physical space three-dimensional geometric data is tetrahedral element mesh partitioning and interface condition loading, and the continuous physical space is converted into a discrete node system. Furthermore, by means of material parameter mapping and acoustic property assignment technology, the frequency variation characteristic loss of the interface material acoustic parameters is defined as a complex impedance boundary condition and a frequency-dependent sound absorption coefficient is fitted. Finally, by means of environmental noise coupling and transient sound field calculation technology, the dynamic interference unmodeled of the real-time environmental noise data is background sound field superposition and excitation source integration, and a high-fidelity acoustic digital twin model is constructed, which integrates static geometric characteristics, material acoustic characteristics, and dynamic environmental interference.

[0053] S4.2, convert each strategy in the preliminary scheduling strategy into an acoustic simulation parameter, set a point sound source in the acoustic digital twin model according to the target audio playing device position in each strategy in the preliminary scheduling strategy, calculate the sound power level according to the broadcast volume parameter in the strategy, load the alarm audio segment to be broadcast, and obtain the sound source signal.

[0054] Further, through the acoustic parameter mapping and sound field modeling technical means, the spatial positioning deviation and energy quantization mismatching defects existing in the discrete device position and broadcast parameters in the preliminary scheduling strategy are processed by three-dimensional coordinate conversion and sound power level standardization, and the strategy configuration is converted into a point sound source with clear spatial coordinates and acoustic energy; further through the audio segment loading and signal generation technical means, the format heterogeneity and loudness inconsistency defects existing in the alarm audio segment are processed by sampling rate unification and amplitude normalization, and a sound source signal with spatial attributes, energy attributes and waveform attributes conforming to the input specification of the acoustic digital twin model is generated.

[0055] S4.3, deploy a virtual microphone array as a receiving point in the user activity area of the acoustic digital twin model, use ray tracing method to calculate the sound wave propagation, reflection and absorption process in the acoustic digital twin model corresponding to each strategy in parallel, and obtain the impulse response data at the receiving point.

[0056] The impulse response data expression is: ; Wherein, is the impulse response data, is the geometric amplitude attenuation coefficient of the m-th ray, is the ray index, is the propagation time delay of the m-th ray, is the reflection coefficient of the sound wave propagation direction in the n-th reflection event, is the sound absorption coefficient on the n-th propagation path, is the propagation path, is the reflection index, is the time variable, is the reflection number of the m-th ray, is the air absorption coefficient, is the distance of the sound wave propagation on the n-th path, is the total propagation time of the m-th ray, is the angle between the sound wave propagation direction and the normal direction of the reflection interface in the n-th reflection event, is the total number of rays, is the medium region with different acoustic characteristics through which the m-th ray passes in the propagation process.

[0057] ​​​​​​​​​Further, in the acoustic digital twin model, the user's key activity area is located, a virtual microphone array is deployed as an acoustic receiving point, a sound ray tracing algorithm is used to calculate the sound wave propagation path of each strategy in the model, the reflection, diffraction and material absorption effect of the sound wave in the propagation process are simulated, the sound pressure response data at each receiving point over time is recorded, and high-precision impulse response data is generated.

[0058] It should be noted that the linear superposition principle summation term The total sound pressure at any receiving point in the sound field is the linear superposition of the sound wave contributions from all propagation paths, each sound ray represents an independent sound wave propagation path, and the physical nature is the superposition property of the wave equation solution, which conforms to the Huygens-Fresnel principle.

[0059] Geometric divergence attenuation Describes the natural attenuation of energy due to the expansion of the wave front when the sound wave propagates in a free field as a spherical wave, the calculation formula = , wherein is the path length of the first sound ray, which is directly derived from the physical law that sound intensity is inversely proportional to the square of the distance, and is a basic characteristic of point source radiation.

[0060] Reflection cumulative loss multiplication term The energy loss of each reflection is described by the reflection coefficient , the value of which depends on the interface material properties, sound wave frequency and incident angle , the modulus of the reflection coefficient is less than or equal to 1, and the loss of multiple reflections is calculated by multiplication, which conforms to the physical fact that energy decays gradually.

[0061] Medium absorption attenuation When the sound wave propagates in the air, energy is absorbed due to viscosity, heat conduction and molecular relaxation effects, the absorption coefficient significantly increases with increasing frequency, and the exponential decay form is the sound wave absorption, which conforms to the Lambert-Beer law.

[0062] Time delay and phase Dirac function The energy of each sound ray reaches the receiving point at a certain time, and the Dirac function accurately describes the time delay effect of the sound wave propagation, which is manifested as a phase shift in frequency domain calculation, and is the core cause of sound field interference and reverberation phenomenon.

[0063] Specifically, the traditional algorithm: numerical solution of wave equation, directly solving Helmholtz equation or sound wave equation, which needs to be discretized in the whole space, and the calculation complexity increases exponentially with the model size.

[0064] Limitations: huge computational resource consumption, difficult to handle multi-strategy scenarios in parallel, and lack of intuitive path-level physical decomposition.

[0065] Present algorithm improvement: high-frequency approximation and energy path decomposition, high-frequency acoustic assumption, when the wavelength of sound wave is much smaller than the geometric size of the environment, ray tracing method can approximate wave propagation, and the physical meaning is clear.

[0066] Energy superposition: the total sound field is regarded as the superposition of the energy of multiple independent sound rays, and each sound ray represents a propagation path.

[0067] S4.4, based on the impulse response data and the alarm audio segment, convolution is performed to simulate the actual audio signal at the receiving point, time-frequency analysis is performed on the simulated actual audio signal, and the mel frequency cepstral coefficient MFCC feature is extracted.

[0068] Further, through convolution operation technical means, the time domain convolution and frequency domain filtering processing of the time domain aliasing and energy distortion of the receiving point simulation audio signal synthesized by the impulse response data and the alarm audio segment, through the short-time Fourier transform and the mel filter bank technical means, the non-stationary characteristics and the high frequency resolution of the simulation audio signal are windowed and framed and the critical band mapping is performed, the wideband spectrum energy is re-distributed to the mel scale conforming to the human ear hearing characteristics, through the discrete cosine transform technical means, the dimension redundancy and the high correlation of the mel spectrum feature are de-correlated and dimension compressed, and the low-dimensional mel frequency cepstral coefficient MFCC feature representing the sound channel characteristics and the excitation source separation is extracted.

[0069] S4.5, based on the impulse response data and the alarm audio segment, the speech transmission index value of the receiving point is calculated.

[0070] The expression of the speech transmission index value is: ; Wherein, is the speech transmission index value, is the number of octave band center frequencies, is the number of modulation frequencies, is the modulation frequency index, is the number of modulation frequencies, is the apparent signal-to-noise ratio under the condition of the combination of the th octave band center frequency and the th modulation frequency, is the reference signal-to-noise ratio, is the curve steepness factor, is the modulation frequency weight.

[0071] Further, by modulating the transmission ratio measurement and the psychoacoustic mapping technique, the physical acoustic parameters existing in the received point simulation signal generated by convolving the impulse response data and the alarm audio segment are subjected to multi-band modulation depth analysis and signal-to-noise ratio nonlinear conversion processing, and the linear time-invariant system characteristics of sound wave propagation are converted into speech transmission index values that conform to the human ear hearing perception characteristics.

[0072] It should be noted that the frequency band processing outer sum The human ear basilar membrane has selective response to different frequency sounds The formula divides the audio spectrum into B octave bands, independently analyzes the contribution of each band to intelligibility, and finally takes the average, which conforms to the human ear frequency domain analysis characteristics.

[0073] Modulation frequency perception inner sum Speech intelligibility depends on the degree of preservation of envelope modulation information, and the formula analyzes the signal-to-noise ratio of M modulation frequencies in each band, covering syllable, vocabulary and other time scale changes, and matching the sensitivity of the auditory nervous system to the timing structure.

[0074] Nonlinear signal-to-noise ratio mapping The human ear has a nonlinear threshold effect on signal-to-noise ratio, and the intelligibility decreases sharply at low signal-to-noise ratio, and the improvement is limited at high signal-to-noise ratio. The Sigmoid function takes the reference signal-to-noise ratio SNR0 as the inflection point, and the steepness factor controls the transition slope, accurately simulating this psychoacoustic characteristic.

[0075] Modulation frequency weighting weight Different modulation frequencies have different contributions to intelligibility, and the weight is set based on hearing experiment data, reflecting the importance difference of the modulation domain.

[0076] Specifically, the calculation of the traditional speech transmission index is based on the measurement of the modulation transfer function MTF, which evaluates the intelligibility by analyzing the influence of the acoustic channel on the envelope modulation depth of the test signal. The standard method needs to measure the signal-to-noise ratio at multiple octave bands and modulation frequencies, and then obtain the value through a series of conversions and weightings. The calculation process is complex and the physical meaning is not intuitive enough.

[0077] S4.6, integrate the speech transmission index value, semantic confusion probability and simulation parameters corresponding to each preliminary scheduling strategy to generate a simulation report of the intelligibility index.

[0078] Further, the integration process first collects the voice transmission index value and semantic confusion probability corresponding to each preliminary scheduling strategy, while recording the simulation parameters including the target audio playback device identifier, the broadcast volume parameter, the broadcast tone parameter and the repetition number parameter; then the voice transmission index value, the semantic confusion probability and the simulation parameters are structured and organized according to the strategy number, to generate a data record containing the strategy number, the voice transmission index value, the semantic confusion probability and the complete simulation parameter list; finally, all strategy data records are summarized into a comprehensibility index simulation report in table form.

[0079] S5. According to the comprehensibility index simulation report, the preliminary scheduling strategy is converted from the priority score to the quantifiable expected propagation effect comparison based on physical acoustic simulation, and the strategy parameters are fine-tuned and optimized according to the acoustic confusion risk assessment, and the preliminary scheduling strategy with the optimal comprehensibility index and the lowest ambiguity risk is selected as the final scheduling execution instruction.

[0080] S5.1, analyze the simulation report to extract the voice transmission index value corresponding to each preliminary scheduling strategy, and generate a mapping list.

[0081] Further, through structured data extraction and key-value mapping technology means, the strategy number matching and field alignment processing are performed on the correlation missing defects of the discrete distributed voice transmission index value and strategy identifier in the simulation report, and the unstructured simulation result data is converted into a mapping list corresponding to the strategy number and the voice transmission index value.

[0082] S5.2, calculate the device resource consumption coefficient based on the preliminary scheduling strategy parameters in the mapping list; The device resource consumption coefficient expression is: ; Wherein, is the device resource consumption coefficient, is the strategy recommended volume, is the maximum available volume of the device, is the alarm audio file, is the upper limit of the device cache, is the strategy recommended repetition broadcast number, is the maximum allowed repetition number, is the strategy recommended volume weight, is the audio file weight, is the strategy recommended repetition broadcast number weight.

[0083] Further, the policy recommendation volume, the alarm audio file size, and the policy recommendation repetition number of times are extracted from the mapping list, and three device capability parameters, i.e., the maximum available volume of the device, the upper limit of the device cache, and the maximum allowed repetition number of times, are obtained; then the policy recommendation volume is divided by the maximum available volume of the device to obtain a volume ratio, the alarm audio file size is divided by the upper limit of the device cache to obtain a file size ratio, and the policy recommendation repetition number of times is divided by the maximum allowed repetition number of times to obtain a repetition number of times ratio; finally, the volume ratio is multiplied by the policy recommendation volume weight, the file size ratio is multiplied by the alarm audio file weight, and the repetition number of times ratio is multiplied by the policy recommendation repetition number of times weight, and the three weighted results are added to obtain the device resource consumption coefficient.

[0084] It should be noted that the power operation simulates the non-linear loss , the volume power consumption : the sound power is proportional to the square of the voltage, and the power consumption increases super-linearly with the increase of the volume, which accurately reflects the physical constraints of the electro-acoustic conversion efficiency.

[0085] File processing overhead Large file transmission involves protocol overhead, cache fragmentation, and other non-linear costs, reflecting the diminishing marginal utility of storage and bandwidth resources.

[0086] Time resource occupation , repeated broadcasting occupies the time window of the device, which may block other tasks and increase the risk of user disturbance, reflecting the cumulative negative effects of repeated behavior.

[0087] Specifically, the traditional device resource evaluation usually adopts a linear weighting model, i.e., the volume, file size, and repetition number of times are linearly superimposed into a comprehensive score by fixed weights. This method ignores the non-linear characteristics of resource consumption and the coupling effect between dimensions, resulting in a deviation between the evaluation result and the real device load.

[0088] S5.3, query the target audio playback device identifier specified by the policy in the mapping list, obtain the device reliability coefficient, and calculate the execution risk coefficient.

[0089] The execution risk coefficient expression is: ; Among them, is the execution risk coefficient, is the basic execution risk coefficient, is the device reliability threshold, is the ambient noise value, is the ambient noise threshold, is the volume redundancy, is the minimum required signal-to-noise ratio, is the basic execution risk coefficient weight, The environmental noise value weight, The volume redundancy weight.

[0090] Further, the target audio playback device identifier specified in the policy in the query mapping list is queried for the corresponding device reliability coefficient, and the current environmental noise value and the policy recommended volume parameter are obtained; the volume redundancy is calculated as the policy recommended volume minus the environmental noise value; the device reliability coefficient and the environmental noise value are compared with the device reliability threshold and the environmental noise threshold, respectively, to obtain a basic execution risk coefficient; the volume redundancy is compared with the minimum required signal-to-noise ratio to obtain a volume redundancy risk term; and finally, the basic execution risk coefficient is multiplied by the basic execution risk coefficient weight, the environmental noise value is multiplied by the environmental noise value weight, and the volume redundancy risk term is multiplied by the volume redundancy weight, to obtain an execution risk coefficient.

[0091] It should be noted that the non-linear amplification power operation of the risk factor The device reliability risk The device reliability coefficient The lower the device reliability risk, the higher the risk of execution failure. When , the ratio 1- 1- > 1, the index > 1 significantly amplifies the risk, reflecting the accelerated rise in failure probability caused by the deterioration of the device state.

[0092] The environmental interference risk The environmental noise value When the environmental noise value exceeds the environmental noise threshold , the ratio is greater than 1, and the index > 1 amplifies the negative impact of noise on audio masking, in line with the Weber-Fechner law in acoustics.

[0093] The task failure risk , the volume redundancy When the volume redundancy is less than the minimum required signal-to-noise ratio , the ratio < 1, but the index < 0 will make the value of this term greater than 1, inversely amplifying the risk and directly simulating the physical law of sound propagation that the probability of failure rises sharply when the signal-to-noise ratio is insufficient.

[0094] Specifically, traditional execution risk assessment usually adopts a linear weighted scoring card model, i.e., by setting fixed weights based on expert experience, the device state, environmental parameters, and other risk indicators are linearly superimposed into a risk score. This method assumes that risk factors are independent and have a linear impact, ignoring the non-linear threshold effect and multi-factor coupling relationship of actual risk transmission, resulting in a deviation between the risk assessment and the true failure probability.

[0095] ​S5.4, the voice transmission index value, semantic confusion probability, device resource consumption coefficient and execution risk coefficient are substituted into the comprehensive performance score formula, and the comprehensive performance score of each strategy is calculated.

[0096] The comprehensive performance score expression is: ; Wherein, is the comprehensive performance score, is, is the intelligibility utility saturation threshold, is the intelligibility benefit amplification index, is the maximum tolerable confusion threshold, is the semantic confusion probability, is the execution risk realization probability.

[0097] Further, through multi-factor nonlinear coupling and psychoacoustic utility mapping technology means, the Sigmoid function thresholding processing is performed on the perceptual saturation characteristics of the voice transmission index value, the linear intelligibility index is converted into a utility value conforming to the marginal effect of human ear hearing, the logistic function conversion is performed on the risk mutation characteristics of the semantic confusion probability, the maximum tolerable confusion threshold is taken as the critical point to realize the nonlinear penalty of semantic fidelity, the exponential decay modeling is performed on the rigidity constraint neglecting of the device resource consumption coefficient, the ratio of resource utilization to maximum tolerable value is converted into an efficiency reduction factor, the Weber distribution function fitting is performed on the probabilistic expression deficiency of the execution risk coefficient, the risk level is mapped into the task failure probability and converted into the success probability complement, and the four types of parameters are integrated into a unified comprehensive performance score through the multiplication coupling mechanism, so that the problem that the traditional linear weighted model cannot reflect the nonlinear interaction and coupling amplification between dimensions is solved.

[0098] It should be noted that the nonlinear saturation of intelligibility benefit power operation , simulates the initial high benefit of intelligibility improvement, when is low, the intelligibility improvement is significant, reflecting the sensitivity of the auditory system to intelligibility improvement, and the Sigmoid saturation function : defines the intelligibility utility saturation point, when exceeds the threshold is the intelligibility utility saturation threshold, and further improvement of the perceived utility decreases, which conforms to the Weber-Fechner law of psychoacoustics.

[0099] Threshold penalty of semantic fidelity , Sigmoid decay function, with the maximum tolerable confusion threshold as the critical point, when > When the function value sharply drops, the semantic confusion probability exceeds the limit, which directly leads to the collapse of the performance score, reflecting the physical constraint that accuracy is prioritized over clarity in information transmission.

[0100] Exponential decay of resource consumption Exponential decay model: resource consumption coefficient Approaching the maximum tolerance value When the performance decreases exponentially with resource consumption, it reflects the rigid constraints of device resource bottlenecks.

[0101] Complement compensation of risk realization probability Risk realization probability The execution risk coefficient is mapped to the task failure probability, and the complement directly represents the task success probability, making the performance score physically associated with risk, in line with the principles of probability risk analysis in reliability engineering.

[0102] S5.5, select the preliminary scheduling strategy with the highest comprehensive performance score as the final scheduling execution instruction.

[0103] Furthermore, compare the comprehensive performance score values of all preliminary scheduling strategies, and select the preliminary scheduling strategy with the highest comprehensive performance score value as the final scheduling execution instruction. The scheduling execution instruction contains a complete set of device identifiers and broadcast parameters.

[0104] S6, issue the final scheduling execution instruction to the target audio playback device specified in the instruction.

[0105] S6.1, encapsulate the final scheduling execution instruction into a device-identifiable standard control protocol data packet, and send the control protocol data packet to the network endpoint corresponding to the target audio playback device identifier through a secure transmission layer protocol encrypted channel.

[0106] Furthermore, the device identifier, broadcast audio file uniform resource identifier, volume parameter value, tone parameter value, and repetition number parameter value contained in the final scheduling execution instruction are structured and encapsulated in JSON format to generate a device-identifiable standard control protocol data packet. The encapsulated control protocol data packet is sent to the network endpoint address registered in the device management platform through a secure transmission layer protocol encrypted channel.

[0107] S6.2, the target audio playback device receives and parses the control protocol data packet to extract the broadcast parameters.

[0108] Further, the target audio playing device receives the control protocol data packet through the network interface, decrypts and parses the data packet, and extracts the broadcast audio file uniform resource identifier, volume parameter value, tone parameter value, and repetition number parameter value.

[0109] S6.3, the target audio playing device executes the local audio broadcast operation according to the broadcast parameters, and the scheduling engine returns a status code to execute a confirmation response.

[0110] Further, the target audio playing device executes the local audio broadcast operation according to the broadcast parameters obtained by parsing, and the operation includes downloading the audio data file corresponding to the broadcast audio file uniform resource identifier from the storage service, processing the audio data into the specified tone parameter value, outputting through the physical loudspeaker at the specified volume parameter value, and repeating the specified repetition number parameter value. After the operation is completed, an execution confirmation response containing a success or failure status code is returned to the scheduling engine.

[0111] The embodiment also provides an Internet of Things alarm audio call scheduling system, which comprises: a fusion module that receives an alarm request message from an Internet of Things sensing terminal, collects and generates a multi-dimensional information report of real-time environmental noise data, audio playing device state information, and user state information according to the device identifier and the area code in the alarm request message; a dynamic priority module that generates a dynamic priority score of the alarm request message by querying a preset strategy library according to the multi-dimensional information report; a preliminary scheduling strategy generation module that sorts the dynamic priority scores of the alarm request messages to be processed, and generates a preliminary scheduling strategy of the suggested playing device and broadcast parameters for the alarm request message with the highest priority; a simulation report module that inputs the preliminary scheduling strategy into an acoustic digital twin model of the alarm occurrence area, and performs high-fidelity acoustic simulation and ambiguity analysis on the preliminary scheduling strategy through the digital twin from static environment modeling to dynamic scheduling strategy prediction and optimization engine to generate a simulation report of intelligibility indicators; a decision module that converts the preliminary scheduling strategy from the priority score to a quantifiable expected propagation effect comparison based on physical acoustic simulation according to the simulation report of intelligibility indicators, fine-tunes and optimizes the strategy parameters according to the acoustic confusion risk assessment, selects the preliminary scheduling strategy with the optimal intelligibility indicators and the lowest ambiguity risk as the final scheduling execution instruction, and issues the final scheduling execution instruction to the target audio playing device.

[0112] The embodiment also provides a computer device suitable for the method of Internet of Things alarm audio call scheduling, which comprises a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions to realize the method of Internet of Things alarm audio call scheduling proposed in the above embodiment.

[0113] The computer device can be a terminal, which includes a processor, a memory, a communication interface, a display screen and an input device connected by a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The communication interface of the computer device is configured to communicate with external terminals in a wired or wireless manner. The wireless manner can be achieved by WIFI, an operator network, NFC (Near Field Communication) or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0114] The embodiment also provides a storage medium having a computer program stored thereon, the program being executed by a processor to implement the method for scheduling an alarm audio call of an Internet of Things as described in the above embodiment. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, a magnetic disk or an optical disk.

[0115] To sum up, the application receives an alarm request message of an Internet of Things sensing terminal, generates a multi-dimensional information report based on device identification and regional coding to collect environmental noise, device state and user state information, and generates a dynamic priority score by using weighted calculation. After sorting the alarm request, a preliminary scheduling strategy is generated for the highest priority alarm, a high-fidelity simulation is generated by using an acoustic digital twin model to generate an intelligibility index report, an optimal strategy is selected based on the simulation result to generate an execution instruction and is delivered to a target audio device, and the whole process of intelligent scheduling from multi-dimensional dynamic sensing to acoustic effect optimization is realized.

[0116] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced, without departing from the spirit and scope of the technical solutions of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. A method for Internet of Things alert audio call dispatching, the method comprising: The method comprises the following steps: receiving an alarm request message from an Internet of Things sensing terminal; collecting and generating a multi-dimensional information report of real-time environmental noise data, audio playing device state information and user state information according to the device identifier and the alarm occurrence area code in the alarm request message; ​ According to the multi-dimensional information report, the dynamic priority score of the alarm request message is generated by querying the preset strategy library, which comprises the following steps: Read the environmental noise value in the multi-dimensional information report and directly convert it into an environmental noise coefficient; Extract the device state data, collect the device state data through multi-dimensional device health indicators, generate a dynamic health score by using weighted fusion, establish a device health score library, and directly obtain a device reliability coefficient; Parse the user activity information into a user state emergency coefficient, generate a basic priority coefficient from the alarm type extracted from the alarm request message, obtain the time stamp in the alarm request message and derive the time difference from the current time, and derive the time attenuation factor through an exponential decay function; 2. The method of IoT alert audio call dispatching of claim 1, wherein: Query the log to count the number of associated events to generate an associated event enhancement factor; Integrate the basic priority coefficient, the environmental noise coefficient, the user state emergency coefficient, the device reliability coefficient, the time attenuation factor and the associated event enhancement factor to generate the dynamic priority score. The method comprises the following steps: receiving an alarm request message from an Internet of Things sensing terminal; collecting and generating a multi-dimensional information report of real-time environmental noise data, audio playing device state information and user state information according to the device identifier and the alarm occurrence area code in the alarm request message; According to the multi-dimensional information report, the dynamic priority score of the alarm request message is generated by querying the preset strategy library, which comprises the following steps:

3. The method of IoT alert audio call dispatching of claim 2, wherein: Read the environmental noise value in the multi-dimensional information report and directly convert it into an environmental noise coefficient; Extract the device state data, collect the device state data through multi-dimensional device health indicators, generate a dynamic health score by using weighted fusion, establish a device health score library, and directly obtain a device reliability coefficient; Parse the user activity information into a user state emergency coefficient, generate a basic priority coefficient from the alarm type extracted from the alarm request message, obtain the time stamp in the alarm request message and derive the time difference from the current time, and derive the time attenuation factor through an exponential decay function; Query the log to count the number of associated events to generate an associated event enhancement factor; Integrate the basic priority coefficient, the environmental noise coefficient, the user state emergency coefficient, the device reliability coefficient, the time attenuation factor and the associated event enhancement factor to generate the dynamic priority score. ​ 4. The method of IoT alert audio call dispatching of claim 3, wherein: The dynamic priority score of the alarm request message to be processed is sorted, and a preliminary scheduling strategy of the recommended playing device and the broadcast parameter is generated for the alarm request message with the highest priority, including the following steps: The alarm request messages to be processed are sorted according to the dynamic priority score from high to low, and an ordered sorted alarm request queue is generated, and the alarm request message at the first position in the sorted alarm request queue is selected to obtain the highest priority alarm to be scheduled; According to the alarm occurrence area code in the alarm request message, the online audio playing device registered in the device management platform is queried to obtain the candidate audio playing device list and the candidate audio playing device list for alarm broadcast; For each audio playing device in the candidate audio playing device list, the device-environment matching degree score is filtered out; The expected broadcast utility value of each audio playing device is generated by combining the device-environment matching degree score, the dynamic priority score of the alarm request message and the environmental noise data in the multi-dimensional information report, and the preliminary scheduling strategy of the recommended volume, tone and repetition number is generated for each device.

5. The method of IoT alert audio call dispatching of claim 4, wherein: The preliminary scheduling strategy is input into the acoustic digital twin model of the alarm occurrence area, and the digital twin is transformed from static environment modeling to dynamic scheduling strategy prediction and optimization engine, the preliminary scheduling strategy is subjected to high-fidelity acoustic simulation and ambiguity analysis, and a simulation report of intelligibility index is generated, including the following steps: Based on the physical space three-dimensional geometric data corresponding to the alarm occurrence area code, the interface material acoustic parameters and the real-time environmental noise data, the acoustic digital twin model is obtained by the finite element method; Each strategy in the preliminary scheduling strategy is converted into an acoustic simulation parameter, the point sound source is set in the acoustic digital twin model according to the target audio playing device position in each strategy in the preliminary scheduling strategy, the sound power level is calculated according to the broadcast volume parameter in the strategy, the alarm audio segment to be broadcast is loaded, and the sound source signal is obtained; A virtual microphone array is deployed as a receiving point in the user activity area in the acoustic digital twin model, the sound wave propagation, reflection and absorption process in the acoustic digital twin model corresponding to each strategy is calculated in parallel by using the ray tracing method, and the impulse response data at the receiving point is obtained; Based on the impulse response data and the alarm audio segment, the actual audio signal at the receiving point is simulated, the mel frequency cepstral coefficient (MFCC) feature is extracted by performing time-frequency analysis on the simulated actual audio signal; Based on the impulse response data and the alarm audio segment, the speech transmission index value of the receiving point is calculated; The speech transmission index value, the semantic confusion probability and the simulation parameters corresponding to each preliminary scheduling strategy are integrated to generate a simulation report of intelligibility index.

6. The method of IoT alert audio call dispatching of claim 5, wherein: According to the simulation report of intelligibility index, the preliminary scheduling strategy is converted from the priority score to the quantifiable expected propagation effect comparison based on physical acoustic simulation, and the strategy parameters are fine-tuned and optimized according to the acoustic confusion risk assessment, the preliminary scheduling strategy with the optimal intelligibility index and the lowest ambiguity risk is selected as the final scheduling execution instruction, including the following steps: The simulation report extracts the voice transmission index value corresponding to each preliminary scheduling strategy, and generates a mapping list; Based on the preliminary scheduling strategy parameters in the mapping list, the device resource consumption coefficient is calculated; Query the target audio playback device identifier specified in the strategy in the mapping list, obtain the device reliability coefficient and calculate the execution risk coefficient; Substitute the voice transmission index value, semantic confusion probability, device resource consumption coefficient and execution risk coefficient into the comprehensive performance score formula to calculate the comprehensive performance score of each strategy; Select the preliminary scheduling strategy with the highest comprehensive performance score and set it as the final scheduling execution instruction.

7. The method of IoT alert audio call dispatching of claim 6, wherein: The final scheduling execution instruction is sent to the target audio playback device, including the following steps: Package the final scheduling execution instruction into a standard control protocol data packet recognizable by the device, and send the control protocol data packet to the network endpoint corresponding to the target audio playback device identifier through a secure transmission layer protocol encryption channel; The target audio playback device receives and parses the control protocol data packet to extract the broadcast parameters; The target audio playback device executes local audio broadcast operation according to the broadcast parameters, and the scheduling engine returns a status code execution confirmation response.

8. A system for dispatching an alarm audio call for an Internet of Things based on the method for dispatching an alarm audio call for an Internet of Things according to any one of claims 1 to 7, characterized in that: The fusion module receives the alarm request message from the Internet of Things sensing terminal, collects and generates a multi-dimensional information report of real-time environmental noise data, audio playback device state information and user state information according to the device identifier and area code in the alarm request message; The dynamic priority module generates a dynamic priority score for the alarm request message by querying the preset strategy library based on the multi-dimensional information report; The preliminary scheduling strategy generation module sorts the dynamic priority scores of the alarm request messages to be processed, and generates a preliminary scheduling strategy of the recommended playback device and broadcast parameters for the alarm request message with the highest priority; The simulation report module inputs the preliminary scheduling strategy into the acoustic digital twin model of the alarm occurrence area, and through digital twinning, the static environment modeling is converted into a dynamic scheduling strategy prediction and optimization engine for high-fidelity acoustic simulation and ambiguity analysis, generating a simulation report of intelligibility indicators; The decision module converts the preliminary scheduling strategy from the priority score to a quantifiable expected propagation effect comparison based on physical acoustic simulation according to the simulation report of intelligibility indicators, and fine-tunes and optimizes the strategy parameters according to the acoustic confusion risk assessment, selects the preliminary scheduling strategy with the optimal intelligibility indicator and the lowest ambiguity risk, and sets it as the final scheduling execution instruction. The final scheduling execution instruction is sent to the target audio playback device. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is characterized in that: The processor executes the computer program to realize the steps of the method of the Internet of Things alarm audio call scheduling of any one of claims 1-7.

10. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by the processor to realize the steps of the method of the Internet of Things alarm audio call scheduling of any one of claims 1-7.

Images