Outdoor disaster emergency broadcast playing method

Abstract

The invention relates to the technical field of emergency broadcast, and discloses an outdoor disaster emergency broadcast playing method. According to the method, multi-type sensing, interaction and terminal equipment are deployed in advance, and full-process emergency response of environmental sensing, feedback identification, rescue scheduling and closed-loop updating is completed according to four stages of t0-t3; a multi-mode feedback signal recognition mechanism is innovatively adopted, accurate matching of rescue resources is achieved in combination with AI intelligent scheduling, and meanwhile efficient and safe promotion of rescue actions is guaranteed through rescue closed-loop instructions and dynamic risk updating. The problems that traditional outdoor emergency broadcast is not timely in feedback, rescue dispatching is blind, risk prevention and control are lagged and the like are solved, the method is particularly suitable for disaster rescue scenes of large-range and complex terrains such as mountain forests, and the response speed and success rate of emergency rescue are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of emergency broadcasting, in particular to an outdoor disaster emergency broadcasting playing method, which is especially suitable for disaster search and rescue work in large-scale, signal unstable scenes such as mountains, forests and wilderness. BACKGROUND

[0002] In outdoor scenes such as mountains, disasters such as earthquakes, mountain torrents, mudslides and forest fires occur frequently. Such scenes often have complex terrain and weak communication infrastructure. Once a disaster occurs, it is easy to have situations such as people trapped, roads blocked and signal interrupted. The existing outdoor emergency broadcasting method has many defects: first, the emergency response is mostly one-way broadcasting, lacking a two-way interactive channel between the trapped people and the rescue system, so it is difficult for rescue personnel to accurately grasp the specific needs and location of the trapped people; second, rescue dispatch relies on human experience and cannot be combined with real-time environmental data and rescue resource status to achieve optimal matching, resulting in low rescue efficiency, and even the rescue team may mistakenly enter a dangerous area; third, the feedback method is single and has poor adaptability. In the scene without network and electronic terminal, the trapped people cannot obtain feedback guidance, and the rescue system also cannot identify the trapped people's signal; fourth, the environmental risk update is lagging behind, and new risks such as temporary rockfall and road collapse discovered during the rescue process cannot be synchronized in time, which may cause secondary disasters.

[0003] Therefore, there is an urgent need for an outdoor disaster emergency broadcasting playing method that can realize two-way interaction, intelligent dispatch, multi-scene adaptation and dynamic risk prevention to meet the actual needs of mountain search and rescue. SUMMARY

[0004] The purpose of the present application is to overcome the shortcomings of the prior art and provide an outdoor disaster emergency broadcasting playing method to achieve the following objectives: 1. Build a two-way interactive channel between the trapped people and the rescue system to accurately identify the needs and location of the trapped people; 2. Optimal matching of rescue resources through AI intelligent dispatch to improve rescue efficiency; 3. Adapt to complex scenes such as network / no network, terminal / no terminal, etc. to ensure full coverage of feedback guidance; 4. Real-time update of environmental risks to ensure the safety of rescue personnel and avoid secondary disasters.

[0005] To solve the above technical problems, the present application adopts the following scheme: An outdoor disaster emergency broadcasting playing method, comprising the following steps: Basic deployment step: dividing the outdoor scene into multiple areas, each area is provided with fixed emergency broadcasting, fixed camera, environmental perception UAV group, trapped person feedback identification module, rescue team intelligent terminal and AI intelligent dispatch module; S1: Environmental Awareness and Initial Broadcast Steps: At time t0, receive emergency commands generated by outdoor disasters, activate the fixed camera to collect images of people in the area, and simultaneously dispatch the environmental awareness drone swarm to collect data and generate a dynamic environmental map; combine the dynamic environmental map to generate a first broadcast list adapted to the current environment, and play the first broadcast list through the fixed emergency broadcast; at the same time, push multimodal feedback guidelines to user terminals in the area, in which the multimodal feedback guidelines clearly define the feedback meanings corresponding to flashing lights, clapping, and whistling respectively; S2: Trapped Person Feedback Identification and Broadcast Update Steps: At time t1, the trapped person feedback identification module identifies the feedback signals transmitted by the trapped person according to the multimodal feedback guidelines and sends out the trapped person's location based on the feedback signals; the feedback signals are divided into medical need feedback, route obstruction feedback, and instruction confirmation feedback according to their priority from high to low; the high-priority feedback is synchronized to the command vehicle control terminal; if the route obstruction feedback is identified, the first broadcast list is updated to an environment-feedback adapted second broadcast list in combination with the dynamic environment map and the route obstruction feedback, and the second broadcast list is played through the fixed emergency broadcast; S3: Rescue Dispatch and Directional Command Issuance Steps: At time t2, the AI ​​intelligent dispatch module associates the following three types of information: the dynamic environment map generated in S1, the location and feedback needs of the trapped person identified in S2, and the real-time location and equipment information of the rescue team reported by the rescue team's intelligent terminal; based on the three types of information, the optimal rescue team is matched as the target rescue team, and the corresponding rescue route is planned, generating customized rescue commands and trapped person comfort commands, forming a third broadcast list; the trapped person comfort commands are broadcast to the trapped person through the fixed emergency broadcast in the area where the trapped person is located, and the customized rescue commands are issued to the rescuers through the intelligent terminal of the target rescue team; S4: Rescue closed-loop and dynamic update steps: At time t3, receive the signal of the trapped person who has been contacted uploaded by the target rescue team through the rescue team's smart terminal, automatically generate a rescue closed-loop instruction and play it through the fixed emergency broadcast; if the rescue team uploads new environmental risk information, update the dynamic environment map and the currently playing broadcast list in combination with the new environmental risk information, and at the same time issue obstacle avoidance prompts to other rescue teams.

[0006] This outdoor disaster emergency broadcasting method first establishes a multi-dimensional collaborative hardware system through basic deployment, encompassing fixed emergency broadcasting equipment, cameras, and environmental sensing drone swarms. Then, it advances the entire process of emergency broadcasting and rescue coordination along the timeline from t0 to t3, sequentially completing environmental sensing and initial broadcasting (generating a dynamic environmental map, a first broadcast list, and pushing multimodal feedback guidelines), identification of trapped individuals' feedback and broadcast updates (processing feedback signals according to priority and generating a second broadcast list), rescue dispatch and directional instruction issuance (integrating 3D information to match rescue teams and generating a third broadcast list), and rescue closure and dynamic updates (generating rescue closure instructions and updating content based on new risks). Its innovation lies in constructing a multimodal feedback system and priority mechanism based on "flashlights + clapping + whistles," breaking through the limitations of traditional one-way broadcasting. It builds an AI-powered intelligent dispatching and directional broadcasting collaborative architecture that integrates 3D information, solving the problem of broadcasting and rescue being disconnected. Simultaneously, it establishes a full-process closed-loop iterative mechanism, overcoming the defects of static push notifications, effectively improving the real-time, targeted, and reliable nature of emergency response, and addressing the core technical pain points of traditional emergency broadcasting.

[0007] A further preferred technical solution is as follows: In step S1, the data collected by the environmental perception drone swarm includes smoke concentration, ambient temperature, water depth, and rockfall risk level; the dynamic environmental map is updated at a frequency of "1 minute / time"; if the environmental parameters of a certain area exceed the safety threshold, the area will be automatically marked as a red high-risk area in the dynamic environmental map, and a warning message "Entry into this red high-risk area is prohibited" will be added to the first broadcast list.

[0008] A further preferred technical solution is as follows: In step S1, the method of pushing multimodal feedback guidelines to user terminals within the area is divided into two cases: When network access is available: push notifications to user terminals via SMS and emergency apps; In the absence of network access: Use a swarm of environmentally aware drones to deliver waterproof paper cards with instructions printed on them at low altitudes, or use a fixed emergency broadcast system to continuously broadcast the instructions; ensure that those stranded without electronic devices can still obtain feedback.

[0009] This further optimized technical solution refines the method of pushing multimodal feedback guidelines to user terminals within the area in step S1, specifically dividing it into differentiated implementations for scenarios with and without network access: when there is network access, the guidelines are directly pushed to user terminals via SMS and an emergency app; when there is no network access, on the one hand, a swarm of environmentally aware drones delivers waterproof paper cards printed with the guidelines at low altitude, and on the other hand, the guidelines are broadcast cyclically via a fixed emergency broadcast system. Through this online and offline collaborative push method, it is effectively ensured that even those trapped without electronic terminals can fully obtain the corresponding feedback.

[0010] A further preferred technical solution is as follows: the step of issuing the location of the trapped person based on the feedback signal in S2 includes: locating the sound source or light source of the flash, clapping or whistling through the image data of the fixed camera and / or the data collected by the environmentally aware drone swarm, thereby determining the location of the trapped person.

[0011] A further preferred technical solution is as follows: In step S2, the specific identification method of the trapped person feedback identification module is as follows: the flashing frequency of the trapped person's flashlight is analyzed by an image recognition algorithm, and the clapping rhythm and whistling frequency of the trapped person are analyzed by a sound feature extraction algorithm; if the confidence level of the identification result of a certain feedback signal is lower than a preset threshold, the environmental perception drone swarm will be automatically dispatched to take a close-up picture or record a sound in the area where the feedback signal was emitted, thereby improving the identification accuracy.

[0012] A further preferred technical solution is as follows: In step S2, the specific correspondence between the multimodal feedback signal and the feedback requirement is as follows: Medical needs feedback: corresponds to "5 flashes of the light" or "3 long whistle". Route disruption feedback: corresponds to "clapping twice (short) and once (long)" or "the flash blinks twice, pauses for 1 second, and then flashes once more"; Command confirmation feedback: corresponds to "flash 3 times" or "clap 1 short 1 long".

[0013] The design of this signal correspondence revolves around the dual objectives of "adaptability to outdoor disaster scenarios" and "accuracy of recognition." On the one hand, considering the differences in equipment conditions that trapped individuals may face (e.g., some have electronic devices with flashlights, while others lack electronic devices and can only provide feedback through limbs / acoustic signals), a dual option of "electronic signal + non-electronic signal" is configured for each type of need. This ensures that trapped individuals without terminals or power sources can effectively transmit information, covering complex scenarios such as "with / without electronic devices" and "with / without light sources." On the other hand, differentiated signal characteristics enhance distinguishability—"5 flashes" and "3-second long beep" for medical need feedback; "2 flashes + pause + 1 flash" and "2 short 1 long clapping" for route obstruction feedback; and "3 flashes" and "1 short 1 long clapping" for instruction confirmation feedback. The "long clapping" signal features significant differences in frequency, rhythm, and duration, effectively avoiding interference from outdoor noise (such as wind and rain, collapse sounds) and stray light (such as firelight, reflections) on recognition. Combined with the customized algorithm of the trapped person feedback recognition module, the signal misidentification rate can be controlled below 5%. Furthermore, the signal design follows the principle of "simple and easy to remember, with no operational barriers," requiring no complex learning from the trapped person. Feedback can be completed using only basic body movements or common objects (such as a mobile phone flash or whistle), making it particularly suitable for the tension and panic that trapped persons may experience in disaster scenarios. This ensures that feedback actions are executed quickly and accurately, providing precise and reliable demand information for subsequent S2 feedback recognition and S3 rescue dispatch.

[0014] A further preferred technical solution is as follows: In step S3, the rule for the AI ​​intelligent dispatch module to match the optimal rescue group is: First priority: Select rescue teams that are "≤1.5 km away from the stranded people" and "carry equipment suitable for the needs of the stranded people"; If multiple rescue teams meet the first priority criteria, they will be selected in the following order: priority will be given to teams with "low risk level of rescue route", and then priority will be given to teams with "remaining physical strength ≥ 60%".

[0015] A further preferred technical solution is as follows: In step S3, the customized rescue instruction, in addition to including the target location and route of the rescue team, also includes two types of information: The latest status of the trapped person's feedback signals; key nodes of the rescue route marked on the dynamic environment map; assisting the rescue team to reach the area where the trapped person is located more efficiently.

[0016] A further preferred technical solution is as follows: In step S4, the rescue team uploads new environmental risk information, including temporary rockfalls, road collapses, and sudden water accumulation; when uploading, the specific location and impact range of the risk must be marked simultaneously; the system update response time for the current broadcast list shall not exceed 30 seconds, and the obstacle avoidance prompts shall clearly state two items: first, the coordinates of the risk area where passage is prohibited, and second, the recommended detour route.

[0017] The beneficial effects of this invention are as follows: 1. Precise two-way interaction: A multimodal feedback system based on flashlights, clapping, and whistles was constructed, combined with image recognition and sound feature extraction algorithms, to achieve precise identification of the needs and location of the trapped person, solving the drawbacks of traditional one-way emergency broadcasting; 2. Intelligent Dispatch of Rescue: By integrating multi-dimensional information through the AI ​​intelligent dispatch module, the optimal rescue team is matched according to scientific rules, avoiding the blindness of manual dispatch, shortening rescue time, and improving rescue efficiency; 3. Comprehensive scenario adaptation: Different feedback guide push methods are designed for scenarios with / without network and with / without terminal. At the same time, the flexible use of drone swarms can make up for the coverage blind spots of fixed equipment and adapt to complex outdoor environments such as mountains and forests. 4. Real-time risk prevention and control: The closed-loop rescue command ensures that each rescue is completed from start to finish. The rapid updating of new environmental risks and the issuance of obstacle avoidance prompts effectively avoid secondary disasters and ensure the safety of rescuers and trapped people. 5. Simplified Operation: All equipment is deployed in a standardized manner, and the feedback signal and instruction generation process is fixed. Rescuers and trapped individuals can operate the equipment without complicated training, which lowers the threshold for emergency response. Attached Figure Description

[0018] Figure 1 This is a flowchart of the outdoor disaster emergency broadcasting method of the present invention; Figure 2 A schematic diagram illustrating the specific correspondence between multimodal feedback signals and feedback requirements in the context of medical demand feedback; Figure 3 A schematic diagram of route blocking feedback in illustrating the specific correspondence between multimodal feedback signals and feedback requirements; Figure 4 This is a schematic diagram illustrating the specific correspondence between multimodal feedback signals and feedback requirements, specifically the instruction confirmation feedback. Detailed Implementation

[0019] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0020] like Figure 1 As shown, Figure 1The flowchart of the outdoor disaster emergency broadcasting method of the present invention includes the following steps: Basic deployment steps: Divide the outdoor scene into multiple areas, and deploy a fixed emergency broadcast, fixed cameras, environmental perception drone swarm, trapped person feedback recognition module, rescue team smart terminal and AI intelligent dispatch module in each area; S1: Environmental Awareness and Initial Broadcast Steps: At time t0, receive emergency commands generated by outdoor disasters, activate the fixed camera to collect images of people in the area, and simultaneously dispatch the environmental awareness drone swarm to collect data and generate a dynamic environmental map; combine the dynamic environmental map to generate a first broadcast list adapted to the current environment, and play the first broadcast list through the fixed emergency broadcast; at the same time, push multimodal feedback guidelines to user terminals in the area, in which the multimodal feedback guidelines clearly define the feedback meanings corresponding to flashing lights, clapping, and whistling respectively; S2: Trapped Person Feedback Identification and Broadcast Update Steps: At time t1, the trapped person feedback identification module identifies the feedback signals transmitted by the trapped person according to the multimodal feedback guidelines and sends out the trapped person's location based on the feedback signals; the feedback signals are divided into medical need feedback, route obstruction feedback, and instruction confirmation feedback according to their priority from high to low; the high-priority feedback is synchronized to the command vehicle control terminal; if the route obstruction feedback is identified, the first broadcast list is updated to an environment-feedback adapted second broadcast list in combination with the dynamic environment map and the route obstruction feedback, and the second broadcast list is played through the fixed emergency broadcast; S3: Rescue Dispatch and Directional Command Issuance Steps: At time t2, the AI ​​intelligent dispatch module associates the following three types of information: the dynamic environment map generated in S1, the location and feedback needs of the trapped person identified in S2, and the real-time location and equipment information of the rescue team reported by the rescue team's intelligent terminal; based on the three types of information, the optimal rescue team is matched as the target rescue team, and the corresponding rescue route is planned, generating customized rescue commands and trapped person comfort commands, forming a third broadcast list; the trapped person comfort commands are broadcast to the trapped person through the fixed emergency broadcast in the area where the trapped person is located, and the customized rescue commands are issued to the rescuers through the intelligent terminal of the target rescue team; S4: Rescue closed-loop and dynamic update steps: At time t3, receive the signal of the trapped person who has been contacted uploaded by the target rescue team through the rescue team's smart terminal, automatically generate a rescue closed-loop instruction and play it through the fixed emergency broadcast; if the rescue team uploads new environmental risk information, update the dynamic environment map and the currently playing broadcast list in combination with the new environmental risk information, and at the same time issue obstacle avoidance prompts to other rescue teams.

[0021] A further preferred technical solution is as follows: In step S1, the data collected by the environmental perception drone swarm includes smoke concentration, ambient temperature, water depth, and rockfall risk level; the dynamic environmental map is updated at a frequency of "1 minute / time"; if the environmental parameters of a certain area exceed the safety threshold, the area will be automatically marked as a red high-risk area in the dynamic environmental map, and a warning message "Entry into this red high-risk area is prohibited" will be added to the first broadcast list.

[0022] A further preferred technical solution is as follows: In step S1, the method of pushing multimodal feedback guidelines to user terminals within the area is divided into two cases: When network access is available: push notifications to user terminals via SMS and emergency apps; In the absence of network access: Use a swarm of environmentally aware drones to deliver waterproof paper cards with instructions printed on them at low altitudes, or use a fixed emergency broadcast system to continuously broadcast the instructions; ensure that those stranded without electronic devices can still obtain feedback.

[0023] A further preferred technical solution is as follows: the step of issuing the location of the trapped person based on the feedback signal in S2 includes: locating the sound source or light source of the flash, clapping or whistling through the image data of the fixed camera and / or the data collected by the environmentally aware drone swarm, thereby determining the location of the trapped person.

[0024] A further preferred technical solution is as follows: In step S2, the specific identification method of the trapped person feedback identification module is as follows: the flashing frequency of the trapped person's flashlight is analyzed by an image recognition algorithm, and the clapping rhythm and whistling frequency of the trapped person are analyzed by a sound feature extraction algorithm; if the confidence level of the identification result of a certain feedback signal is lower than a preset threshold, the environmental perception drone swarm will be automatically dispatched to take a close-up picture or record a sound in the area where the feedback signal was emitted, thereby improving the identification accuracy.

[0025] like Figures 2-4 As shown, a further preferred technical solution is as follows: In step S2, the specific correspondence between the multimodal feedback signal and the feedback requirement is as follows: Medical needs feedback: corresponds to "5 flashes of the light" or "3 long whistle". Route disruption feedback: corresponds to "clapping twice (short) and once (long)" or "the flash blinks twice, pauses for 1 second, and then flashes once more"; Command confirmation feedback: corresponds to "flash 3 times" or "clap 1 short 1 long".

[0026] A further preferred technical solution is as follows: In step S3, the rule for the AI ​​intelligent dispatch module to match the optimal rescue group is: First priority: Select rescue teams that are "≤1.5 km away from the stranded people" and "carry equipment suitable for the needs of the stranded people"; If multiple rescue teams meet the first priority criteria, they will be selected in the following order: priority will be given to teams with "low risk level of rescue route", and then priority will be given to teams with "remaining physical strength ≥ 60%".

[0027] A further preferred technical solution is as follows: In step S3, the customized rescue instruction, in addition to including the target location and route of the rescue team, also includes two types of information: The latest status of the trapped person's feedback signals; key nodes of the rescue route marked on the dynamic environment map; assisting the rescue team to reach the area where the trapped person is located more efficiently.

[0028] A further preferred technical solution is as follows: In step S4, the rescue team uploads new environmental risk information, including temporary rockfalls, road collapses, and sudden water accumulation; when uploading, the specific location and impact range of the risk must be marked simultaneously; the system update response time for the current broadcast list shall not exceed 30 seconds, and the obstacle avoidance prompts shall clearly state two items: first, the coordinates of the risk area where passage is prohibited, and second, the recommended detour route.

[0029] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Based on the technical essence of the present invention, any simple modifications, equivalent substitutions, and improvements made to the above embodiments within the spirit and principles of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An outdoor disaster emergency broadcast play method, characterized by, Comprising the following steps: Base deployment step: divide the outdoor scene into multiple areas, each of which is deployed with fixed emergency broadcasting, fixed camera, environmental perception UAV group, trapped person feedback identification module, rescue group intelligent terminal and AI intelligent scheduling module; S1: environmental perception and initial broadcasting step: at t0, receive the emergency instruction generated by the outdoor disaster, start the fixed camera to collect the images of the personnel in the area, and simultaneously dispatch the environmental perception UAV group to collect data and generate a dynamic environment map; combine the dynamic environment map to generate a first broadcast list adapted to the current environment, and play the first broadcast list through the fixed emergency broadcasting; at the same time, push a multi-modal feedback guide to the user terminal in the area, and the multi-modal feedback guide clearly indicates the feedback meanings corresponding to the flashlight, hand clapping and whistling respectively; S2: trapped person feedback identification and broadcast updating step: at t1, identify the feedback signals transmitted by the trapped persons according to the multi-modal feedback guide through the trapped person feedback identification module, and output the trapped person positions based on the feedback signals; sort the feedback signals from high to low priority as medical demand feedback, route blocking feedback and instruction confirmation feedback; synchronize the high priority feedback to the control end of the command vehicle; if the route blocking feedback is identified, update the first broadcast list to an environment-feedback adaptive second broadcast list in combination with the dynamic environment map and the route blocking feedback, and play the second broadcast list through the fixed emergency broadcasting; S3: rescue scheduling and directional instruction issuing step: at t2, associate the following three types of information through the AI intelligent scheduling module: the dynamic environment map generated in S1, the trapped person positions and feedback demands identified in S2, and the real-time positions and carried equipment information of the rescue groups reported on the rescue group intelligent terminal; based on the three types of information, match the optimal rescue group as the target rescue group, and plan the corresponding rescue route to generate customized rescue instructions and trapped person calming instructions, forming a third broadcast list; play the trapped person calming instructions to the trapped persons through the fixed emergency broadcasting in the area where the trapped persons are located, and issue the customized rescue instructions to the rescue personnel through the rescue group intelligent terminal of the target rescue group; S4: rescue closed loop and dynamic updating step: at t3, receive the contacted trapped person signals uploaded by the target rescue group through the rescue group intelligent terminal, automatically generate rescue closed loop instructions and play them through the fixed emergency broadcasting; if the rescue group uploads new environmental risk information, update the dynamic environment map and the currently played broadcast list in combination with the new environmental risk information, and simultaneously issue an obstacle avoidance prompt to other rescue groups.

2. The outdoor disaster emergency broadcasting play method according to claim 1, characterized in that, In step S1, the data collected by the environmental perception UAV group includes smoke concentration, environmental temperature, water depth, and rockfall risk level; the dynamic environment map is updated at a frequency of "1 minute / time"; if the environmental parameters of a certain area exceed the safety threshold, the area will be automatically marked as a red high-risk area in the dynamic environment map, and the warning content "prohibit entering the red high-risk area" will be added to the first broadcast list.

3. The outdoor disaster emergency broadcasting play method according to claim 1, characterized in that, In step S1, the multi-modal feedback guide pushed to the user terminal in the area is in two cases: With network: push to the user terminal through SMS and emergency APP; Without network: through the low-altitude delivery of waterproof paper cards printed with guides by the environmental perception UAV group, or through the fixed emergency broadcast to play the guide content in a loop; ensure that the trapped people without electronic terminals can also obtain the feedback method.

4. The outdoor disaster emergency broadcasting play method according to claim 1, characterized in that, The step of issuing the trapped person's location based on the feedback signal in S2 includes: positioning the sound or light source of the flashlight, hand clapping or whistling through the image data of the fixed camera and / or the data collected by the environmental perception UAV group, to determine the location of the trapped person.

5. The outdoor disaster emergency broadcasting play method according to claim 1, characterized in that, In step S2, the specific identification method of the trapped person feedback recognition module is: analyzing the flashlight blinking frequency of the trapped person through image recognition algorithm, and analyzing the hand clapping rhythm and whistle frequency of the trapped person through sound feature extraction algorithm; if the recognition result of a feedback signal has a confidence level lower than the preset threshold, the environmental perception UAV group will be automatically dispatched to take close-up images or record sounds to improve the recognition accuracy.

6. The outdoor disaster emergency broadcasting play method according to claim 1, characterized in that, In step S2, the specific correspondence between multi-modal feedback signals and feedback needs is as follows: Medical demand feedback: corresponds to "flashlight flashes 5 times" or "whistle sounds for 3 seconds"; Route blocking feedback: corresponds to "clap 2 short 1 long" or "flashlight flashes 2 times, pauses for 1 second, and flashes again"; Instruction confirmation feedback: corresponds to "flashlight flashes 3 times" or "clap 1 short 1 long".

7. The outdoor disaster emergency broadcasting play method according to claim 1, characterized in that, In step S3, the rules for the AI intelligent scheduling module to match the optimal rescue team are: First priority: select the rescue team that is "within 1.5 kilometers of the trapped person" and "carries equipment suitable for the trapped person's needs"; If there are multiple rescue teams that meet the first priority, select them in the following order: first, select the team with "low risk level of rescue route"; second, select the team with "remaining physical strength ≥ 60%".

8. The outdoor disaster emergency broadcasting play method according to claim 1, characterized in that, In step S3, the customized rescue instructions include not only the target position and route of the rescue team, but also two additional types of information: The latest state of the trapped person's feedback signal; the rescue route key nodes marked in the dynamic environment map; assist the rescue team to arrive at the area where the trapped person is located more efficiently.

9. The outdoor disaster emergency broadcasting play method according to claim 1, characterized in that, In step S4, the rescue team uploads new environmental risk information including temporary rockfall, road collapse, and sudden water accumulation. The specific location and impact range of the risk need to be labeled synchronously. The system updates the current broadcast list within 30 seconds, and the obstacle avoidance prompt needs to clearly indicate two items: one is the risk area coordinates that are prohibited from passing through, and the other is the recommended detour route.

Images