Interphone auxiliary rescue method and system based on rescue scene fall detection

Abstract

The invention discloses an interphone auxiliary rescue method and system based on rescue scene fall detection, and the method comprises the steps: obtaining rescue state collection information, carrying out the preprocessing, and constructing an operation environment data sequence, a communication state data sequence, and a motion state data sequence; constructing an anomaly detection model, and inputting the motion state data sequence to analyze abnormal motion state data in the real-time rescue task; performing fall detection on the target interphone user, and judging whether a fall condition exists or not; if the tumble condition exists, tumble risk assessment is carried out in combination with the tumble early warning information, and whether collaborative rescue needs to be carried out or not is judged; and if cooperative rescue needs to be carried out, interphones of other rescuers are searched in the current communication environment for cooperative rescue, and communication optimization is carried out in combination with the operation environment data sequence and the communication state data sequence so as to expand the communication range. Therefore, safety accidents caused by the fact that rescue workers cannot ask for help autonomously in emergency situations in rescue actions are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of walkie-talkie-assisted rescue, and in particular to a walkie-talkie-assisted rescue method and system based on fall detection in rescue scenarios. Background Art

[0002] In outdoor rescue operations, industrial operations, and high-risk environments, practitioners and individuals are increasingly using walkie-talkies, relying on their convenient voice communication capabilities to promptly transmit important information. However, in these scenarios, the possibility of accidental falls or loss of mobility is high, especially in complex terrain or extreme environments. When a fall occurs, failure to call for help in a timely manner can put rescuers in danger, leading to serious consequences. When a user is in a fallen state and their ability to call for help is limited, traditional communication devices are unable to effectively locate and identify the risk and autonomously call for help. This isolated communication problem is common in real-world applications and poses a potential threat to the safety of rescuers. Therefore, improving the safety and searchability of rescuers during rescue missions is an urgent issue that needs to be addressed. Summary of the Invention

[0003] The present invention overcomes the defects of the prior art and provides a walkie-talkie assisted rescue method and system based on fall detection in rescue scenarios, the important purpose of which is to improve the safety and searchability of rescue personnel when performing rescue missions.

[0004] To achieve the above objectives, the present invention provides a first aspect of a walkie-talkie-assisted rescue method based on fall detection in a rescue scenario, comprising: Collect rescue status data for real-time rescue missions to obtain rescue status information and preprocess it, time-sequence the preprocessed data, and construct operating environment data sequence, communication status data sequence, and motion status data sequence; An anomaly detection model is constructed based on a generative adversarial network, and the motion state data sequence is input to analyze abnormal motion state data in real-time rescue missions to obtain abnormal motion state analysis information; Performing fall detection on the target intercom user based on the abnormal motion state analysis information, determining whether there is a fall condition, and generating fall warning information; If a fall is detected for the current intercom user, a fall risk assessment is performed based on the fall warning information to determine whether collaborative rescue is required; If collaborative rescue is required, the walkie-talkies of other rescuers will be searched for in the current communication environment for collaborative rescue, and communication optimization will be performed in combination with the operating environment data sequence and the communication status data sequence to expand the communication range.

[0005] In this solution, the rescue status data of the real-time rescue mission is collected to obtain the rescue status information and pre-process it, the pre-processed data is time-sequenced, and the operation environment data sequence, communication status data sequence and motion status data sequence are constructed, specifically including: The sensor array set based on the target walkie-talkie performs rescue status data collection for real-time rescue tasks to obtain rescue status collection information, wherein the rescue status collection information includes walkie-talkie operating environment data, walkie-talkie communication status data and walkie-talkie motion status data; Performing data preprocessing on the rescue status collected information, calculating the data standard deviation of the rescue status collected information, and comparing the calculated data standard deviation with a preset standard deviation range; The rescue state collected data corresponding to the data standard deviation that is not within the preset standard deviation range is defined as abnormal data, outliers are removed and linear interpolation method is introduced to supplement the difference to obtain preprocessing result information; Time series features are extracted based on the preprocessing result information, and time series processing is performed on the preprocessing result information according to the extracted time series features to generate an operating environment data sequence, a communication status data sequence, and a motion status data sequence.

[0006] In this solution, the anomaly detection model is constructed based on the generative adversarial network, and the motion state data sequence is input to analyze the abnormal motion state data in the real-time rescue mission to obtain abnormal motion state analysis information, which specifically includes: Building an anomaly detection model based on a generative adversarial network, obtaining a motion state data sequence, and inputting the motion state data sequence as an original sequence into the anomaly detection model for analysis; The generator extracts features from the input motion state data sequence, reconstructs the sequence based on the extracted real-time intercom motion state features, generates a reconstructed motion state sequence, and inputs the reconstructed motion state sequence into the discriminator for judgment; The discriminator determines whether the current reconstructed motion state sequence can be constructed. If the discriminator determines that the current reconstructed motion state sequence is unacceptable, the current reconstructed motion state sequence is marked as an incorrect reconstructed motion state sequence, and the next reconstructed motion state sequence is obtained for judgment. The reconstructed motion state sequence generated by the generator is iteratively and repeatedly judged to select a reconstructed motion state sequence that meets a preset acceptance criterion, thereby obtaining a final reconstructed motion state sequence; Calculating a reconstruction error between the final reconstructed motion state sequence and the original sequence, and comparing the calculated reconstruction error with a preset threshold. If the calculated reconstruction error is greater than the preset threshold, it indicates that the current real-time motion state is abnormal. The final reconstructed motion state sequence and the original sequence are time-series regularized, the matching difference between the reconstructed motion state sequence and the original sequence is analyzed, and the abnormal motion state data and time nodes are determined through the matching difference to obtain abnormal motion state analysis information.

[0007] In this solution, the fall detection of the target intercom user based on the abnormal motion state analysis information to determine whether there is a fall condition specifically includes: Acquiring abnormal motion state analysis information, performing feature extraction on the abnormal motion state analysis information, extracting abnormal motion state time features, and marking the running state data sequence with the extracted abnormal motion state time features; According to the marked motion state data sequence, a motion state data sequence of a preset time length is intercepted to obtain several motion state data subsequences, and a sliding window method is introduced to perform fall detection; Performing window sliding on the motion state data subsequence according to a preset sliding window size, extracting motion state data features within each window, and obtaining first feature extraction information; Preset a fall determination rule, compare the first feature extraction information with the fall determination rule, and if the feature extracted in the current window meets the fall determination rule, determine the current window as a potential fall event window; Analyze whether the motion state characteristics in the next window of the potential fall event window meet the fall determination rules. If they also meet the rules, mark the corresponding window as a potential fall event window. Perform iterative sliding analysis and fall detection until all motion state data subsequences are detected, output all potential fall event windows and their time positions in the current motion state subsequence, and obtain potential fall event window detection information; Based on the potential fall event window detection information, it is analyzed whether there are continuous potential fall event windows. If there are continuous fall event windows and the number is greater than a preset threshold, the potential fall event window with the earliest time feature is marked as the fall starting point, and fall warning information is generated.

[0008] In this solution, if a fall is detected for the current intercom user, a fall risk assessment is performed in combination with the fall warning information to determine whether collaborative rescue is required, specifically including: Obtaining an operating environment data sequence and fall warning information, extracting a motion state data subsequence when a fall occurs using the fall warning information, and determining a time point at which the intercom user falls based on the fall warning information, and intercepting an operating environment data subsequence after the corresponding time point from the acquired operating environment data sequence; Performing feature extraction on the motion state data subsequence to obtain motion state features of the intercom after the fall occurs, calculating the impact intensity and posture change amplitude based on the extracted motion state features, and generating first analysis information; Construct a three-dimensional coordinate system, perform feature extraction based on the intercepted operating environment data subsequence, obtain the characteristics of the operating environment changes of the walkie-talkie after the fall occurs, and perform fall analysis; Representing the acquired operating environment change characteristics using the constructed three-dimensional coordinate system to generate a three-dimensional coordinate change graph after the walkie-talkie user falls, and analyzing the fall direction, fall distance, and fall location of the walkie-talkie user based on the three-dimensional coordinate change graph to obtain second analysis information; Constructing a risk assessment model, inputting the first analysis information and the second analysis information into the model for analysis, determining the real-time fall risk of the intercom user, and obtaining fall risk assessment information; A risk rescue threshold is preset, and the fall risk assessment information is compared with the risk rescue threshold. If it is greater than the preset threshold, it means that the current intercom user needs to be collaboratively rescued.

[0009] In this solution, if collaborative rescue is required, the walkie-talkies of other rescuers are searched for in the current communication environment for collaborative rescue, and communication optimization is performed in combination with the operating environment data sequence and the communication status data sequence to expand the communication range, specifically including: Obtaining a walkie-talkie operating environment data sequence and a walkie-talkie communication status data sequence, and extracting the walkie-talkie spatial position features before the time node when the target person falls based on the walkie-talkie operating environment data sequence; The extracted spatial position features of the walkie-talkie are converted into three-dimensional coordinates and associated with the corresponding time features; redundancy operations are performed on the associated three-dimensional coordinates and the movement trajectory data of the target walkie-talkie user is generated based on the time change characteristics; Generate collaborative rescue information based on the target walkie-talkie user's movement trajectory data, search for other rescuers' walkie-talkies in the current communication environment for collaborative rescue, and if other rescuers cannot be found in the current environment, optimize communication to expand the autonomous search range; Obtaining specification information of the target walkie-talkie, extracting communication status features of the target walkie-talkie based on the walkie-talkie communication status data sequence, and calculating the current communication frequency of the target walkie-talkie based on the extracted communication status features; The maximum communication frequency feature is extracted from the target walkie-talkie's specifications, and a restricted interval is generated based on the current communication frequency. An improved sparrow algorithm is introduced to optimize the communication frequency, initialize the population, and construct constraint conditions using the restricted interval. Calculate the fitness of the sparrow individuals in the initial population, select a corresponding number of sparrow individuals as discoverers based on the fitness ranking and the proportion of discoverers, and take the remaining sparrow individuals as joiners to update their positions; Randomly select alerts from the sparrows according to the alert ratio, and update their positions. Calculate and update the fitness of each sparrow. In the position update, introduce the golden sine algorithm to guide the sparrows to update to the new position. Update the optimal position of the sparrows through repeated iterations and output the optimized control parameters. Based on the optimized control parameters, a control instruction is generated to control the communication frequency of the target walkie-talkie to expand the communication range.

[0010] A second aspect of the present invention provides a walkie-talkie assisted rescue system based on fall detection in a rescue scenario, the system comprising: a memory and a processor, the memory containing a walkie-talkie assisted rescue method program based on fall detection in a rescue scenario, the walkie-talkie assisted rescue method program based on fall detection in a rescue scenario, when executed by the processor, implements the following steps: Collect rescue status data for real-time rescue missions to obtain rescue status information and preprocess it, time-sequence the preprocessed data, and construct operating environment data sequence, communication status data sequence, and motion status data sequence; An anomaly detection model is constructed based on a generative adversarial network, and the motion state data sequence is input to analyze abnormal motion state data in real-time rescue missions to obtain abnormal motion state analysis information; Performing fall detection on the target intercom user based on the abnormal motion state analysis information, determining whether there is a fall condition, and generating fall warning information; If a fall is detected for the current intercom user, a fall risk assessment is performed based on the fall warning information to determine whether collaborative rescue is required; If collaborative rescue is required, the walkie-talkies of other rescuers will be searched for in the current communication environment for collaborative rescue, and communication optimization will be performed in combination with the operating environment data sequence and the communication status data sequence to expand the communication range.

[0011] The present invention discloses a walkie-talkie-assisted rescue method and system based on fall detection in rescue scenarios, comprising: obtaining and preprocessing rescue status information, and constructing an operating environment data sequence, a communication status data sequence, and a motion status data sequence; constructing an anomaly detection model, inputting the motion status data sequence to analyze abnormal motion status data in real-time rescue tasks; performing fall detection on a target walkie-talkie user to determine whether a fall condition exists; if a fall condition exists, performing a fall risk assessment based on the fall warning information to determine whether collaborative rescue is required; if collaborative rescue is required, searching for the walkie-talkies of other rescuers in the current communication environment for collaborative rescue, and optimizing communication based on the operating environment data sequence and the communication status data sequence to expand the communication range. This avoids safety accidents caused by rescuers being unable to independently call for help in the event of an emergency during the rescue operation, thereby improving the safety and searchability of rescuers when performing rescue tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or exemplary descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to these drawings without paying any creative work.

[0013] Figure 1 A flow chart of a walkie-talkie-assisted rescue method based on fall detection in a rescue scenario provided by one embodiment of the present invention; Figure 2 A flow chart of a method for autonomously requesting assistance by a walkie-talkie provided in one embodiment of the present invention; Figure 3 A block diagram of a walkie-talkie-assisted rescue system based on fall detection in rescue scenarios provided by one embodiment of the present invention; The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0014] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present application and the features therein can be combined with each other.

[0015] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited to the specific embodiments disclosed below.

[0016] Figure 1 A flow chart of a walkie-talkie-assisted rescue method based on fall detection in a rescue scenario provided by one embodiment of the present invention; like Figure 1 As shown, the present invention provides a flow chart of a walkie-talkie assisted rescue method based on fall detection in a rescue scenario, including: S102, collecting rescue status data for real-time rescue tasks to obtain rescue status information and preprocess it, time-sequencing the preprocessed data to construct an operating environment data sequence, a communication status data sequence, and a motion status data sequence; S104, constructing an anomaly detection model based on a generative adversarial network, inputting the motion state data sequence to analyze abnormal motion state data in a real-time rescue mission, and obtaining abnormal motion state analysis information; S106, performing fall detection on the target walkie-talkie user based on the abnormal motion state analysis information, determining whether there is a fall condition, and generating fall warning information; S108, if it is detected that the current intercom user has fallen, a fall risk assessment is performed based on the fall warning information to determine whether coordinated rescue is required; S110, if collaborative rescue is required, then search for the walkie-talkies of other rescuers in the current communication environment for collaborative rescue, and optimize the communication based on the operating environment data sequence and the communication status data sequence to expand the communication range.

[0017] It should be noted that this walkie-talkie-assisted rescue method and system based on fall detection in rescue scenarios is primarily used for abnormal motion state detection, fall risk assessment, and coordinated rescue implementation during real-time rescue missions. It aims to improve rescue efficiency and reliability through data collection and analysis technologies. First, various state data during the rescue process, including the operating environment, communication status, and the user's motion state, are collected to generate multidimensional data for subsequent analysis. Next, three key time series data sequences are constructed: the operating environment data sequence, the communication status data sequence, and the motion state data sequence, laying the foundation for subsequent analysis. During the anomaly detection phase, a generative adversarial network (GAN) model is used to process the motion state data sequence. Through adversarial training of generators and discriminators, GANs can model the latent distribution of input data and efficiently identify anomalies. The motion state of the walkie-talkie user during the rescue mission is analyzed in real time to capture abnormal motion patterns that may be associated with falls and generate abnormal motion state analysis information. Based on this abnormal motion state analysis information, a sliding window method is introduced to detect whether the walkie-talkie user has fallen. Once a fall is confirmed, a risk assessment is conducted based on the fall warning information. A comprehensive analysis of the severity of the fall, the impact of the surrounding environment, and the required rescue time is used to determine whether collaborative rescue mode should be initiated. If collaborative rescue is determined to be necessary, the system proactively searches for other rescuers' walkie-talkie devices within the current communication environment and automatically establishes a communication link. Simultaneously, the system optimizes the communication process by combining the operating environment data sequence and the communication status data sequence to expand the communication range and improve communication stability, thereby increasing the efficiency and effectiveness of collaborative rescue and enhancing the safety of rescuers in the event of a fall.

[0018] Furthermore, in a preferred embodiment of the present invention, the rescue status data collection for real-time rescue tasks is performed to obtain rescue status collection information and perform preprocessing, and the preprocessed data is time-sequenced to construct an operating environment data sequence, a communication status data sequence, and a motion status data sequence, specifically including: The sensor array set based on the target walkie-talkie performs rescue status data collection for real-time rescue tasks to obtain rescue status collection information, wherein the rescue status collection information includes walkie-talkie operating environment data, walkie-talkie communication status data and walkie-talkie motion status data; Performing data preprocessing on the rescue status collected information, calculating the data standard deviation of the rescue status collected information, and comparing the calculated data standard deviation with a preset standard deviation range; The rescue state collected data corresponding to the data standard deviation that is not within the preset standard deviation range is defined as abnormal data, outliers are removed and linear interpolation method is introduced to supplement the difference to obtain preprocessing result information; Time series features are extracted based on the preprocessing result information, and time series processing is performed on the preprocessing result information according to the extracted time series features to generate an operating environment data sequence, a communication status data sequence, and a motion status data sequence.

[0019] It should be noted that the collected information covers three main aspects: the radio's operating environment data, communication status data, and motion status data. This data reflects the physical state of the radio's current environment, the quality of communication with other devices, and the radio's own motion status, providing a foundation for subsequent intelligent analysis. After completing the initial data collection, the rescue status information undergoes a data preprocessing phase to improve data validity and consistency, thereby enhancing the accuracy and robustness of the analysis. After preprocessing, time series features are further extracted. This identifies potential patterns in the data that change over time, captures important information that reflects dynamic changes, and performs time series analysis to generate operating environment data series, communication status data series, and motion status data series. This data reflects the temporal changes of data in different dimensions, providing a more organized and logical infrastructure for subsequent in-depth analysis.

[0020] Furthermore, in a preferred embodiment of the present invention, the anomaly detection model is constructed based on the generative adversarial network, and the motion state data sequence is input to analyze the abnormal motion state data in the real-time rescue mission to obtain abnormal motion state analysis information, which specifically includes: Building an anomaly detection model based on a generative adversarial network, obtaining a motion state data sequence, and inputting the motion state data sequence as an original sequence into the anomaly detection model for analysis; The generator extracts features from the input motion state data sequence, reconstructs the sequence based on the extracted real-time intercom motion state features, generates a reconstructed motion state sequence, and inputs the reconstructed motion state sequence into the discriminator for judgment; The discriminator determines whether the current reconstructed motion state sequence can be constructed. If the discriminator determines that the current reconstructed motion state sequence is unacceptable, the current reconstructed motion state sequence is marked as an incorrect reconstructed motion state sequence, and the next reconstructed motion state sequence is obtained for judgment. The reconstructed motion state sequence generated by the generator is iteratively and repeatedly judged to select a reconstructed motion state sequence that meets a preset acceptance criterion, thereby obtaining a final reconstructed motion state sequence; Calculating a reconstruction error between the final reconstructed motion state sequence and the original sequence, and comparing the calculated reconstruction error with a preset threshold. If the calculated reconstruction error is greater than the preset threshold, it indicates that the current real-time motion state is abnormal. The final reconstructed motion state sequence and the original sequence are time-series regularized, the matching difference between the reconstructed motion state sequence and the original sequence is analyzed, and the abnormal motion state data and time nodes are determined through the matching difference to obtain abnormal motion state analysis information.

[0021] It should be noted that the anomaly detection model is constructed using a generative adversarial network (GAN). By analyzing motion state data sequences, it accurately identifies abnormal motion states. First, the collected motion state data sequence is used as the original sequence and input into the anomaly detection model for processing. During the GAN operation, the generator reconstructs the sequence based on the motion state features of the real-time walkie-talkie, generating a reconstructed motion state sequence. This sequence is then fed into the discriminator for judgment and evaluation of the reconstruction results. During the discrimination phase, the discriminator analyzes the generator's reconstructed sequence and determines whether it meets structural expectations. If the discriminator evaluates the current reconstructed sequence as "unacceptable," indicating that the generator's output does not meet the established criteria, the reconstructed sequence is marked as an incorrectly reconstructed motion state sequence and enters the next generation-discrimination cycle until a reconstructed motion state sequence that meets the pre-set acceptance criteria is selected. Through adversarial training between the generator and the discriminator, the accuracy and reliability of the reconstructed sequence are continuously improved. After obtaining the final reconstructed motion state sequence, the reconstruction error between the reconstructed sequence and the original sequence is calculated to assess the generator's degree of restoration of the original data features. If the reconstruction error exceeds a preset threshold, it indicates an anomaly in the current motion state. The final reconstructed motion state sequence is time-aligned with the original sequence, and the matching discrepancies between the reconstructed and original sequences are analyzed. The abnormal data points and their corresponding time points are then located, enabling precise analysis of the abnormal motion state. Ultimately, the abnormal motion state analysis information is output, providing crucial data support for subsequent fall detection and rescue decisions.

[0022] Furthermore, in a preferred embodiment of the present invention, the step of performing fall detection on the target intercom user based on the abnormal motion state analysis information to determine whether a fall condition exists specifically includes: Acquiring abnormal motion state analysis information, performing feature extraction on the abnormal motion state analysis information, extracting abnormal motion state time features, and marking the running state data sequence with the extracted abnormal motion state time features; According to the marked motion state data sequence, a motion state data sequence of a preset time length is intercepted to obtain several motion state data subsequences, and a sliding window method is introduced to perform fall detection; Performing window sliding on the motion state data subsequence according to a preset sliding window size, extracting motion state data features within each window, and obtaining first feature extraction information; Preset a fall determination rule, compare the first feature extraction information with the fall determination rule, and if the feature extracted in the current window meets the fall determination rule, determine the current window as a potential fall event window; Analyze whether the motion state characteristics in the next window of the potential fall event window meet the fall determination rules. If they also meet the rules, mark the corresponding window as a potential fall event window. Perform iterative sliding analysis and fall detection until all motion state data subsequences are detected, output all potential fall event windows and their time positions in the current motion state subsequence, and obtain potential fall event window detection information; Based on the potential fall event window detection information, it is analyzed whether there are continuous potential fall event windows. If there are continuous fall event windows and the number is greater than a preset threshold, the potential fall event window with the earliest time feature is marked as the fall starting point, and fall warning information is generated.

[0023] It should be noted that, first, key features are extracted from the abnormal motion state analysis information, particularly the temporal features of the abnormal motion state, to identify the time points associated with the abnormality. Based on the labeled motion state data sequence, the sequence is divided into several motion state data subsequences of preset time lengths. A sliding window method is introduced, gradually sliding a window of a preset sliding window size over each motion state data subsequence, analyzing the features of the motion state data extracted from each window. The extracted features include various motion parameters such as velocity, acceleration, and posture change, comprehensively reflecting the motion state within that time period. Subsequently, the features extracted from each window are evaluated using a preset fall detection rule. If the features within a window meet the fall detection rule, the window is marked as a potential fall event window, indicating that a fall may have occurred during that time period. For each potential fall event window, the motion state features within subsequent windows are further analyzed. If the features of subsequent windows also meet the fall detection rule, the window is also marked as a potential fall event window. This is expanded across the entire subsequence using an iterative sliding method until all windows are detected. All potential fall event windows in the current subsequence and their corresponding time locations are output to form potential fall event window detection information. Finally, based on the potential fall event window detection information, the system analyzes whether there are consecutive potential fall event windows. If the number of consecutive windows exceeds a preset threshold, the anomaly is likely a fall event. The potential fall event window with the earliest temporal signature is marked as the fall starting point, and a fall warning is generated.

[0024] It should be noted that the setting of the fall judgment rules includes acceleration peak value, posture angle change, acceleration change rate, etc. By setting the corresponding thresholds, it is possible to determine whether a fall occurs.

[0025] Furthermore, in a preferred embodiment of the present invention, if a fall is detected in the current intercom user, a fall risk assessment is performed in combination with the fall warning information to determine whether collaborative rescue is required, specifically including: Obtaining an operating environment data sequence and fall warning information, extracting a motion state data subsequence when a fall occurs using the fall warning information, and determining a time point at which the intercom user falls based on the fall warning information, and intercepting an operating environment data subsequence after the corresponding time point from the acquired operating environment data sequence; Performing feature extraction on the motion state data subsequence to obtain motion state features of the intercom after the fall occurs, calculating the impact intensity and posture change amplitude based on the extracted motion state features, and generating first analysis information; Construct a three-dimensional coordinate system, perform feature extraction based on the intercepted operating environment data subsequence, obtain the characteristics of the operating environment changes of the walkie-talkie after the fall occurs, and perform fall analysis; Representing the acquired operating environment change characteristics using the constructed three-dimensional coordinate system to generate a three-dimensional coordinate change graph after the walkie-talkie user falls, and analyzing the fall direction, fall distance, and fall location of the walkie-talkie user based on the three-dimensional coordinate change graph to obtain second analysis information; Constructing a risk assessment model, inputting the first analysis information and the second analysis information into the model for analysis, determining the real-time fall risk of the intercom user, and obtaining fall risk assessment information; A risk rescue threshold is preset, and the fall risk assessment information is compared with the risk rescue threshold. If it is greater than the preset threshold, it means that the current intercom user needs to be collaboratively rescued.

[0026] It should be noted that the fall warning information is first used to locate the time point at which the walkie-talkie user fell, and the corresponding motion state data subsequences are extracted from it. These subsequences collectively display the detailed motion characteristics of the user during the fall. Simultaneously, a subsequence of environmental data corresponding to the time point of the fall is extracted from the acquired operating environment data sequence. This data reflects the dynamic changes in the user's environment after the fall. Feature extraction is performed on the extracted motion state data subsequences. Based on the extracted features, the impact intensity and posture change amplitude during the user's fall are calculated to generate first analysis information related to the motion state. These data are important parameters for measuring the severity of the fall. The impact intensity can reflect the force exerted by the user during the fall, while the posture change amplitude provides information about the posture changes during the fall. Furthermore, feature extraction is performed on the operating environment data subsequences to construct a three-dimensional coordinate system to characterize the changes in the operating environment after the user falls. By visualizing these environmental change characteristics in the three-dimensional coordinate system, a three-dimensional coordinate change map of the walkie-talkie user's fall condition is generated. The three-dimensional change map provides spatial information, including the direction, distance, and specific location of the fall. These indicators are used to generate second analysis information related to environmental changes. The first and second analysis information are input into the risk assessment model for calculation. The user's real-time fall risk is accurately assessed to obtain fall risk assessment information. This is then compared with a preset risk rescue threshold. If the risk assessment result exceeds the threshold, it indicates that the user's situation requires attention and may require timely coordinated rescue. This ensures that the risk level of a fall incident can be quickly and accurately determined in real-world rescue scenarios, providing a foundation for ensuring the safety of rescuers in a timely manner.

[0027] Furthermore, in a preferred embodiment of the present invention, if collaborative rescue is required, searching for the walkie-talkies of other rescuers in the current communication environment for collaborative rescue, and optimizing communication to expand the communication range by combining the operating environment data sequence and the communication status data sequence, specifically including: Obtaining a walkie-talkie operating environment data sequence and a walkie-talkie communication status data sequence, and extracting the walkie-talkie spatial position features before the time node when the target person falls based on the walkie-talkie operating environment data sequence; The extracted spatial position features of the walkie-talkie are converted into three-dimensional coordinates and associated with the corresponding time features; redundancy operations are performed on the associated three-dimensional coordinates and the movement trajectory data of the target walkie-talkie user is generated based on the time change characteristics; Generate collaborative rescue information based on the target walkie-talkie user's movement trajectory data, search for other rescuers' walkie-talkies in the current communication environment for collaborative rescue, and if other rescuers cannot be found in the current environment, optimize communication to expand the autonomous search range; Obtaining specification information of the target walkie-talkie, extracting communication status features of the target walkie-talkie based on the walkie-talkie communication status data sequence, and calculating the current communication frequency of the target walkie-talkie based on the extracted communication status features; The maximum communication frequency feature is extracted from the target walkie-talkie's specifications, and a restricted interval is generated based on the current communication frequency. An improved sparrow algorithm is introduced to optimize the communication frequency, initialize the population, and construct constraint conditions using the restricted interval. Calculate the fitness of the sparrow individuals in the initial population, select a corresponding number of sparrow individuals as discoverers based on the fitness ranking and the proportion of discoverers, and take the remaining sparrow individuals as joiners to update their positions; Randomly select alerts from the sparrows according to the alert ratio, and update their positions. Calculate and update the fitness of each sparrow. In the position update, introduce the golden sine algorithm to guide the sparrows to update to the new position. Update the optimal position of the sparrows through repeated iterations and output the optimized control parameters. Based on the optimized control parameters, a control instruction is generated to control the communication frequency of the target walkie-talkie to expand the communication range.

[0028] It should be noted that the spatial location characteristics of the target walkie-talkie before the target person fell are extracted based on the target walkie-talkie's operating environment data sequence. This location characteristic characterizes the dynamic behavior of the walkie-talkie in space. Subsequently, the spatial location characteristics are converted into three-dimensional coordinates and correlated with time characteristics to form three-dimensional coordinate data with a time dimension. Redundancy is removed from the correlated three-dimensional coordinates, and based on the time-varying characteristics, movement trajectory data of the target walkie-talkie user is generated. This trajectory data clearly demonstrates the walkie-talkie's movement path, providing important reference for rescue operations. Within the current communication environment, the system attempts to search for other rescuers' walkie-talkies to achieve coordinated rescue. If other rescue equipment cannot be successfully found within the current communication environment, communication optimization is automatically performed to expand the search range to ensure the effectiveness of the coordinated rescue operation. Furthermore, the communication status characteristics of the target walkie-talkie are extracted using the walkie-talkie's specifications and communication status data sequence, and its current communication frequency is calculated based on this information. Furthermore, the maximum communication frequency characteristics are extracted based on the specifications, and a limit range is defined based on the current communication frequency. An improved sparrow algorithm optimizes communication frequency. Constraints are established within a restricted range during population initialization to ensure that the optimization process remains within a feasible range. During the optimization phase, the algorithm calculates the fitness of individual sparrows in the initial population and selects the best-performing individuals as discoverers based on their fitness ranking. The remaining individuals are classified as joiners. Discoverers are responsible for exploring the global optimal solution, while joiners conduct a more localized search. Furthermore, a specific proportion of individuals are selected as vigilants. These vigilants utilize a special update mechanism to enhance the algorithm's global search capabilities. To improve convergence and accuracy, the golden sine algorithm guides the migration of individual sparrows during position updates. Through iterative optimization, the algorithm continuously searches for the optimal position, ultimately outputting optimized control parameters and generating control commands. This algorithm then adjusts the communication frequency of the target walkie-talkie in real time, effectively expanding the communication range. This not only improves the safety of rescuers who have fallen, but also enhances communication quality in complex environments, ensuring the success of rescue missions.

[0029] It should be noted that the location update calculation process is: ; in, Indicates the The sparrow individual is in sequence The first iteration The position on the dimension, Represents a random number that determines the moving distance and direction of the individual sparrow. Indicates the The global optimal position of the iteration is Indicates the position update parameter combined with the golden section coefficient, where the golden section coefficient is By combining the position update parameters with the golden section coefficient, the search space is narrowed and the individual sparrows are guided to move to the optimal position.

[0030] Figure 2 A flow chart of a method for autonomously requesting assistance by a walkie-talkie provided in one embodiment of the present invention; like Figure 2 As shown, the present invention provides a flow chart of a method for autonomously requesting assistance by a walkie-talkie, including: S202, obtaining a walkie-talkie operating environment data sequence and a walkie-talkie communication status data sequence, and extracting the walkie-talkie spatial position features before the time node when the target person falls based on the walkie-talkie operating environment data sequence; S204, converting the extracted walkie-talkie spatial position features into three-dimensional coordinates and associating them with corresponding time features, performing a redundancy operation on the associated three-dimensional coordinates and generating movement trajectory data of the target walkie-talkie user based on the time change features; S206, generating collaborative rescue information based on the target walkie-talkie user's movement trajectory data, searching for other rescuers' walkie-talkies in the current communication environment for collaborative rescue, and if no other rescuers are found in the current environment, performing communication optimization to expand the autonomous search range; S208, extracting the communication status characteristics of the target intercom through the intercom communication status data sequence, calculating the current communication frequency of the target intercom, and generating a restriction interval based on the maximum communication frequency characteristics extracted from the target intercom; S210, introduce the improved sparrow algorithm to optimize the communication frequency, update the optimal position of the sparrow through repeated iterations, output the optimized control parameters to generate a control instruction to control the communication frequency of the target walkie-talkie to expand the communication range.

[0031] Furthermore, the walkie-talkie assisted rescue method based on fall detection in a rescue scenario provided by the present invention further includes the following steps: Based on data retrieval, several historical collaborative rescue instances of historical fall scenarios are obtained, and features of the historical collaborative rescue instances are extracted to obtain communication environment features and communication optimization scheme features of each historical collaborative rescue instance. Based on the communication environment characteristics of each historical collaborative rescue instance, a communication environment portrait of the corresponding historical rescue instance is constructed as a representation of the corresponding historical rescue instance, and the similarity value between the communication environment portraits of each historical rescue instance is calculated; The similarity values ​​calculated are used to classify each historical collaborative rescue instance using a clustering algorithm, and the extracted communication optimization scheme features are associated with the corresponding categories to build a historical collaborative rescue database. When the fall detection determines that the current intercom user needs to be rescued collaboratively, the rescue status information collected at the current moment is matched and analyzed with the communication environment portrait stored in the historical collaborative database; If there is a matching communication environment profile, the communication optimization scheme associated with the target communication environment profile corresponding to the collaborative rescue category is extracted. Based on the extracted communication optimization scheme, corresponding control instructions are generated to optimize communication and expand the communication rescue range. If there is no matching communication environment profile, an optimization plan is formulated based on the current communication frequency and the maximum communication frequency of the walkie-talkie.

[0032] It should be noted that in the event of a rescuer falling, in more serious situations, such as when the rescuer has fallen from a height and lost the ability to rescue themselves, it is necessary to find someone who can assist in the rescue in a short time to ensure the safety of the rescuer. By analyzing historical data and adopting communication optimization solutions in similar environments, the walkie-talkie's ability to call for help is further improved, ensuring the safety of fallen rescuers and improving searchability, thereby avoiding safety issues for rescuers caused by carrying out rescue missions in harsh environments.

[0033] Figure 3 An embodiment of the present invention provides a walkie-talkie assisted rescue system 3 based on fall detection in a rescue scenario. The system includes: a memory 31 and a processor 32. The memory 31 contains a walkie-talkie assisted rescue method program based on fall detection in a rescue scenario. When the walkie-talkie assisted rescue method program based on fall detection in a rescue scenario is executed by the processor 32, the following steps are implemented: Collect rescue status data for real-time rescue missions to obtain rescue status information and preprocess it, time-sequence the preprocessed data, and construct operating environment data sequence, communication status data sequence, and motion status data sequence; An anomaly detection model is constructed based on a generative adversarial network, and the motion state data sequence is input to analyze abnormal motion state data in real-time rescue missions to obtain abnormal motion state analysis information; Performing fall detection on the target intercom user based on the abnormal motion state analysis information to determine whether there is a fall condition; If a fall is detected for the current intercom user, a fall risk assessment is performed based on the fall warning information to determine whether collaborative rescue is required; If collaborative rescue is required, the walkie-talkies of other rescuers will be searched for in the current communication environment for collaborative rescue, and communication optimization will be performed in combination with the operating environment data sequence and the communication status data sequence to expand the communication range.

[0034] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0035] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed across multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0036] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0037] Those skilled in the art will appreciate that all or part of the steps of the above-mentioned method embodiments may be implemented by hardware associated with program instructions, and the aforementioned program may be stored in a computer-readable storage medium. When the program is executed, the program executes the steps of the above-mentioned method embodiments. The aforementioned storage medium includes various media that can store program codes, such as mobile storage devices, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0038] Alternatively, if the integrated units described above are implemented as software modules and sold or used as standalone products, they can also be stored on a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present invention, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product, stored on a storage medium, includes instructions for enabling a computer device (such as a personal computer, server, or network device) to execute all or part of the methods described in various embodiments of the present invention. The aforementioned storage media include various media capable of storing program code, such as removable storage devices, ROM, RAM, magnetic disks, or optical disks.

[0039] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A walkie-talkie assisted rescue method based on fall detection in rescue scenarios, characterized in that: include: Collect rescue status data for real-time rescue tasks to obtain rescue status collection information and preprocess it, time-sequence the preprocessed data, and construct operating environment data sequence, communication status data sequence and motion status data sequence; Building an anomaly detection model based on a generative adversarial network, inputting the motion state data sequence to analyze abnormal motion state data in real-time rescue missions, and obtaining abnormal motion state analysis information; Based on the abnormal motion state analysis information, a fall detection is performed on the target intercom user to determine whether there is a fall condition, and a fall warning information is generated; If it is detected that the current intercom user has fallen, a fall risk assessment is performed in combination with the fall warning information to determine whether collaborative rescue is required; If collaborative rescue is required, the intercoms of other rescuers are searched for in the current communication environment for collaborative rescue, and communication optimization is performed in combination with the operating environment data sequence and the communication status data sequence to expand the communication range.

2. According to claim 1, a walkie-talkie assisted rescue method based on fall detection in rescue scenarios is characterized in that: The rescue status data collection for the real-time rescue task obtains the rescue status collection information and performs preprocessing, performs time series processing on the preprocessed data, and constructs the operation environment data sequence, the communication status data sequence and the motion status data sequence, specifically including: Based on the sensor array set by the target walkie-talkie, rescue status data collection is performed on the real-time rescue task to obtain rescue status collection information, wherein the rescue status collection information includes walkie-talkie operating environment data, walkie-talkie communication status data and walkie-talkie motion status data; Performing data preprocessing on the rescue status collected information, calculating the data standard deviation of the rescue status collected information, and comparing the calculated data standard deviation with a preset standard deviation range; Define the rescue state collected data corresponding to the data standard deviation that is not within the preset standard deviation range as abnormal data, remove the abnormal values ​​and introduce the linear interpolation method to supplement the difference to obtain the preprocessing result information; Based on the preprocessing result information, time series features are extracted, and time series processing is performed on the preprocessing result information according to the extracted time series features to generate an operating environment data sequence, a communication status data sequence, and a motion status data sequence.

3. According to claim 1, a walkie-talkie assisted rescue method based on fall detection in rescue scenarios is characterized in that: The abnormality detection model is constructed based on the generative adversarial network, and the motion state data sequence is input to analyze the abnormal motion state data in the real-time rescue mission to obtain abnormal motion state analysis information, which specifically includes: Building an anomaly detection model based on a generative adversarial network, obtaining a motion state data sequence, and inputting the motion state data sequence as an original sequence into the anomaly detection model for analysis; The generator extracts features from the input motion state data sequence, reconstructs the sequence according to the extracted real-time intercom motion state features, generates a reconstructed motion state sequence, and inputs the reconstructed motion state sequence into the discriminator for judgment; The discriminator determines whether the current reconstructed motion state sequence can be constructed. If the discriminator determines that the current reconstructed motion state sequence is unacceptable, the current reconstructed motion state sequence is marked as an erroneous reconstructed motion state sequence, and the next reconstructed motion state sequence is obtained for judgment; The reconstructed motion state sequence generated by the iteratively repeated discriminant generator is selected to meet the preset acceptance criteria to obtain a final reconstructed motion state sequence; Calculating the reconstruction error between the final reconstructed motion state sequence and the original sequence, and comparing the calculated reconstruction error with a preset threshold. If the error is greater than the preset threshold, it indicates that the current real-time motion state is abnormal. The final reconstructed motion state sequence and the original sequence are time-ordered, the matching difference between the reconstructed motion state sequence and the original sequence is analyzed, and the abnormal motion state data and time nodes are determined through the matching difference to obtain abnormal motion state analysis information.

4. The walkie-talkie assisted rescue method based on fall detection in rescue scenarios according to claim 1, characterized in that: The step of performing fall detection on the target intercom user based on the abnormal motion state analysis information to determine whether there is a fall condition specifically includes: Acquire abnormal motion state analysis information, perform feature extraction on the abnormal motion state analysis information, extract abnormal motion state time features, and mark the running state data sequence with the extracted abnormal motion state time features; According to the marked motion state data sequence, a motion state data sequence of a preset time length is intercepted to obtain several motion state data subsequences, and a sliding window method is introduced to perform fall detection; Performing window sliding on the motion state data subsequence according to a preset sliding window size, extracting motion state data features in each window, and obtaining first feature extraction information; Preset a fall determination rule, and judge the first feature extraction information against the fall determination rule. If the feature extracted in the current window meets the fall determination rule, the current window is judged as a potential fall event window; Analyze whether the motion state characteristics in the next window of the potential fall event window meet the fall determination rule. If they also meet the rule, mark the corresponding window as a potential fall event window. Perform iterative sliding analysis and fall determination until all motion state data subsequences are detected, output all potential fall event windows and their time positions of the current motion state subsequence, and obtain potential fall event window detection information; Whether there are continuous potential fall event windows is analyzed according to the potential fall event window detection information. If there are continuous fall event windows and the number is greater than a preset threshold, the potential fall event window with the earliest time feature is marked as the fall starting point, and fall warning information is generated.

5. The walkie-talkie assisted rescue method based on fall detection in rescue scenarios according to claim 1, characterized in that: If it is detected that the current intercom user has fallen, a fall risk assessment is performed in combination with the fall warning information to determine whether collaborative rescue is required, specifically including: Acquire an operating environment data sequence and fall warning information, use the fall warning information to extract a motion state data subsequence when a fall occurs, and determine the time node when the intercom user falls according to the fall warning information, and intercept the operating environment data subsequence after the corresponding time node from the acquired operating environment data sequence; Extracting features from the motion state data subsequence, obtaining motion state features of the intercom after the fall occurs, calculating the impact intensity and posture change amplitude based on the extracted motion state features, and generating first analysis information; Construct a three-dimensional coordinate system, perform feature extraction based on the intercepted operating environment data subsequence, obtain the operating environment change characteristics of the intercom after the fall occurs, and perform fall analysis; The acquired operating environment change characteristics are represented by using the constructed three-dimensional coordinate system, and a three-dimensional coordinate change graph after the walkie-talkie user falls is generated; and the fall direction, fall distance and fall position of the walkie-talkie user are analyzed based on the three-dimensional coordinate change graph to obtain second analysis information; Constructing a risk assessment model, inputting the first analysis information and the second analysis information into the model for analysis, determining the real-time fall risk of the intercom user, and obtaining fall risk assessment information; A risk rescue threshold is preset, and the fall risk assessment information is compared with the risk rescue threshold. If it is greater than the preset threshold, it means that the current intercom user needs collaborative rescue.

6. The walkie-talkie assisted rescue method based on fall detection in rescue scenarios according to claim 1, characterized in that: If collaborative rescue is required, the intercoms of other rescuers are searched for in the current communication environment for collaborative rescue, and communication optimization is performed in combination with the operating environment data sequence and the communication status data sequence to expand the communication range, specifically including: Acquire a walkie-talkie operating environment data sequence and a walkie-talkie communication status data sequence, and extract the walkie-talkie spatial position features before the time node when the target person falls based on the walkie-talkie operating environment data sequence; The extracted spatial position features of the walkie-talkie are converted into three-dimensional coordinates and associated with the corresponding time features; the associated three-dimensional coordinates are de-redundant and the movement trajectory data of the target walkie-talkie user is generated according to the time change characteristics; Generate collaborative rescue information based on the target walkie-talkie user's movement trajectory data, search for other rescuers' walkie-talkies in the current communication environment for collaborative rescue, and if other rescuers cannot be found in the current environment, optimize communication to expand the autonomous search range; Acquire specification information of the target intercom, extract communication status features of the target intercom based on the intercom communication status data sequence, and calculate the current communication frequency of the target intercom according to the extracted communication status features; The maximum communication frequency feature is extracted through the specification information of the target walkie-talkie, and the restricted interval is generated in combination with the current communication frequency. The improved sparrow algorithm is introduced to optimize the communication frequency, the population is initialized, and the constraint conditions are constructed through the restricted interval. Calculate the fitness of the sparrow individuals in the initial population, select the corresponding number of sparrow individuals as discoverers based on the fitness ranking and the proportion of discoverers, and take the remaining sparrow individuals as joiners to update their positions; Randomly select alerters from individual sparrows according to the alerter ratio, and update the position. Calculate and update the fitness of each individual sparrow. In the position update, introduce the golden sine algorithm to guide the individual sparrow to update to the new position. Update the optimal position of the sparrow through repeated iterations and output the optimized control parameters. Based on the optimized control parameters, a control instruction is generated to control the communication frequency of the target walkie-talkie to expand the communication range.

7. A walkie-talkie assisted rescue system based on fall detection in rescue scenarios, characterized in that: The system includes: a memory and a processor, wherein the memory contains a walkie-talkie assisted rescue method program based on rescue scene fall detection, and when the walkie-talkie assisted rescue method program based on rescue scene fall detection is executed by the processor, the following steps are implemented: Collect rescue status data for real-time rescue tasks to obtain rescue status collection information and preprocess it, time-sequence the preprocessed data, and construct operating environment data sequence, communication status data sequence and motion status data sequence; Building an anomaly detection model based on a generative adversarial network, inputting the motion state data sequence to analyze abnormal motion state data in real-time rescue missions, and obtaining abnormal motion state analysis information; Based on the abnormal motion state analysis information, a fall detection is performed on the target intercom user to determine whether there is a fall condition, and a fall warning information is generated; If it is detected that the current intercom user has fallen, a fall risk assessment is performed in combination with the fall warning information to determine whether collaborative rescue is required; If collaborative rescue is required, the intercoms of other rescuers are searched for in the current communication environment for collaborative rescue, and communication optimization is performed in combination with the operating environment data sequence and the communication status data sequence to expand the communication range.

Images