A machine learning based method and system for identifying anomalous data

Abstract

The embodiment of the application provides a kind of based on machine learning's abnormal data identification method and system. Wherein, first receive user and early warning system triggered emergency event signal;Utilize situational awareness algorithm to analyze these parameters, match the personalized broadcast content suitable for specific geographic location, and adjust the applicability and priority order of emergency response plan library, form emergency response strategy;According to this strategy, integrate text, voice and visual signal to build multimodal early warning information;Optimal transmission path is calculated by path planning algorithm, and information is processed using adaptive modulation and coding technology to generate transmission scheme;After terminal receives information, through modal selection algorithm adaptation and select appropriate presentation mode, introduce augmented reality technology to support visual information presentation, collect user feedback data, evaluate communication effect.The technical scheme provided by the embodiment of the application improves the speed, efficiency and accuracy of emergency information transmission.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a method and system for identifying abnormal data based on machine learning. Background Technology

[0002] In modern emergency management systems, it is crucial to quickly and accurately transmit emergency information; especially in complex and ever-changing scenarios such as natural disasters, public health emergencies, and industrial accidents, timely issuance of early warning information can effectively reduce casualties and property losses.

[0003] Currently, most emergency communication systems rely on traditional single-modal information dissemination methods, such as SMS, radio, or television notifications. These systems typically have a pre-set database of emergency response plans, and when an emergency signal is received, they directly select the corresponding warning information from the database for dissemination. In addition, some more advanced systems have begun to introduce path planning algorithms to optimize information transmission paths and adopt adaptive modulation and coding techniques to improve transmission efficiency.

[0004] Existing emergency communication solutions suffer from several shortcomings: First, single-modal information dissemination methods cannot meet the diverse needs of different users, especially in complex and ever-changing emergency environments, where a single information format makes it difficult to ensure that all users can quickly understand and react. Second, traditional systems lack context awareness and cannot dynamically adjust the content and priority of warning information based on real-time environmental parameters, resulting in insufficient targeting and effectiveness of information delivery. Finally, while existing path planning and coding technologies improve transmission efficiency, they are still insufficient in terms of information presentation and user feedback, lacking interactive mechanisms with users and unable to assess the effectiveness of information delivery in real time and adjust the transmission range accordingly. These problems limit the overall performance of emergency communication systems and affect the speed and accuracy of emergency response. Summary of the Invention

[0005] This application provides a machine learning-based method and system for identifying abnormal data, which addresses the problems of slow speed, low efficiency, and low accuracy in emergency information transmission in the prior art.

[0006] In a first aspect, embodiments of this application provide a method for identifying abnormal data based on machine learning, including:

[0007] Receive emergency event signals triggered by users and the early warning information release system, wherein the emergency event signals carry event type information, geographical location information, and dynamic environmental parameters;

[0008] The dynamic environmental parameters are analyzed using a context-aware algorithm to obtain abnormal data identification results. Based on the event type information and the abnormal data identification results, personalized broadcast content suitable for the geographical location information is matched from a preset emergency response plan library. The applicability and priority order of the emergency response plan library are adjusted to obtain an emergency response strategy.

[0009] Based on the aforementioned emergency response strategy, text, voice, and visual signals from the emergency response plan database are integrated to construct multimodal early warning information, and the multimodal early warning information is optimized according to specific scenarios.

[0010] Based on the characteristics of the multimodal early warning information and the network status of the multimodal early warning broadcasting terminals in the affected area, the optimal transmission path is calculated using a path planning algorithm, and the multimodal early warning information is encoded and processed using adaptive modulation and coding technology. Based on the processed multimodal early warning information and the optimal transmission path, a transmission scheme is generated, and the processed multimodal early warning information is sent to the selected multimodal early warning broadcasting terminals through the transmission scheme.

[0011] After the multimodal warning broadcasting terminal acquires the processed multimodal warning information, it adapts and selects appropriate multimodal warning information based on the modality selection algorithm for visual information presentation, and introduces augmented reality technology to support the visual information presentation to obtain feedback data, wherein the feedback data includes user understanding and reaction speed.

[0012] Collect and analyze the feedback data, evaluate the transmission effect of the multimodal early warning information, and adjust the transmission range of the multimodal early warning information based on the transmission effect.

[0013] Optionally, the step of calculating the optimal transmission path using a path planning algorithm based on the characteristics of the multimodal early warning information and the network status of the multimodal early warning broadcasting terminals in the affected area, encoding the multimodal early warning information using adaptive modulation and coding technology, and generating a transmission scheme based on the processed multimodal early warning information and the optimal transmission path includes:

[0014] The information entropy algorithm is used to evaluate the content complexity of the multimodal early warning information, and the characteristics of the multimodal early warning information are analyzed in combination with the data volume and urgency of the multimodal early warning information. The priority, required bandwidth and redundancy strategy of the transmission of the multimodal early warning information are determined to obtain a transmission optimization scheme. The transmission optimization scheme is weighted by the quality of service parameters to obtain the information transmission strategy.

[0015] Based on the information transmission strategy, machine learning algorithms are applied to monitor the status of the multimodal early warning broadcasting terminals in the affected area in real time, predict the changing trend of the multimodal early warning broadcasting terminals, and combine deep learning models to identify network congestion points and high-latency areas of the multimodal early warning broadcasting terminals, form a network topology map, adjust the sudden situation of the multimodal early warning broadcasting terminal topology map, and generate a dynamic network status map.

[0016] Based on the dynamic network state diagram and the information transmission strategy, the Dijkstra algorithm or A* search algorithm in the path planning algorithm is used to calculate the transmission path from the early warning information release system to the selected multimodal early warning broadcast terminal, and a genetic algorithm is introduced to optimize the transmission path to obtain the optimal transmission path.

[0017] Based on the segment characteristics of the multimodal early warning broadcasting terminal on the optimal transmission path, adaptive modulation and coding technology and channel state information feedback mechanism are used to encode and optimize the multimodal early warning information to generate optimized multimodal early warning information;

[0018] The operation instructions, transmission time window, and error retransmission mechanism for sending the optimized multimodal early warning information are obtained by sending the selected multimodal early warning broadcasting terminal;

[0019] Based on the transmission instruction set, a transmission scheme document is compiled, and a risk assessment and countermeasures are added to the transmission scheme document to obtain the transmission scheme. The transmission scheme document includes: the optimal transmission path, encoding method, transmission instructions, and emergency response plan.

[0020] Optionally, the step of using the information entropy algorithm to evaluate the content complexity of the multimodal early warning information, and combining the data volume and urgency of the multimodal early warning information to analyze its characteristics, determines the transmission priority, required bandwidth, and redundancy strategy of the multimodal early warning information, and obtains a transmission optimization scheme, includes:

[0021] The multimodal early warning information is processed using the information entropy algorithm complexity assessment to obtain a quantitative result of the content complexity;

[0022] The quantitative results of the content complexity are analyzed using the data volume and urgency of the multimodal early warning information to obtain a specific characteristic description of each multimodal early warning information.

[0023] Based on the specific characteristic description, the priority processing of the multimodal early warning information transmission is determined, and an information transmission priority list is generated;

[0024] Based on the information transmission priority list, the required bandwidth is estimated and the minimum bandwidth requirement scheme is obtained.

[0025] For the minimum bandwidth requirement scheme, a redundancy strategy is formulated to ensure reliability and generate a redundant transmission plan; the multimodal early warning information will be transmitted through other redundancy mechanisms to cope with network failures and congestion, and a transmission optimization scheme will be formed based on the redundant transmission plan.

[0026] Optionally, the step of calculating the transmission path from the early warning information dissemination system to the selected multimodal early warning broadcasting terminal based on the dynamic network state diagram and the information transmission strategy, using Dijkstra's algorithm or A* search algorithm in path planning algorithms, and introducing a genetic algorithm to optimize the transmission path to obtain the optimal transmission path, includes:

[0027] The status of the multimodal early warning broadcasting terminal network within the affected area is comprehensively evaluated using the dynamic network status diagram and the information transmission strategy to obtain basic data;

[0028] Based on the aforementioned basic data, the path from the early warning information release system to the selected multimodal early warning broadcasting terminal is calculated using the Dijkstra algorithm or A* search algorithm in the path planning algorithm, and a preliminary transmission path is generated.

[0029] A genetic algorithm is introduced to iteratively optimize the initial transmission path, select the transmission path with the best performance, and generate candidate transmission paths.

[0030] By evaluating the performance of the candidate transmission paths under network congestion points, high latency areas, and sudden events, the stability and anti-interference capabilities of the candidate transmission paths are analyzed, and important transmission paths with high efficiency and stability are obtained.

[0031] Based on the aforementioned important transmission paths, and considering the network topology, the status of the multimodal early warning broadcasting terminals, the urgency of the event, and the information transmission priority, the optimal transmission path is generated.

[0032] Optionally, it also includes:

[0033] The transmission optimization scheme is weighted and adjusted using the service quality parameters. A fault prediction and health management system is introduced to obtain an information transmission strategy and monitor the health status during transmission in real time. When an anomaly is detected, the health management system can respond quickly, take preventive measures, monitor potential risks during transmission, and automatically adjust the information transmission strategy.

[0034] Optionally, monitoring potential risks during the transmission process and automatically adjusting the information transmission strategy includes:

[0035] Continuously monitor each stage of the transmission execution plan, collect and analyze network performance indicators and terminal feedback data to obtain potential risk factors. Among them, the multimodal early warning broadcast terminal performance indicators include latency, packet loss rate and bandwidth utilization.

[0036] Machine learning algorithms are used to predict the probability of occurrence and the scope of impact of potential risks in the terminal feedback data, provide early warning of fault points or bottlenecks, and generate risk assessment reports.

[0037] The information transmission strategy is automatically adjusted based on the risk assessment report. After high-risk areas are detected, the transmission path is replanned or redundant transmission channels are added. Service quality parameters are dynamically adjusted to prioritize the transmission of the multimodal early warning information.

[0038] Optionally, the step of using a context-aware algorithm to analyze the dynamic environmental parameters to obtain abnormal data identification results, and matching personalized broadcast content suitable for the geographical location information from a preset emergency response plan database based on the event type information and the abnormal data identification results, includes:

[0039] Receive emergency event signals triggered by users and the early warning information release system, extract and parse the event type information, geographical location information and dynamic environmental parameters in the emergency event signals, wherein the dynamic environmental parameters include weather conditions, traffic flow and population density;

[0040] The dynamic environmental parameters are analyzed using a context-aware algorithm to assess the impact of current environmental conditions on emergency response and obtain environmental anomaly data identification results.

[0041] Based on the event type information and the environmental anomaly data identification results, personalized broadcast content suitable for the geographical location information is matched from the preset emergency response plan library, and the emergency response plan library is filtered to obtain personalized broadcast content.

[0042] Optionally, the Dijkstra algorithm or A* search algorithm in the application path planning algorithm calculates the transmission path from the early warning information dissemination system to the selected multimodal early warning broadcasting terminal, including:

[0043] The dynamic network state diagram and the information transmission strategy are used to comprehensively evaluate the status of the multimodal early warning broadcasting terminal network in the affected area to obtain basic data. This basic data is then used to calculate node reliability. The node reliability is obtained; It is obtained by calculation using the following formula: in, Represents a node Reliability, Represents a node transmission rate Represents a node The delay time, Represents a node The degree of congestion, These are weight parameters that are adaptively adjusted based on the network status of the multimodal early warning broadcasting terminal. These are the adjustment coefficients that affect the index;

[0044] Using Dijkstra's algorithm or A* search algorithm in combination to assess node reliability The calculated basic data is used to calculate the preliminary transmission path, which is obtained by calculating each path. Total cost Obtain; wherein, the total cost It is obtained by calculation using the following formula: in, Represent each path Total cost , Representing a path upper node To the node The distance between them Represents a node To the node The delay between Indicates the maximum possible distance. These are the adjustment coefficients that affect the index. Representing a path The total number of nodes on;

[0045] Introducing a genetic algorithm to iteratively optimize each path Total cost and the reliability of the nodes The optimal transmission path with the best performance is selected by evaluating each path. The fitness function is calculated based on the performance. Obtain; where, fitness function It is obtained by calculation using the following formula: in, Representing a path Adaptability, These are the weighting parameters for total cost, average node reliability, and priority, respectively. These are the adjustment coefficients that affect the index. Additional weighted values ​​indicating the urgency of the event and the priority of information transmission.

[0046] Secondly, embodiments of this application provide an anomaly data identification system based on machine learning, comprising:

[0047] The receiving module is used to receive emergency event signals triggered by users and the early warning information release system, wherein the emergency event signals carry event type information, geographical location information, and dynamic environmental parameters;

[0048] The analysis module uses a context-aware algorithm to analyze the dynamic environmental parameters, obtain abnormal data identification results, and matches personalized broadcast content suitable for the geographical location information from the preset emergency response plan library based on the event type information and the abnormal data identification results, and adjusts the applicability and priority order of the emergency response plan library to obtain an emergency response strategy.

[0049] The module is used to integrate text, voice, and visual signals from the emergency response plan library based on the emergency response strategy, construct multimodal early warning information, and optimize the multimodal early warning information according to specific scenarios.

[0050] The calculation module is used to calculate the optimal transmission path using a path planning algorithm based on the characteristics of the multimodal early warning information and the network status of the multimodal early warning broadcasting terminals in the affected area, encode and process the multimodal early warning information using adaptive modulation and coding technology, generate a transmission scheme based on the processed multimodal early warning information and the optimal transmission path, and send the processed multimodal early warning information to the selected multimodal early warning broadcasting terminals through the transmission scheme.

[0051] The feedback module is used to, after the multimodal warning information is acquired and processed by the multimodal warning broadcasting terminal, adapt and process it based on the modality selection algorithm and select appropriate multimodal warning information for visual information presentation, and introduce augmented reality technology to support the visual information presentation to obtain feedback data, wherein the feedback data includes user understanding and reaction speed.

[0052] An evaluation module is used to collect and analyze the feedback data, evaluate the transmission effect of the multimodal early warning information, and adjust the transmission range of the multimodal early warning information based on the transmission effect.

[0053] Thirdly, embodiments of this application provide a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement a machine learning-based abnormal data identification method as described in the first aspect above.

[0054] Fourthly, embodiments of this application provide a computer storage medium storing a computer program, which, when executed by a computer, implements a machine learning-based abnormal data identification method as described in the first aspect.

[0055] In this embodiment, an emergency event signal triggered by a user and an early warning information dissemination system is received. The emergency event signal carries event type information, geographic location information, and dynamic environmental parameters. A context-aware algorithm is used to analyze the dynamic environmental parameters to obtain abnormal data identification results. Based on the event type information and the abnormal data identification results, personalized broadcast content suitable for the geographic location information is matched from a preset emergency response plan library. The applicability and priority order of the emergency response plan library are adjusted to obtain an emergency response strategy. Based on the emergency response strategy, text, voice, and visual signals from the emergency response plan library are integrated to construct multimodal early warning information. This multimodal early warning information is then optimized according to specific scenarios. Finally, based on the characteristics of the multimodal early warning information and the multimodal early warning broadcasting terminals within the affected area… Based on the network conditions, the optimal transmission path is calculated using a path planning algorithm. Adaptive modulation and coding techniques are used to encode the multimodal warning information. A transmission scheme is generated based on the processed multimodal warning information and the optimal transmission path. The processed multimodal warning information is then sent to a selected multimodal warning broadcasting terminal via this scheme. After the multimodal warning broadcasting terminal receives the processed multimodal warning information, it adapts and selects appropriate multimodal warning information for visual presentation based on a modality selection algorithm. Augmented reality technology is introduced to support the visual presentation, resulting in feedback data, including user understanding and reaction speed. The feedback data is collected and analyzed to evaluate the transmission effect of the multimodal warning information, and the transmission range of the multimodal warning information is adjusted based on the transmission effect.

[0056] The technical solution of this application has the following beneficial effects:

[0057] By rapidly analyzing dynamic environmental parameters and matching personalized broadcast content, information preparation time is shortened; multimodal early warning information combining text, voice, and visual signals optimizes user experience and improves user comprehension and reaction speed; path planning algorithms and adaptive modulation and coding techniques ensure efficient and stable information transmission in complex network environments; adjusting the transmission range of multimodal early warning information based on real-time feedback data enhances the system's flexibility and adaptability; the introduction of augmented reality technology to support visual information presentation increases the intuitiveness and interactivity of information, further improving user engagement and response speed; and dynamically adjusting the information transmission range by evaluating the delivery effect and rationally allocating network resources improves the overall system performance and resource utilization.

[0058] Furthermore, by evaluating the characteristics of multimodal early warning information and the network conditions within the affected area, the optimal transmission path is calculated using a path planning algorithm, and adaptive modulation and coding technology is employed to process the early warning information, generating a transmission scheme. Specifically, this includes: using an information entropy algorithm to assess information complexity, determining transmission priority, required bandwidth, and redundancy strategies based on data volume and urgency, forming a transmission optimization scheme, and obtaining an information transmission strategy through weighted processing using quality of service parameters; based on this strategy, applying machine learning and deep learning models to monitor and predict terminal conditions in real time, identifying network congestion points and high-latency areas, and generating a dynamic network state diagram; according to the dynamic network state diagram and transmission strategy, using Dijkstra's or A* search algorithms to calculate and optimize the transmission path, ensuring optimal path selection; considering the characteristics of terminals on the optimal path, using adaptive modulation and coding technology and a channel state feedback mechanism to optimize the early warning information, generating optimized multimodal early warning information; finally, by selecting terminal operation instructions for sending optimized information, transmission time windows, and error retransmission mechanisms, a transmission instruction set is compiled, and risk assessment and countermeasures are supplemented to the transmission scheme document to ensure the comprehensiveness and reliability of the transmission scheme.

[0059] The above methods significantly improve the transmission efficiency and reliability of multimodal early warning information in complex network environments. By comprehensively assessing information complexity, network conditions, and event urgency, transmission priorities and required resources are accurately determined, ensuring that critical information can be delivered to target users quickly and reliably.

[0060] Real-time monitoring and prediction of terminal status, identification of network congestion points and high-latency areas, generation of dynamic network status diagrams, further optimization of path selection and information encoding, and improvement of system flexibility and response speed.

[0061] Furthermore, by introducing adaptive modulation and coding techniques and a channel state feedback mechanism, the robustness of information transmission was enhanced, ensuring the effective delivery of early warning information even in emergencies. The final transmission scheme document not only covers the optimal path, coding method, and transmission instructions, but also includes a detailed emergency response plan, providing solid technical support for emergency communications.

[0062] These or other aspects of this application will become more apparent from the description of the following embodiments. Attached Figure Description

[0063] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0064] Figure 1 A flowchart of a machine learning-based anomaly data identification method provided in this application is shown;

[0065] Figure 2 A schematic diagram of the structure of an anomaly data identification system based on machine learning provided in this application is shown;

[0066] Figure 3 A schematic diagram of the structure of a computing device provided in this application is shown. Detailed Implementation

[0067] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0068] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0069] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0070] Figure 1 A flowchart illustrating a machine learning-based anomaly data identification method is provided in this application embodiment, as follows: Figure 1 As shown, the method includes:

[0071] 101. Receive emergency event signals triggered by users and the early warning information release system, wherein the emergency event signals carry event type information, geographical location information, and dynamic environmental parameters;

[0072] Emergency event signals are notifications from user equipment or early warning systems that include event type information, geographic location information, and dynamic environmental parameters. They are used to initiate emergency response procedures and guide subsequent handling steps.

[0073] Event type information: This indicates the nature of the emergency event, such as fire, earthquake, flood, etc. This information is used to determine the response strategy;

[0074] Geographic location information: including precise GPS coordinates or address descriptions to help pinpoint the exact location where the event occurred;

[0075] Dynamic environmental parameters: Real-time collected data, such as temperature, humidity, air quality, wind speed, etc., are used to assess the current environmental conditions and their changing trends.

[0076] In the field of communication technology, when an emergency is detected, the user or the early warning system will send an emergency event signal carrying the above information. After the signal is received, the system will be activated immediately to prepare for the next step of analysis and processing to ensure timely response to the emergency.

[0077] For example, a network of smart sensors deployed in a city detects abnormally high temperatures and increased smoke concentration, triggering a fire alarm. This alarm, along with location and other environmental parameters, is sent to a central early warning system, providing foundational data for subsequent context-aware analysis.

[0078] 102. Analyze the dynamic environmental parameters using a context-aware algorithm to obtain abnormal data identification results. Based on the event type information and the abnormal data identification results, match personalized broadcast content suitable for the geographical location information in a preset emergency response plan library, and adjust the applicability and priority order of the emergency response plan library to obtain an emergency response strategy.

[0079] Among them, context-aware algorithms are a technology that combines machine learning and data analysis, which can identify the current environmental state based on real-time data and predict future trends.

[0080] Anomaly data identification results: The data analysis results obtained after processing dynamic environmental parameters through context-aware algorithms reflect the current environmental status and potential risks;

[0081] Personalized broadcast content: Early warning information customized according to event type and geographical environment, aiming to improve the relevance and effectiveness of information and ensure that information can be optimized for the characteristics of specific regions and groups of people;

[0082] Emergency Response Plan Library: A series of pre-set emergency plans covering different types of emergency events and their response measures;

[0083] Applicability and Priority: Adjust the application scope and importance ranking of each plan in the contingency plan library according to the current situation to adapt to the latest risk assessment results;

[0084] Emergency response strategy: A response plan formed after comprehensively considering all factors, which guides the subsequent construction of multimodal early warning information and the planning of transmission paths.

[0085] In practice, advanced machine learning models are used to analyze dynamic environmental parameters, understand the current situation, and select the most appropriate broadcast content from a pre-set emergency response plan library based on the event type. At the same time, the priority order of the plan library is adjusted to reflect the latest risk assessment results, thus forming an emergency response strategy.

[0086] According to the above embodiments, the context-aware algorithm identifies that this is a serious fire incident, and selects the most suitable fire escape guidelines and safety suggestions from the contingency plan library based on the location of the fire and the characteristics of the surrounding environment, while also increasing the priority of fire-related contingency plans.

[0087] 103. Based on the emergency response strategy, integrate the text, voice, and visual signals of the emergency response plan library to construct multimodal early warning information, and optimize the multimodal early warning information according to specific situations;

[0088] Among them, multimodal early warning information is a comprehensive information package that integrates text, voice, and visual signals. It can be optimized according to specific circumstances to adapt to different media and audience needs.

[0089] Optimization: Based on the specific needs of the context, further adjust the content and format of the information to ensure the best effect of information delivery;

[0090] Text signals: Information presented in text form, such as text messages, emails, or text prompts on screen;

[0091] Voice signals: Information transmitted through sound, such as telephone notifications, broadcasts, or messages played by voice assistants;

[0092] Visual signals: Information conveyed through images or videos, such as maps, animations, or augmented reality (AR) displays.

[0093] In practice, based on the established emergency response strategy, this stage involves creating a multimodal early warning message that includes all necessary warning information, presented in multiple formats to ensure that as many people as possible can receive and understand the information. The content and format of the message are further optimized according to the needs of specific scenarios.

[0094] Continuing with the above embodiments, in response to fire incidents, the system generates multimodal early warning information, including fire alarm sounds, fire prevention knowledge videos, and escape route maps. This information has been optimized to ensure that it can be correctly displayed on various terminal devices and is easy for people of different ages and abilities to understand.

[0095] 104. Based on the characteristics of the multimodal early warning information and the network status of the multimodal early warning broadcasting terminals in the affected area, the optimal transmission path is calculated using a path planning algorithm, the multimodal early warning information is encoded and processed using adaptive modulation and coding technology, a transmission scheme is generated based on the processed multimodal early warning information and the optimal transmission path, and the processed multimodal early warning information is sent to the selected multimodal early warning broadcasting terminals through the transmission scheme.

[0096] Among them, path planning algorithms are used to calculate the optimal path from the early warning information release system to the multimodal early warning broadcast terminal to ensure the efficiency and reliability of information transmission. Commonly used algorithms include Dijkstra's algorithm and A* search algorithm.

[0097] Adaptive modulation and coding technology: automatically adjusts the coding method according to channel conditions to ensure efficient transmission of information in complex network environments;

[0098] Transmission scheme: including the optimal transmission path, encoding method and specific transmission instructions, to ensure that information can be accurately transmitted to the target terminal;

[0099] Bandwidth: refers to the maximum amount of data that a communication channel can carry, usually measured in bits per second (bps). It determines the speed and quality of information transmission.

[0100] Redundancy strategy: To improve the reliability and fault tolerance of information transmission, multiple paths or other backup mechanisms are adopted.

[0101] In practice, considering the characteristics of multimodal early warning information and the network conditions within the target area, an appropriate path planning algorithm is selected to calculate the optimal transmission path, and adaptive modulation and coding technology is applied to encode the early warning information. Finally, a transmission scheme is generated to ensure that the information can be accurately transmitted to the target terminal.

[0102] According to the above embodiments, for the previously constructed multimodal early warning information, the system selects the A* search algorithm to calculate the optimal transmission path based on the network topology of the affected area, and applies an adaptive modulation and coding technique suitable for the current channel conditions, so that the early warning information can be stably transmitted even in the case of network congestion.

[0103] 105. After the multimodal warning information is processed by the multimodal warning broadcasting terminal, the appropriate multimodal warning information is selected for visual information presentation based on the modality selection algorithm, and augmented reality technology is introduced to support the visual information presentation to obtain feedback data, wherein the feedback data includes user understanding and reaction speed.

[0104] Among them, the modality selection algorithm determines how to adapt and process multimodal warning information and selects the most appropriate way to present visual information in order to enhance user experience;

[0105] Augmented Reality (AR) technology supports: making information more intuitive and easier to understand, improving users' comprehension and reaction speed;

[0106] Feedback data: This includes data on user comprehension and reaction speed, used to evaluate the effectiveness of information delivery and improve future early warning mechanisms;

[0107] User understanding and response speed: A time metric that measures whether users can quickly understand the content of an alert and take appropriate action after receiving it.

[0108] In practice, after receiving the processed warning information, the multimodal warning broadcasting terminal uses a modality selection algorithm to determine the best display method and may use augmented reality technology to assist in the presentation of visual information. Then, it collects user feedback data as the basis for improvement.

[0109] According to the above embodiments, once the multimodal warning information reaches the user's smartphone, the modality selection algorithm will play an escape route animation with AR markers to guide the user to the nearest safe exit. The system also records the time and actions of the user viewing the information, using this as a basis for subsequent analysis.

[0110] 106. Collect and analyze the feedback data, evaluate the transmission effect of the multimodal early warning information, and adjust the transmission range of the multimodal early warning information according to the transmission effect.

[0111] Among them, feedback data analysis: by analyzing user feedback data (such as comprehension and reaction speed), the actual effect of the early warning information is evaluated;

[0112] Distribution Scope Adjustment: Based on the results of feedback data analysis, adjust the future warning information release strategy to ensure that the information covers all groups that need to know and achieves the best dissemination effect;

[0113] Risk assessment: Evaluate the effectiveness of the existing early warning mechanism, identify potential risk points, and propose improvement suggestions;

[0114] Countermeasures: Develop specific solutions to address the problems identified in the assessment in order to improve the overall performance of the early warning system.

[0115] In practice, by analyzing the feedback data, we can understand the actual effect of the early warning information, and then adjust the scope of information dissemination to ensure that the early warning information can cover all groups that need to know and achieve the best dissemination effect.

[0116] Continuing with the above embodiments, by analyzing the feedback data collected from users' mobile phones, it was found that some users failed to understand the provided AR escape routes in a timely manner. Therefore, the system decided to expand the sending range and send simplified text and image warning information to more types of devices to ensure that a wider range of people can quickly obtain key information.

[0117] Through the implementation of steps 101 to 106, the entire process not only achieves complete closed-loop management from receiving emergency event signals, situational awareness analysis, construction and optimization of multimodal early warning information, efficient and stable transmission, adaptation and presentation to the final effect evaluation and feedback mechanism, but also greatly improves the accuracy, timeliness and coverage of early warning information transmission, ensuring that the safety of public life and property can be protected to the greatest extent under any circumstances.

[0118] To address the challenge of efficient transmission of multimodal early warning information in complex network environments and further improve the accuracy and timeliness of early warning information delivery, in some embodiments, step 104 involves calculating the optimal transmission path using a path planning algorithm based on the characteristics of the multimodal early warning information and the network conditions of the multimodal early warning broadcasting terminals within the affected area, encoding the multimodal early warning information using adaptive modulation and coding techniques, and generating a transmission scheme based on the processed multimodal early warning information and the optimal transmission path. This includes:

[0119] The information entropy algorithm is used to evaluate the content complexity of the multimodal early warning information. Combined with the data volume and urgency of the multimodal early warning information, its characteristics are analyzed to determine the transmission priority, required bandwidth, and redundancy strategy, resulting in a transmission optimization scheme. This scheme is then weighted using quality of service parameters to obtain an information transmission strategy. Based on this strategy, machine learning algorithms are applied to monitor the status of the multimodal early warning broadcasting terminals in the affected area in real time, predicting their changing trends. A deep learning model is used to identify network congestion points and high-latency areas of the multimodal early warning broadcasting terminals, forming a network topology map. Sudden events in the multimodal early warning broadcasting terminal topology map are adjusted to generate a dynamic network state map. Based on the dynamic network state map and the information transmission strategy, the Di algorithm in path planning is applied. The JKStra algorithm or A* search algorithm is used to calculate the transmission path from the early warning information dissemination system to the selected multimodal early warning broadcasting terminal, and a genetic algorithm is introduced to optimize the transmission path to obtain the optimal transmission path. Based on the segment characteristics of the multimodal early warning broadcasting terminal on the optimal transmission path, adaptive modulation and coding technology and a channel state information feedback mechanism are used to encode and optimize the multimodal early warning information, generating optimized multimodal early warning information. Operation instructions, transmission time windows, and error retransmission mechanisms for sending the optimized multimodal early warning information are sent through the selected multimodal early warning broadcasting terminal to obtain a transmission instruction set. Based on the transmission instruction set, a transmission scheme document is compiled, and risk assessment and countermeasures are added to the transmission scheme document to obtain the transmission scheme. The transmission scheme document includes: the optimal transmission path, encoding method, transmission instructions, and emergency response plan.

[0120] In this embodiment, multimodal early warning information includes data in various forms such as text, images, and videos, used to convey emergency or early warning information. This type of information usually needs to be quickly and accurately delivered to the public or relevant agencies in the affected area.

[0121] Information entropy algorithm: A mathematical method for evaluating data complexity to measure the uncertainty of information. In this context, it is used to evaluate the content complexity of multimodal early warning information to determine information transmission priority and required resources.

[0122] Quality of Service (QoS) parameters: a set of standards and service capabilities in network communication, such as latency, jitter, packet loss rate, etc., used to ensure that specific types of data traffic can obtain the necessary transmission quality;

[0123] Machine learning algorithms: a class of algorithms that automatically analyze data to obtain models and use the models to predict unknown data. Here, they are used to monitor and predict the changing trends of multimodal early warning broadcasting terminals in real time.

[0124] Deep learning models: a subset of machine learning that use multi-layered neural networks to handle complex pattern recognition tasks. In this approach, they help identify network congestion points and high-latency regions.

[0125] Path planning algorithms, such as Dijkstra's algorithm or A* search algorithm, are used to find the shortest or optimal path in a network topology graph. Genetic algorithms can further optimize these paths.

[0126] Adaptive modulation and coding techniques: techniques that dynamically adjust the modulation scheme and coding strategy according to channel conditions to optimize data transmission efficiency and reliability;

[0127] Channel state information feedback mechanism: The receiver feeds back information about the current channel state to the transmitter, so that the transmitter can adjust the transmission strategy according to the actual channel state.

[0128] Transmission instruction set: A series of operational instructions that define how and when to send optimized multimodal warning information, including time window settings and error retransmission mechanisms;

[0129] Emergency response plan: A pre-established action plan to deal with possible transmission failures or other unforeseen circumstances.

[0130] In this embodiment of the application, the aim is to calculate an optimal transmission path by comprehensively considering the characteristics of multimodal early warning information and the network conditions of broadcast terminals in the affected area, and to use appropriate encoding techniques to process the early warning information so as to ensure that the information can be efficiently and accurately conveyed to the target audience.

[0131] First, the information complexity is assessed using the information entropy algorithm, and the information transmission characteristics are determined by combining the data volume and urgency.

[0132] Then, an information transmission strategy is formed based on QoS parameters to guide subsequent steps. Next, machine learning and deep learning technologies are applied to monitor and predict the status of broadcasting terminals, generating a dynamic network status diagram.

[0133] Subsequently, based on this diagram and the transmission strategy, a suitable path planning algorithm is selected to calculate the optimal path, and then optimized using a genetic algorithm.

[0134] Finally, based on the characteristics of the selected path, adaptive modulation and coding techniques and channel state information feedback mechanisms are used to process the information. At the same time, a transmission instruction set is prepared and compiled into a transmission scheme document, which includes risk assessment and contingency plans.

[0135] Here is a specific example:

[0136] Imagine a city where the meteorological department detects an impending extreme weather event and needs to quickly issue a warning to residents. This warning includes text descriptions, charts, and short videos, forming a multimodal warning system. Because some areas of the city may experience network instability, an intelligent system is needed to ensure the effective transmission of this information.

[0137] First, the system uses the information entropy algorithm to evaluate the complexity of the warning information and, in combination with the amount of information and the degree of urgency, determines which information should be transmitted first, as well as the required bandwidth and redundancy strategies. Next, the system adjusts the transmission strategy according to QoS parameters to ensure that critical information can be delivered within the specified time.

[0138] The system's machine learning module begins monitoring the status of each broadcasting terminal in real time, such as network connection speed and stability, and predicts potential future problems. If network congestion is anticipated in a certain area, the system will make adjustments in advance to avoid affecting information transmission.

[0139] The system then uses path planning algorithms to find the optimal path from the publishing system to each broadcast terminal, and introduces a genetic algorithm to optimize this path, ensuring the fastest and most stable information transmission. For paths with poor network conditions, the system employs adaptive modulation and coding techniques to enhance the signal and ensure information integrity.

[0140] Ultimately, the system generated a detailed set of transmission instructions, including when to send information, how to handle transmission errors, and prepared contingency plans in case of unforeseen circumstances. All of these were documented in the transmission plan document to ensure the entire process proceeded smoothly.

[0141] To address the challenges of high complexity, large data volume, and varying urgency levels encountered during the transmission of multimodal early warning information, and to further improve the efficiency and reliability of information transmission, in one embodiment described above, the information entropy algorithm is used to evaluate the content complexity of the multimodal early warning information. Combined with the data volume and urgency level of the multimodal early warning information, the characteristics of the multimodal early warning information are analyzed to determine the transmission priority, required bandwidth, and redundancy strategy, resulting in a transmission optimization scheme, including:

[0142] The multimodal early warning information is processed using an information entropy algorithm to evaluate its complexity, resulting in a quantified content complexity. The quantified complexity is then analyzed based on the data size and urgency of the multimodal early warning information to obtain a specific characteristic description for each piece of information. Based on these descriptions, the transmission priority of the multimodal early warning information is determined, generating an information transmission priority list. Based on this priority list, the required bandwidth is estimated to determine the minimum bandwidth requirement. For this minimum bandwidth requirement, a redundancy strategy is implemented to ensure reliability, generating a redundant transmission plan. The multimodal early warning information will be transmitted through other redundancy mechanisms to address network failures and congestion, and a transmission optimization scheme is formed based on the redundant transmission plan.

[0143] In this embodiment, the characteristic analysis process involves comprehensively analyzing the data volume and urgency of the multimodal early warning information, and combining the quantitative results of content complexity to derive a specific characteristic description for each piece of information; these characteristic descriptions are used to guide the formulation of subsequent transmission strategies.

[0144] Information transmission priority list: Based on the specific characteristics of multimodal early warning information, a list of information transmission order is generated according to importance and urgency; this list ensures that the most critical information is processed and sent first.

[0145] Bandwidth requirement estimation: Based on the information transmission priority list, calculate the minimum bandwidth required to ensure efficient transmission; this step aims to optimize network resource utilization and avoid unnecessary bandwidth waste.

[0146] Redundancy strategy: To improve the reliability and fault tolerance of information transmission, multiple paths or other backup mechanisms are adopted; for example, through different communication channels or by adding additional data copies.

[0147] Redundancy transmission plan: This plan outlines how to ensure the reliability of information transmission through redundancy mechanisms, including selecting backup paths and setting retransmission mechanisms, in order to cope with possible network failures and congestion.

[0148] Transmission optimization plan: Integrate the results of all the above steps to form a comprehensive transmission plan document, which includes the optimal transmission path, encoding method, transmission instructions and emergency response plan.

[0149] In this embodiment, efficient and reliable transmission in complex network environments is ensured through in-depth analysis of multimodal early warning information. First, the information entropy algorithm is used to evaluate the content complexity of each early warning message, and its characteristics are analyzed in conjunction with its data volume and urgency to generate a specific characteristic description.

[0150] Then, based on these characteristic descriptions, the priority of information transmission is determined, and a priority list is generated.

[0151] Next, based on the priority list, estimate the minimum bandwidth required and formulate a reasonable bandwidth allocation scheme.

[0152] Finally, for the minimum bandwidth requirement scheme, a redundancy strategy is designed to enhance transmission reliability, and all these steps are integrated into a complete transmission optimization scheme to ensure that critical early warning information can be effectively delivered even in the event of network failure or congestion.

[0153] Here is a specific example:

[0154] According to the above embodiment, the system first uses the information entropy algorithm to evaluate the content complexity of each warning message. The quantification results show that some video segments have high complexity.

[0155] Next, taking into account the data volume (e.g., large video files) and the level of urgency (e.g., immediate evacuation instructions), the system generates detailed characteristic descriptions, indicating which information is most critical. Based on these characteristic descriptions, the system establishes an information transmission priority list, ensuring that the most important evacuation guidelines are sent first.

[0156] Subsequently, the system estimated the minimum required bandwidth and formulated a bandwidth allocation scheme to ensure that critical information could be delivered in the shortest possible time. Considering the potential instability in urban networks, the system also developed a redundant transmission plan, including using backup paths and adding data copies to ensure high reliability of information transmission.

[0157] Ultimately, all these steps were integrated into a transmission optimization scheme, documented, to ensure the entire transmission process proceeded smoothly and effectively addressed potential network problems.

[0158] To address the optimization of multimodal early warning information transmission paths in complex network environments and further improve transmission efficiency and stability, in another embodiment, the transmission path from the early warning information dissemination system to the selected multimodal early warning broadcasting terminal is calculated using Dijkstra's algorithm or A* search algorithm in path planning algorithms based on the dynamic network state diagram and the information transmission strategy. A genetic algorithm is then introduced to optimize the transmission path to obtain the optimal transmission path, including:

[0159] The status of the multimodal early warning broadcasting terminal network within the affected area is comprehensively evaluated using the dynamic network state diagram and the information transmission strategy to obtain basic data. Based on this basic data, the path from the early warning information release system to the selected multimodal early warning broadcasting terminal is calculated using Dijkstra's algorithm or A* search algorithm in path planning, generating a preliminary transmission path. A genetic algorithm is introduced to iteratively optimize the preliminary transmission path, selecting the transmission path with the best performance to generate candidate transmission paths. The stability and anti-interference capability of the candidate transmission paths are analyzed using performance evaluation under network congestion points, high-latency areas, and sudden events to obtain important transmission paths with high efficiency and stability. Based on the important transmission paths, the transmission path is selected by combining the network topology, the status of the multimodal early warning broadcasting terminal, the urgency of the event, and the information transmission priority to generate the optimal transmission path.

[0160] In this embodiment, the dynamic network state diagram includes real-time network topology, connection status between nodes, and quality data (such as bandwidth, latency, packet loss rate, etc.) of each link segment; it is used to describe the current state of the network environment between the early warning information release system and the multimodal early warning broadcast terminal, providing basic information for path planning.

[0161] Information transmission strategy: refers to a series of rules and methods formulated according to the characteristics and requirements of the communication network to ensure that information can be transmitted in accordance with predetermined requirements (such as priority, security, reliability, etc.); these strategies may involve routing selection, congestion control, error detection and correction, etc.

[0162] Dijkstra's algorithm and A* search algorithm are two commonly used shortest path algorithms. Dijkstra's algorithm is suitable for calculating the single-source shortest path problem in weighted graphs, while A* is a heuristic search algorithm that combines the advantages of best-first search and Dijkstra's algorithm. It predicts the distance to the target while considering the actual distance, thus finding the optimal path faster.

[0163] Genetic algorithm: An optimization algorithm that simulates natural selection and genetic mechanisms. It iteratively solves problems by simulating selection, crossover and mutation operations in the process of biological evolution. It is particularly suitable for solving complex and nonlinear optimization problems and can effectively avoid getting trapped in local optima.

[0164] In this embodiment, the status of the multimodal early warning broadcasting terminal network in the affected area is first assessed using a dynamic network state diagram and information transmission strategy to generate basic data.

[0165] Then, based on this basic data, the preliminary transmission path from the early warning information dissemination system to the selected terminal is calculated using Dijkstra's or A* algorithm.

[0166] Next, in order to further improve the quality of the path, a genetic algorithm is introduced to iteratively optimize the initial path and select candidate paths with better performance.

[0167] Finally, by evaluating the performance of candidate paths in network congestion points, high-latency areas, and sudden events, and taking into account factors such as network structure, terminal status, event urgency, and information transmission priority, the optimal transmission path is determined.

[0168] Here is a specific example:

[0169] For example, during a natural disaster early warning system deployment, a system needs to quickly transmit alert information to multimodal early warning broadcasting terminals scattered throughout the disaster area. Because disasters can damage some communication infrastructure, causing changes in network topology, and because network quality varies significantly between different regions, the aforementioned solution is necessary to ensure efficient and stable information transmission.

[0170] The system first collects the latest dynamic network state graph to understand which areas of the network are severely affected and which paths may become bottlenecks.

[0171] Next, based on existing information transmission strategies, such as assigning higher transmission priority to particularly urgent information, an A* search algorithm was used to calculate a preliminary path from the publishing system to a terminal in a remote area. Considering that multiple paths may have similar theoretical transmission efficiencies, but their actual performance could vary due to instantaneous network changes, a genetic algorithm was used to iteratively optimize the preliminary path multiple times, selecting several candidate paths. After testing the performance of these paths in actual transmission at network congestion points and high-latency areas, the optimal path was finally selected. This path avoided known problem areas while meeting the requirement of rapid delivery of urgent information, successfully achieving the timely and accurate dissemination of early warning information.

[0172] To address potential network anomalies and risks during transmission and further improve the reliability and stability of information transmission, one or more of the above embodiments further include:

[0173] The transmission optimization scheme is weighted and adjusted using the service quality parameters. A fault prediction and health management system is introduced to obtain an information transmission strategy and monitor the health status during transmission in real time. When an anomaly is detected, the health management system can respond quickly, take preventive measures, monitor potential risks during transmission, and automatically adjust the information transmission strategy.

[0174] In this embodiment, Quality of Service (QoS) parameters are a set of standards and service capabilities in network communication, such as latency, jitter, and packet loss rate, used to ensure that specific types of data traffic can obtain the necessary transmission quality; in this scheme, QoS parameters are used to evaluate and adjust information transmission strategies.

[0175] Fault Prediction and Health Management System (PHM): A system that combines real-time monitoring, data analysis, and predictive models to identify potential faults in advance and manage the health status of the system; it predicts possible faults by monitoring key performance indicators and takes preventative measures to avoid service interruptions.

[0176] Weighted adjustment processing: This involves comprehensively considering multiple factors based on different weights to optimize the decision-making process; in this context, it refers to using different weights of QoS parameters to fine-tune the transmission optimization scheme to ensure that the transmission strategy better meets actual needs.

[0177] In this embodiment, not only is the optimal transmission path calculated, but a quality of service (QoS) parameter is also introduced to weight and adjust the transmission optimization scheme, ensuring that the transmission strategy can better adapt to changes in the network environment. Simultaneously, a fault prediction and health management system (PHM) is introduced to monitor the health status during transmission in real time. When the PHM detects anomalies, it can respond quickly and take preventative measures, such as automatically switching to an alternative path or adjusting transmission parameters, to monitor and address potential risks during transmission, thereby ensuring the efficiency and stability of information transmission.

[0178] Here is a specific example:

[0179] Continuing with the above embodiments, firstly, the optimal transmission path between the early warning information release system and each multimodal early warning broadcast terminal is calculated using a dynamic network state diagram and information transmission strategy.

[0180] Then, based on the Quality of Service (QoS) parameters, the path was weighted and adjusted to ensure that high-priority information could be delivered to the target user in the shortest possible time.

[0181] In addition, the system incorporates a Predictive Health Management (PHM) system to monitor various performance metrics during transmission in real time, such as bandwidth utilization, latency, and packet loss rate. Once PHM detects abnormal fluctuations or potential fault signs in a network area, such as a sudden increase in latency at a node, the system immediately triggers an alarm and automatically takes preventative measures, such as switching to a preset redundant path or adjusting the data transmission rate, to avoid potential service interruptions. Throughout the transmission process, PHM continuously monitors potential risks and dynamically adjusts information transmission strategies to ensure the timeliness and accuracy of early warning information, even under poor network conditions.

[0182] To address the issue of timely detection and response to potential risks during transmission, and to further improve the reliability and stability of information transmission, in one or more of the above embodiments, monitoring potential risks during transmission and automatically adjusting the information transmission strategy includes:

[0183] The system continuously monitors each stage of the transmission execution plan, collects and analyzes network performance indicators and terminal feedback data to identify potential risk factors. The multimodal early warning broadcast terminal performance indicators include latency, packet loss rate, and bandwidth utilization. Machine learning algorithms are applied to predict the probability of occurrence and the scope of impact of potential risks in the terminal feedback data, providing early warnings of fault points or bottlenecks and generating risk assessment reports. Based on these risk assessment reports, the system automatically adjusts information transmission strategies, replanning transmission paths or adding redundant transmission channels after detecting high-risk areas, and dynamically adjusting service quality parameters to prioritize the delivery of the multimodal early warning information.

[0184] In this embodiment, network performance metrics refer to a series of parameters used to evaluate network transmission quality, such as latency, packet loss rate, and bandwidth utilization. These metrics reflect the network's performance in actual operation and are an important basis for judging the network's health status.

[0185] Terminal feedback data: Data collected from multimodal early warning broadcast terminals, including but not limited to user understanding and reaction speed, information reception confirmation, etc., to evaluate the information transmission effect and the performance status of terminal devices;

[0186] Machine learning algorithms: techniques that automatically learn patterns from large amounts of data by building models and are able to predict or classify new data; in this context, it is used to predict the probability of occurrence of potential risks and the scope of their impact.

[0187] Risk assessment report: A document based on collected data and anomaly identification results, which details the potential risks, their probability of occurrence, and the possible scope of impact, providing a basis for subsequent decision-making.

[0188] In this embodiment, the system first continuously monitors each stage of the transmission execution plan, collects and analyzes network performance indicators and terminal feedback data in real time, and identifies potential risk factors.

[0189] Next, machine learning algorithms are applied to conduct in-depth analysis of terminal feedback data, predict the probability of potential risks and their impact range, identify fault points or bottlenecks in advance, and generate risk assessment reports.

[0190] Finally, based on this report, the system automatically adjusted its information transmission strategy, replanned transmission paths or added redundant transmission channels for high-risk areas, and dynamically adjusted service quality parameters to prioritize the delivery of multimodal early warning information.

[0191] Here is a specific example:

[0192] According to the above embodiments, firstly, the system continuously monitors the entire transmission process from the release of the warning information to its final presentation to the user, and collects network performance indicators (such as latency, packet loss rate and bandwidth utilization) and terminal feedback data (such as user comprehension speed and information reception confirmation) in real time.

[0193] When the system detects abnormal fluctuations in network performance in a certain area, such as a sudden increase in latency or a rise in packet loss rate, it immediately applies machine learning algorithms to analyze the data and predict the probability of potential risks and their scope of impact.

[0194] In this way, the system can identify potential failure points or bottlenecks in several key nodes in advance and generate detailed risk assessment reports.

[0195] Based on the report's recommendations, the system automatically adjusted its information transmission strategy. For high-risk areas detected, it replanned transmission paths, added redundant transmission channels, and dynamically adjusted service quality parameters to ensure that even under poor network conditions, it could prioritize the efficient delivery of multimodal early warning information.

[0196] In addition, the system continuously optimizes the transmission strategy based on terminal feedback data, further improving the effectiveness of information delivery and user experience.

[0197] To address the issue of accurately assessing the impact of environmental factors on emergency response measures and further improve the personalization and effectiveness of early warning information, as another embodiment, step 102 involves using a context-aware algorithm to analyze the dynamic environmental parameters, obtaining abnormal data identification results, and matching personalized broadcast content suitable for the geographical location information from a pre-set emergency response plan database based on the event type information, including:

[0198] The system receives emergency event signals triggered by users and the early warning information dissemination system, extracts and parses event type information, geographical location information, and dynamic environmental parameters from the emergency event signals, wherein the dynamic environmental parameters include weather conditions, traffic flow, and population density; it analyzes the dynamic environmental parameters using a context-aware algorithm to assess the impact of current environmental conditions on emergency response and obtains environmental anomaly data identification results; based on the event type information and the environmental anomaly data identification results, it matches personalized broadcast content suitable for the geographical location information from a preset emergency response plan library, filters the emergency response plan library, and obtains personalized broadcast content.

[0199] In this embodiment, the context-aware algorithm, a technique combining machine learning and data analysis, is capable of identifying the current environmental state based on real-time data and predicting future trends; in this context, it is used to analyze dynamic environmental parameters and assess the impact of these factors on emergency response.

[0200] Dynamic environmental parameters include real-time data such as weather conditions, traffic flow, and crowd density, used to assess the impact of current environmental conditions on emergency response; these parameters help the system understand the specific circumstances when an event occurs, thereby making more accurate decisions.

[0201] Personalized broadcast content: Early warning information is customized based on event type and geographical environment, aiming to improve the relevance and effectiveness of information and ensure that information can be optimized for the characteristics of specific regions and populations.

[0202] In this embodiment, emergency event signals from users or early warning systems are first received and parsed, extracting event type information, geographical location information, and dynamic environmental parameters. Then, a context-aware algorithm is used to conduct in-depth analysis of the dynamic environmental parameters, assessing the impact of current environmental conditions on emergency response and obtaining environmental anomaly data identification results. Next, based on the event type information and the environmental anomaly data identification results, personalized broadcast content most suitable for the current geographical location is selected from a pre-set emergency response plan library, ensuring that the early warning information is both accurate and targeted.

[0203] Here is a specific example:

[0204] For example, in a smart city project, when seismic activity is detected in a certain area, the early warning system immediately receives the triggered emergency event signal. The system first analyzes the signal to extract the event type (such as earthquake), geographical location (such as the epicenter), and dynamic environmental parameters (such as the weather conditions, traffic flow, and population density at the time).

[0205] Next, context-aware algorithms are used to analyze these dynamic environmental parameters and assess the impact of current environmental conditions on emergency response. For example, if an earthquake occurs during peak hours with heavy traffic, the system will consider potential traffic congestion during evacuation. Based on these anomaly data identification results, the system matches the most suitable personalized broadcast content for the geographical location from a pre-set emergency response plan library. Specifically, the system selects content including shelter locations, optimal evacuation routes, and safety advice on how to deal with aftershocks, and further filters and optimizes this information to ensure it is not only accurate but also easy to understand and implement.

[0206] Ultimately, these personalized broadcasts were quickly delivered to residents in the affected areas, helping them to take appropriate action swiftly and minimize the damage caused by the disaster.

[0207] This application addresses the issue that traditional path planning algorithms fail to adequately consider dynamic network changes and the characteristics of multimodal early warning information in emergency communication scenarios, leading to low transmission efficiency and insufficient reliability. Especially in complex and ever-changing network environments, ensuring the rapid and stable delivery of critical early warning information to target users is a significant challenge. Therefore, a new alternative solution is proposed, which includes:

[0208] The application path planning algorithm uses either Dijkstra's algorithm or A* search algorithm to calculate the transmission path from the early warning information dissemination system to the selected multimodal early warning broadcasting terminal, including:

[0209] The dynamic network state diagram and the information transmission strategy are used to comprehensively evaluate the status of the multimodal early warning broadcasting terminal network in the affected area to obtain basic data. This basic data is then used to calculate node reliability. The node reliability is obtained; It is obtained by calculation using the following formula: in, Represents a node Reliability, Represents a node transmission rate Represents a node The delay time, Represents a node The degree of congestion, These are weight parameters that are adaptively adjusted based on the network status of the multimodal early warning broadcasting terminal. These are the adjustment coefficients that affect the index;

[0210] The following is a detailed explanation of each parameter:

[0211] Node reliability Node reliability, parameters are comprehensively evaluated based on transmission rate. Delay time and congestion level To obtain;

[0212] : is a weight parameter that is adaptively adjusted based on the network status of the multimodal early warning broadcast terminal, used to measure the impact of transmission rate on node reliability; this parameter is trained from a large amount of historical data through machine learning algorithms to ensure that it can be dynamically adjusted under different network conditions to reflect the actual network status;

[0213] : Represents a node The transmission rate is usually measured in Mbps; it is data collected in real time by network monitoring tools and reflects the data transmission speed between nodes; a higher transmission rate means that information can reach the target user faster, thereby improving the reliability of the nodes.

[0214] The adjustment coefficient of the index is used to adjust the degree of influence of transmission rate on node reliability; its value can be determined by experiments or simulations, aiming to capture the trend of transmission rate changes over time and ensure that the formula can adapt to different network environments and conditions.

[0215] : This is another weight parameter that is adaptively adjusted based on the network status of the multimodal early warning broadcast terminal. It is used to measure the impact of latency on node reliability. It is also trained from historical data through machine learning algorithms to ensure that it can be dynamically adjusted to reflect the actual network conditions.

[0216] : Represents a node The latency, usually measured in milliseconds, is data collected in real time by network monitoring tools and reflects the data transmission latency between nodes. Lower latency means that information can be delivered to users more promptly, thereby improving node reliability.

[0217] The adjustment coefficient affecting the index is used to adjust the degree of impact of latency on node reliability; its value can be determined through experiments or simulations, aiming to capture the trend of latency changing over time and ensure that the formula can adapt to different network environments and conditions.

[0218] : is a weight parameter that is adaptively adjusted based on the network status of the multimodal early warning broadcast terminal, used to measure the impact of congestion on node reliability; it is trained from historical data through machine learning algorithms to ensure that it can be dynamically adjusted to reflect the actual network conditions;

[0219] : Represents a node The congestion level is usually a value between 0 and 1, where 0 represents no congestion and 1 represents complete congestion. It is data collected in real time by network monitoring tools and reflects the network traffic between nodes. A lower congestion level means smoother information transmission, thereby improving the reliability of the nodes.

[0220] The adjustment coefficient affecting the index is used to adjust the degree of congestion on node reliability; its value can be determined through experiments or simulations, aiming to capture the trend of congestion over time and ensure that the formula can adapt to different network environments and conditions.

[0221] The adjustment coefficient affecting the index is used to further adjust the impact of congestion on node reliability, especially under high congestion conditions, to ensure the sensitivity and accuracy of the formula; its value can also be determined through experiments or simulations.

[0222] The following is a brief introduction to the reasons for each sub-item design:

[0223] This sub-item is used to evaluate the impact of transmission rate on node reliability; as the transmission rate increases, the speed of information transmission accelerates, and the reliability of the node increases accordingly; [Introduction] and The purpose is to capture the non-linear growth trend of transmission rate and ensure that the formula can maintain accuracy and sensitivity at different rates;

[0224] This sub-item is used to evaluate the impact of latency on node reliability; the lower the latency, the higher the timeliness of information transmission, and the stronger the node reliability; (Introduction) and The purpose is to capture the square root effect of delay time and ensure that the formula can better reflect the actual impact of delay time on reliability;

[0225] This sub-item is used to assess the impact of congestion on node reliability; the lower the congestion level, the smoother the information transmission, and the higher the node reliability; (Introducing...) and The purpose is to capture the exponential decay effect of congestion and ensure that the formula accurately reflects the sharp decline in reliability under high congestion conditions.

[0226] The formula adds up the components to comprehensively consider the impact of transmission rate, latency, and congestion on node reliability. Each component represents the contribution of different factors to reliability, and a comprehensive and balanced node reliability assessment can be obtained through weighted summation. Within each component, multiplication is used to combine weight parameters and other variables (such as transmission rate, latency, and congestion) to emphasize the interaction between different factors. For example, combining transmission rate and its influence index through multiplication can more accurately capture the impact of these factors on node reliability.

[0227] The overall design of the formula is based on the introduction of three key factors: transmission rate, latency, and congestion level. By combining their respective influence indices and weighting parameters, the formula comprehensively reflects the performance of nodes in the network. This method not only considers the impact of each factor on reliability but also ensures the comprehensiveness and accuracy of the evaluation results through weighted summation. Furthermore, the formula's design allows for dynamic adjustment of the weighting parameters based on actual network conditions, enhancing the system's flexibility and adaptability, thus providing a reliable guarantee for emergency communication.

[0228] Here is a specific example:

[0229] In a simulated urban emergency response system, the early warning information dissemination system needs to rapidly transmit multimodal early warning information (such as text, voice, and images) to multimodal early warning broadcasting terminals distributed throughout the city. To ensure the efficiency and reliability of information transmission, a path planning method based on Dijkstra's algorithm or A* search algorithm combined with a genetic algorithm was adopted, and the following specific calculations were performed.

[0230] Parameter settings:

[0231] Set a node transmission rate Delay time Congestion level ;

[0232] Weight parameters ;

[0233] Impact Index Adjustment Coefficient ;

[0234] Substitute the values: Calculation results: Node reliability The calculation results show that the node exhibits high reliability in network transmission performance. Specifically, this value reflects the node's reliability after comprehensively considering factors such as transmission rate, latency, and congestion level. The node can maintain high efficiency and stability during the transmission of multimodal early warning information. A high reliability value (close to 1) means that the node has a low risk of failure and a high quality of service in data transmission, and can effectively transmit key early warning information in complex network environments.

[0235] Using Dijkstra's algorithm or A* search algorithm in combination to assess node reliability The calculated basic data is used to calculate the preliminary transmission path, which is obtained by calculating each path. Total cost Obtain; wherein, the total cost It is obtained by calculation using the following formula: in, Represent each path Total cost , Representing a path upper node To the node The distance between them Represents a node To the node The delay between Indicates the maximum possible distance. These are the adjustment coefficients that affect the index. Representing a path The total number of nodes on;

[0236] The following is a detailed explanation of each parameter:

[0237] : Indicates the path from the early warning information dissemination system to the selected multimodal early warning broadcasting terminal. The total cost; this cost takes into account factors such as the distance between nodes on the path, latency, and node reliability, and is used to evaluate the merits of different transmission paths; a lower total cost means a more efficient transmission path;

[0238] : Indicates a path upper node To the node The physical or logical distance between them is usually measured in meters; it is obtained through network topology maps and Geographic Information System (GIS) data; shorter distances generally mean lower transmission latency and higher transmission efficiency.

[0239] : Represents a node To the node The transmission latency between nodes is usually measured in milliseconds; it is data collected in real time by network monitoring tools and reflects the time required for information to be transmitted between two nodes; lower latency means more timely information transmission, thereby improving the communication efficiency between nodes.

[0240] : Represents the maximum possible distance between any two nodes on the path, usually a preset value, depending on the specific application scenario and network size; it is used to normalize the impact of the distance between nodes, ensuring that the formula remains consistent and comparable in networks of different scales;

[0241] The distance impact index adjustment coefficient is used to adjust the degree of influence of distance on the total path cost; its value can be determined through experiments or simulations, aiming to capture the nonlinear impact of distance on transmission costs; a higher value indicates a higher impact. The value increases the impact of distance on the total cost, significantly increasing the cost of long-distance transmission;

[0242] The delay time impact adjustment coefficient is used to adjust the degree of influence of delay time on the total path cost; its value can also be determined experimentally or through simulation, aiming to capture the nonlinear effect of delay time on transmission cost; a higher value indicates a higher impact. The value increases the impact of latency on total cost, significantly increasing the cost of high-latency paths;

[0243] Node reliability impact index adjustment coefficient, used to adjust the degree of influence of node reliability on total path cost; its value can be determined through experiments or simulations, aiming to capture the nonlinear impact of node reliability on transmission cost; higher values ​​indicate a higher impact. The value increases the impact of low-reliability nodes on the total cost, causing the cost of paths containing unreliable nodes to rise significantly.

[0244] : Indicates a path The total number of nodes on the path, including the starting and ending points; it is calculated by a path planning algorithm and reflects the complexity and length of the path; a longer path (i.e. more nodes) usually means higher transmission costs because each node may introduce additional latency and uncertainty.

[0245] The overall design of the formula is to provide a comprehensive method for evaluating the total cost of a transmission path, taking into full account the impact of distance between nodes, delay time, and node reliability on information transmission efficiency. By introducing a nonlinear adjustment coefficient, the formula can capture the complex interactions between various factors, ensuring that it accurately reflects the actual transmission performance of the path under different network conditions.

[0246] Here is a specific example:

[0247] According to the above embodiment, the parameters are set as follows:

[0248] Set path Distance between the two adjacent nodes ;

[0249] Maximum possible distance ;

[0250] Delay ;

[0251] Impact Index Adjustment Coefficient ;

[0252] path Include 1 node

[0253] Substitute the values: Calculation results:

[0254] Total path cost The calculation results show that the transmission path from the early warning information dissemination system to the selected multimodal early warning broadcasting terminal is... After comprehensively considering factors such as distance between nodes, latency, and node reliability, the transmission cost is relatively high. Specifically, this figure reflects that information transmission on this path needs to overcome a large physical distance and a certain transmission delay. At the same time, the reliability of some nodes may be slightly lower, increasing the potential risks and complexity in the transmission process.

[0255] Introducing a genetic algorithm to iteratively optimize each path Total cost and the reliability of the nodes The optimal transmission path with the best performance is selected by evaluating each path. The fitness function is calculated based on the performance. Obtain; where, fitness function It is obtained by calculation using the following formula:

[0256] in, Representing a path Adaptability, These are the weighting parameters for total cost, average node reliability, and priority, respectively. These are the adjustment coefficients that affect the index. Additional weighted values ​​indicating the urgency of the event and the priority of information transmission;

[0257] The following is a detailed explanation of each parameter:

[0258] : Indicates a path The fitness value is used to comprehensively evaluate factors such as the total cost of a path, node reliability, and priority. It is obtained by weighted summation of the components and reflects the overall performance of the path in multimodal early warning information transmission. The lower the fitness value, the better the path.

[0259] : A weight parameter that is adaptively adjusted based on the network status of the multimodal early warning broadcasting terminal, used to measure the total path cost. Impact on path fitness; This parameter is trained from a large amount of historical data using machine learning algorithms to ensure that it can be dynamically adjusted under different network conditions to reflect the actual network situation;

[0260] : This is another weight parameter that is adaptively adjusted based on the network status of the multimodal early warning broadcasting terminal. It is used to measure the impact of average node reliability on path fitness. It is also trained from historical data through machine learning algorithms to ensure that it can be dynamically adjusted to reflect the actual network conditions.

[0261] It is a weighted parameter that adaptively adjusts based on the urgency of the event and the priority of information transmission, used to measure... Impact on path adaptability; This parameter is set based on the event type and the importance of the warning information to ensure that high-priority information is processed first.

[0262] The reliability impact index adjustment coefficient is used to adjust the degree of influence of node reliability on path fitness; its value can be determined through experiments or simulations, aiming to capture the trend of reliability changing over time and ensure that the formula can adapt to different network environments and conditions.

[0263] The reliability impact index adjustment coefficient is used to further adjust the influence of node reliability on path fitness, especially in high reliability cases, to ensure the sensitivity and accuracy of the formula; its value can also be determined through experiments or simulations.

[0264] : This represents an additional weighted value that indicates the urgency of the event and the priority of information transmission, used to ensure that critical information can be delivered preferentially in complex network environments; this value is preset or dynamically adjusted according to the type of event (such as natural disasters, public health events, etc.).

[0265] The following is a brief introduction to the reasons for each sub-item design:

[0266] This sub-item is used to evaluate the impact of total path cost on path fitness; total path cost Taking into account factors such as distance, latency, and node reliability, higher costs imply lower transmission efficiency or greater risk; therefore, the introduction of... The weighting parameter is used to emphasize the importance of cost in path selection, ensuring that the lowest cost path is selected when resources are limited. This sub-item is used to evaluate the impact of the average reliability of all nodes on the path on path fitness; node reliability This reflects the transmission performance of each node in the network; higher reliability means lower failure risk and higher service quality. Weighting parameters and nonlinear adjustment coefficients This is to capture the nonlinear effect of reliability on fitness and ensure that the formula accurately reflects the importance of node reliability;

[0267] This sub-item assesses the impact of event urgency and information transmission priority on path adaptability; high-priority information (such as emergency evacuation instructions) needs to be communicated first to ensure users receive the most critical information in a timely manner; [Introduction] The weighting parameter is used to emphasize the importance of priority in path selection, ensuring that high-priority information can be delivered to the target user in the shortest possible time;

[0268] The formula sums the components to comprehensively consider the impact of total path cost, node reliability, and priority on path fitness. Each component represents the contribution of different factors to fitness, and a weighted summation yields a comprehensive and balanced path fitness assessment. Within each component, multiplication is used to combine weight parameters with other variables (such as total cost, reliability, and priority) to emphasize the interaction between different factors. For example, combining total cost and its weight parameters through multiplication can more accurately capture the impact of these factors on path fitness.

[0269] The overall design of the formula aims to provide a comprehensive method for evaluating path fitness, taking into full account factors such as total path cost, node reliability, and priority. By introducing weighted parameters and nonlinear adjustment coefficients, the formula can flexibly adapt to different network conditions and event requirements, ensuring the selection of the optimal transmission path in emergency communication scenarios. This method not only improves the efficiency and reliability of information transmission but also enhances the robustness and flexibility of the system, providing a solid technical guarantee for emergency response.

[0270] Here is a specific example:

[0271] Continuing with the above embodiments, the parameters are set as follows:

[0272] Weight parameters ;

[0273] Impact Index Adjustment Coefficient ;

[0274] Additional weighting of event urgency and information transmission priority ;

[0275] Substitute the values: Calculation results: Path fitness The calculation results show that the transmission path from the early warning information dissemination system to the selected multimodal early warning broadcasting terminal is... After comprehensively considering factors such as total cost, node reliability, and information priority, the overall performance is quite reasonable. Specifically, this figure reflects that although there are certain transmission costs and delays on this path, the node reliability is high and key information can be prioritized, ensuring the efficiency and stability of information transmission.

[0276] Figure 2 This application provides a schematic diagram of the structure of an anomaly data identification device (or system) based on machine learning, as shown in the embodiments of this application. Figure 2 As shown, the device includes:

[0277] The receiving module 21 is used to receive emergency event signals triggered by users and the early warning information release system, wherein the emergency event signals carry event type information, geographical location information and dynamic environmental parameters;

[0278] Analysis module 22 uses a context-aware algorithm to analyze the dynamic environmental parameters, obtain abnormal data identification results, and matches personalized broadcast content suitable for the geographical location information in a preset emergency response plan library based on the event type information and the abnormal data identification results, and adjusts the applicability and priority order of the emergency response plan library to obtain an emergency response strategy.

[0279] The construction module 23 is used to integrate text, voice and visual signals from the emergency response plan library based on the emergency response strategy to construct multimodal early warning information, and optimize the multimodal early warning information according to specific situations;

[0280] The calculation module 24 is used to calculate the optimal transmission path using a path planning algorithm based on the characteristics of the multimodal early warning information and the network status of the multimodal early warning broadcasting terminals in the affected area, encode the multimodal early warning information using adaptive modulation and coding technology, generate a transmission scheme based on the processed multimodal early warning information and the optimal transmission path, and send the processed multimodal early warning information to the selected multimodal early warning broadcasting terminal through the transmission scheme.

[0281] The feedback module 25 is used to, after the multimodal warning information is acquired and processed by the multimodal warning broadcast terminal, adapt and process it based on the modality selection algorithm and select appropriate multimodal warning information for visual information presentation, and introduce augmented reality technology to support the visual information presentation to obtain feedback data, wherein the feedback data includes user understanding and reaction speed.

[0282] Evaluation module 26 is used to collect and analyze the feedback data, evaluate the transmission effect of the multimodal early warning information, and adjust the transmission range of the multimodal early warning information according to the transmission effect.

[0283] Figure 2 The aforementioned machine learning-based anomaly data identification device can perform... Figure 1 The implementation principle and technical effects of the machine learning-based anomaly data identification method described in the illustrated embodiment will not be repeated here. The specific methods by which each module and unit of the machine learning-based anomaly data identification device in the above embodiments perform their operations have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0284] In one possible design, Figure 2 The machine learning-based anomaly data identification device shown in the embodiment can be implemented as a computing device, such as... Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0285] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 32.

[0286] The processing component 32 is used for the above Figure 1 The embodiment describes a machine learning-based method for identifying abnormal data.

[0287] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0288] Storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0289] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0290] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0291] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0292] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0293] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The embodiment shown is a machine learning-based method for identifying abnormal data.

[0294] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0295] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0296] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0297] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for identifying abnormal data based on machine learning, characterized in that, include: It receives emergency event signals triggered by users and the early warning information release system. These emergency event signals carry event type information, geographical location information, and dynamic environmental parameters. By using context-aware algorithms to analyze dynamic environmental parameters, abnormal data identification results are obtained. Based on event type information and abnormal data identification results, personalized broadcast content suitable for geographical location information is matched from the preset emergency response plan library. The applicability and priority order of the emergency response plan library are adjusted to obtain emergency response strategies. Based on emergency response strategies, text, voice, and visual signals from the emergency response plan database are integrated to construct multimodal early warning information, and the multimodal early warning information is optimized according to specific scenarios; The information entropy algorithm is used to evaluate the content complexity of multimodal early warning information. The characteristics of multimodal early warning information are analyzed in combination with the data volume and urgency of the information. The priority, required bandwidth and redundancy strategy for the transmission of multimodal early warning information are determined to obtain a transmission optimization scheme. The transmission optimization scheme is weighted by the quality of service parameters to obtain the information transmission strategy. Based on information transmission strategies, machine learning algorithms are applied to monitor the status of multimodal early warning broadcast terminals in the affected area in real time, predict the changing trends of multimodal early warning broadcast terminals, and combine deep learning models to identify network congestion points and high-latency areas of multimodal early warning broadcast terminals, forming a network topology map, adjusting the sudden situation of the multimodal early warning broadcast terminal topology map, and generating a dynamic network status map. Based on the dynamic network state diagram and information transmission strategy, the Dijkstra algorithm or A* search algorithm in the path planning algorithm is used to calculate the transmission path from the early warning information release system to the selected multimodal early warning broadcast terminal, and a genetic algorithm is introduced to optimize the transmission path to obtain the optimal transmission path. To address the segment characteristics of multimodal early warning broadcasting terminals on the optimal transmission path, adaptive modulation and coding techniques and a channel state information feedback mechanism are employed to encode and optimize multimodal early warning information, thereby generating optimized multimodal early warning information. The selected multimodal early warning broadcasting terminal sends operation instructions, transmission time windows, and error retransmission mechanisms to optimize multimodal early warning information, thereby obtaining a transmission instruction set. Based on the transmission instruction set, a transmission scheme document is compiled, and a risk assessment and countermeasures are added to the transmission scheme document to obtain a transmission scheme. The transmission scheme document includes: the optimal transmission path, encoding method, transmission instructions and emergency response plan. The processed multimodal early warning information is sent to the selected multimodal early warning broadcasting terminal through the transmission scheme. Based on the modality selection algorithm, appropriate multimodal warning information is selected for visual information presentation, and augmented reality technology is introduced to support the visual information presentation to obtain feedback data, including user understanding and reaction speed. Collect and analyze feedback data, evaluate the effectiveness of multimodal early warning information dissemination, and adjust the dissemination scope of multimodal early warning information based on the dissemination effectiveness.

2. The method according to claim 1, characterized in that, The method involves evaluating the content complexity of the multimodal early warning information using an information entropy algorithm, analyzing its characteristics in conjunction with the data volume and urgency, determining the transmission priority, required bandwidth, and redundancy strategy of the multimodal early warning information, and obtaining a transmission optimization scheme, including: The multimodal early warning information is processed using the information entropy algorithm complexity assessment to obtain a quantitative result of the content complexity; The quantitative results of the content complexity are analyzed using the data volume and urgency of the multimodal early warning information to obtain a specific characteristic description of each multimodal early warning information. Based on the specific characteristic description, the priority processing of the multimodal early warning information transmission is determined, and an information transmission priority list is generated; Based on the information transmission priority list, the required bandwidth is estimated and the minimum bandwidth requirement scheme is obtained. For the minimum bandwidth requirement scheme, a redundancy strategy is formulated to ensure reliability and generate a redundant transmission plan; the multimodal early warning information will be transmitted through other redundancy mechanisms to cope with network failures and congestion, and a transmission optimization scheme will be formed based on the redundant transmission plan.

3. The method according to claim 2, characterized in that, The step involves calculating the transmission path from the early warning information dissemination system to the selected multimodal early warning broadcasting terminal based on the dynamic network state diagram and the information transmission strategy, using either Dijkstra's algorithm or A* search algorithm in path planning algorithms, and then introducing a genetic algorithm to optimize the transmission path to obtain the optimal transmission path, including: The status of the multimodal early warning broadcasting terminal network within the affected area is comprehensively evaluated using the dynamic network status diagram and the information transmission strategy to obtain basic data; Based on the aforementioned basic data, the path from the early warning information release system to the selected multimodal early warning broadcasting terminal is calculated using the Dijkstra algorithm or A* search algorithm in the path planning algorithm, and a preliminary transmission path is generated. A genetic algorithm is introduced to iteratively optimize the initial transmission path, select the transmission path with the best performance, and generate candidate transmission paths. By evaluating the performance of the candidate transmission paths under network congestion points, high latency areas, and sudden events, the stability and anti-interference capabilities of the candidate transmission paths are analyzed, and important transmission paths with high efficiency and stability are obtained. Based on the important transmission paths, the optimal transmission path is generated by combining the network topology, the status of the multimodal early warning broadcasting terminal, the urgency of the event, and the information transmission priority.

4. The method according to claim 3, characterized in that, Also includes: The transmission optimization scheme is weighted and adjusted using the service quality parameters. A fault prediction and health management system is introduced to obtain an information transmission strategy and monitor the health status during transmission in real time. When an anomaly is detected, the health management system can respond quickly, take preventive measures, monitor potential risks during transmission, and automatically adjust the information transmission strategy.

5. The method according to claim 4, characterized in that, The monitoring of potential risks during the transmission process and the automatic adjustment of information transmission strategies include: Continuously monitor each stage of the transmission execution plan, collect and analyze network performance indicators and terminal feedback data to obtain potential risk factors. Among them, the multimodal early warning broadcast terminal performance indicators include latency, packet loss rate and bandwidth utilization. Machine learning algorithms are used to predict the probability of occurrence and the scope of impact of potential risks in the terminal feedback data, provide early warning of fault points or bottlenecks, and generate risk assessment reports. The information transmission strategy is automatically adjusted based on the risk assessment report. After high-risk areas are detected, the transmission path is replanned or redundant transmission channels are added. Service quality parameters are dynamically adjusted to prioritize the transmission of the multimodal early warning information.

6. The method according to claim 1, characterized in that, The process involves analyzing the dynamic environmental parameters using a context-aware algorithm to obtain abnormal data identification results. Based on the event type information and the abnormal data identification results, personalized broadcast content suitable for the geographical location information is matched from a preset emergency response plan database, including: Receive emergency event signals triggered by users and the early warning information release system, extract and parse the event type information, geographical location information and dynamic environmental parameters in the emergency event signals, wherein the dynamic environmental parameters include weather conditions, traffic flow and population density; The dynamic environmental parameters are analyzed using a context-aware algorithm to assess the impact of current environmental conditions on emergency response and obtain environmental anomaly data identification results. Based on the event type information and the environmental anomaly data identification results, personalized broadcast content suitable for the geographical location information is matched from the preset emergency response plan library, and the emergency response plan library is filtered to obtain personalized broadcast content.

7. A machine learning-based anomaly data identification system, used to implement the machine learning-based anomaly data identification method as described in any one of claims 1 to 6, characterized in that, include: The receiving module is used to receive emergency event signals triggered by users and the early warning information release system, wherein the emergency event signals carry event type information, geographical location information, and dynamic environmental parameters; The analysis module uses a context-aware algorithm to analyze the dynamic environmental parameters, obtain abnormal data identification results, and matches personalized broadcast content suitable for the geographical location information in a preset emergency response plan library based on the event type information and the abnormal data identification results, and adjusts the applicability and priority order of the emergency response plan library to obtain an emergency response strategy. The module is used to integrate text, voice, and visual signals from the emergency response plan library based on the emergency response strategy, construct multimodal early warning information, and optimize the multimodal early warning information according to specific scenarios. The calculation module is used to calculate the optimal transmission path using a path planning algorithm based on the characteristics of the multimodal early warning information and the network status of the multimodal early warning broadcasting terminals in the affected area, encode and process the multimodal early warning information using adaptive modulation and coding technology, generate a transmission scheme based on the processed multimodal early warning information and the optimal transmission path, and send the processed multimodal early warning information to the selected multimodal early warning broadcasting terminals through the transmission scheme. The feedback module is used to, after the multimodal warning information is acquired and processed by the multimodal warning broadcasting terminal, adapt and process it based on the modality selection algorithm and select appropriate multimodal warning information for visual information presentation, and introduce augmented reality technology to support the visual information presentation to obtain feedback data, wherein the feedback data includes user understanding and reaction speed. An evaluation module is used to collect and analyze the feedback data, evaluate the transmission effect of the multimodal early warning information, and adjust the transmission range of the multimodal early warning information based on the transmission effect.

8. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement a machine learning-based abnormal data identification method as described in any one of claims 1 to 6.

9. A computer storage medium, characterized in that, The device contains a computer program that, when executed by a computer, implements a machine learning-based anomaly data identification method as described in any one of claims 1 to 6.

Images