A rescue method and system of a rescue robot for exploration

Abstract

The application discloses a rescue method and system of a rescue robot for exploration, and relates to the technical field of underground space rescue, and comprises the following steps: acquiring multi-path acoustic echo data and robot motion trajectory data of a karst cave, and constructing a three-dimensional point cloud model based on the data; identifying unmatched data in the multi-path acoustic echo data and the robot motion trajectory data, and taking a region where the unmatched data is located as an abnormal region. The application enhances the spatial resolution of signal collection through a multi-microphone array, extracts robust acoustic fingerprints by combining a noise reduction algorithm and a short-time Fourier transform, and screens out human body sound source signals with high confidence by using a feature template matching mechanism, so that accurate extraction and identification of human body acoustic characteristics in a complex noise environment of a karst cave are realized, and the misjudgment problem caused by confusion of sound source characteristics and environmental noise in a traditional method is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of underground space rescue technology, specifically to a rescue method and system for an exploration rescue robot. Background Technology

[0002] With the continuous development of underground space and geological exploration activities, the need for rescue of missing persons in complex environments such as underground caves and mines is becoming increasingly prominent. Traditional manual rescue methods often suffer from many problems such as slow response speed, high personnel safety risks, and low positioning accuracy due to complex spatial environments, low visibility, and unclear geological structures. In recent years, with the development of technologies such as robotics, environmental perception, acoustic detection, and 3D modeling, research on the search and rescue of missing persons based on rescue robots has become a hot topic in related fields.

[0003] However, in the process of implementing the technical solution of the invention in the embodiments of this application, it was found that the above-mentioned technology has at least the following technical problems:

[0004] Existing methods often fail to effectively combine 3D environmental models with acoustic features, resulting in inaccurate estimations of sound source locations. They also lack correlation analysis between sound source locations and abnormal environmental areas, leading to insufficient safety in generating rescue paths and hindering rescue robots from making optimal decisions. This restricts the efficiency and success rate of rescue operations in karst environments. Summary of the Invention

[0005] The purpose of this invention is to provide a rescue method and system for an exploration rescue robot to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows:

[0007] In a first aspect, this invention discloses a rescue method for an exploration rescue robot, applied to the rescue of people who have lost contact during cave exploration, comprising the following steps:

[0008] Acquire multipath acoustic echo data and robot motion trajectory data from the cave, and use them to construct a three-dimensional point cloud model;

[0009] Identify mismatched data in the multipath acoustic echo data and robot motion trajectory data, and use the area where the mismatched data is located as an abnormal area.

[0010] Acquire audio signals from the cave and extract acoustic fingerprint features from the audio signals;

[0011] The acoustic fingerprint features are compared with a preset feature template to calculate similarity, and the target audio signal is determined based on the similarity calculation results.

[0012] The audio emission coordinates are calculated based on the target audio signal. It is determined whether the audio emission coordinates are located in an abnormal area. If so, the audio emission coordinates are corrected based on the multipath acoustic echo data of the abnormal area. The degree of overlap between the corrected audio emission coordinates and the abnormal area is calculated.

[0013] If the overlap is greater than a preset threshold, the corrected audio coordinates are determined as the coordinates of the potential missing person, and the spatial confidence score is obtained by nonlinear calculation of the overlap and similarity results.

[0014] If the spatial confidence level is greater than a preset value, then the corrected audio coordinates are determined to be the coordinates of the missing person.

[0015] A rescue route for the missing persons is generated based on their coordinates and the robot's trajectory data.

[0016] Secondly, the present invention discloses a rescue system for an exploration rescue robot, comprising:

[0017] The 3D point cloud model building module is used to acquire multipath acoustic echo data and robot motion trajectory data of the cave, and to build a 3D point cloud model based on these data.

[0018] An abnormal region identification module is used to identify mismatched data in the multipath acoustic echo data and robot motion trajectory data, and to use the area where the mismatched data is located as an abnormal region.

[0019] The audio signal acquisition module is used to acquire audio signals from the cave.

[0020] An audio signal processing module is used to extract acoustic fingerprint features from the audio signal, perform similarity calculation between the acoustic fingerprint features and a preset feature template, and determine the target audio signal based on the similarity calculation result.

[0021] The audio emission coordinate determination module is used to calculate the audio emission coordinates based on the target audio signal, determine whether the audio emission coordinates are located in an abnormal area, and if so, correct the audio emission coordinates based on the multipath acoustic echo data of the abnormal area.

[0022] The missing persons coordinate determination module is used to calculate the overlap between the corrected audio emission coordinates and the abnormal area;

[0023] If the overlap is greater than a preset threshold, the corrected audio coordinates are determined as the coordinates of the potential missing person, and the spatial confidence score is obtained by nonlinear calculation of the overlap and similarity results.

[0024] If the spatial confidence level is greater than a preset value, then the corrected audio coordinates are determined to be the coordinates of the missing person.

[0025] The rescue path generation module is used to generate a rescue path for the missing person based on the coordinates of the missing person and the robot's motion trajectory data.

[0026] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0027] 1. This application enhances the spatial resolution of signal acquisition by using a multi-microphone array, combines noise reduction algorithms with short-time Fourier transform to extract robust acoustic fingerprints, and uses a feature template matching mechanism to screen out high-confidence human sound source signals. This achieves accurate extraction and identification of human acoustic features in the complex noise environment of caves, effectively solving the problem of misjudgment caused by confusion between sound source features and environmental noise in traditional methods.

[0028] 2. This application innovatively combines system error compensation based on environmental structure modeling with dynamic coordinate correction based on signal feature learning. Through the dual correction mechanism, it effectively overcomes the limitations of a single correction method under complex geological conditions, effectively solves the problem of decreased positioning accuracy caused by multipath propagation of sound waves in the abnormal area of ​​the karst cave, and provides a reliable spatial coordinate data foundation for rescue route planning.

[0029] 3. This solution establishes a real-time correlation mechanism between path status and environmental changes through dynamic signal monitoring and three-dimensional spatial mapping, realizing dynamic safety control of rescue paths in complex karst cave environments and effectively avoiding robot jamming or crash accidents caused by sudden changes in geological structure. Attached Figure Description

[0030] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts. Wherein:

[0031] Figure 1 This is a flowchart of the steps of the present invention;

[0032] Figure 2 A schematic diagram illustrating the process of determining the coordinates of audio emission provided by the present invention;

[0033] Figure 3 A schematic diagram of the process for correcting audio emission coordinates provided by the present invention;

[0034] Figure 4 A flowchart illustrating the process for determining the priority of rescuing missing persons, provided by this invention;

[0035] Figure 5 This is a schematic diagram of a scenario provided by the present invention;

[0036] Figure 6 A functional diagram of the system modules provided by the present invention. Detailed Implementation

[0037] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0038] Application Overview:

[0039] In existing technologies, with the continuous development of underground space and geological exploration activities, the rescue of missing persons in complex environments faces severe challenges. Traditional rescue methods rely on manual operation, which has problems of response delay and positioning deviation in low visibility environments such as caves. Existing technologies mostly use single sensor data, and sound source positioning is easily affected by multipath effects. Furthermore, the positioning results are not correlated with environmental structural risks, resulting in safety hazards in rescue route planning.

[0040] To address the aforementioned issues, and considering the distortion of sound wave propagation in complex geological structures, it was found that relying solely on acoustic signals for localization leads to coordinate shifts. Further research into the structural characteristics of karst caves revealed that 3D environmental modeling can provide spatial constraints. Through multiple experimental verifications, it was found that fusing multipath acoustic echo data with motion trajectory data for modeling can effectively restore the true spatial structure. When the sound source coordinates are detected to be located in an abnormal area, it was inferred that conventional localization algorithms have systematic errors, leading to the proposal of an adaptive correction mechanism based on environmental characteristics.

[0041] After introducing the basic concept of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0042] Example 1:

[0043] Please see Figures 1-5 A rescue method using an exploration rescue robot, applied to the rescue of people who have lost contact while exploring caves, includes the following steps:

[0044] Acquire multipath acoustic echo data and robot motion trajectory data from the cave, and use them to construct a three-dimensional point cloud model;

[0045] Identify mismatched data in multipath acoustic echo data and robot motion trajectory data, and use the areas where the mismatched data is located as abnormal areas.

[0046] Acquire audio signals from the cave and extract acoustic fingerprint features from the audio signals;

[0047] The similarity between the acoustic fingerprint features and the preset feature template is calculated, and the target audio signal is determined based on the similarity calculation results.

[0048] The audio emission coordinates are calculated based on the target audio signal. It is determined whether the audio emission coordinates are located in the abnormal area. If so, the audio emission coordinates are corrected based on the multipath acoustic echo data of the abnormal area, and the overlap between the corrected audio emission coordinates and the abnormal area is calculated.

[0049] If the overlap is greater than a preset threshold, the corrected audio coordinates are determined as the coordinates of the potential missing person. The spatial confidence score is then obtained through non-linear calculation of the overlap and similarity results. The specific calculation formula is as follows:

[0050]

[0051] In the formula, Indicates spatial confidence. Indicates the degree of overlap. Indicates similarity. Indicates the adaptive weighting coefficient. Represents the nonlinear enhancement coefficient. Represents the base of the natural logarithm;

[0052] If the spatial confidence level is greater than the preset value, then the corrected audio coordinates are determined to be the coordinates of the missing person.

[0053] A rescue route for the missing persons is generated based on their coordinates and the robot's movement trajectory data.

[0054] Among them, the three-dimensional point cloud model refers to the set of spatial points generated by the propagation delay of sound waves and the spatial coordinate data of the robot. For example, it can be realized by using the ToF ranging algorithm combined with SLAM technology to accurately characterize the internal structure of the cave.

[0055] Anomaly regions refer to spatial areas with abnormal point cloud distributions, such as those identified by density mutation detection algorithms, which can be used to provide early warning of geological structural risks.

[0056] Acoustic fingerprint features refer to the spectral feature vector of an audio signal, which can be extracted, for example, through MFCC coefficients, and is used to distinguish human sound sources from environmental noise.

[0057] Overlap refers to the spatial overlap ratio between the corrected coordinates and the abnormal region. For example, it can be obtained by calculating through voxelized mesh and is used to evaluate the reliability of the coordinates.

[0058] Spatial confidence is a composite index that combines acoustic matching degree and spatial overlap degree. For example, it can be calculated using a weighted product method and used for final positioning decision.

[0059] Specifically, the system first collects multipath echo data through the sonar equipment on the robot, combines it with IMU sensors to record motion trajectory data, and uses point cloud reconstruction algorithms to generate a three-dimensional environment model.

[0060] When the microphone array captures suspicious audio, its Mel frequency cepstral coefficients are extracted as feature templates for matching, and signals that match human voiceprint characteristics are filtered out.

[0061] After initially calculating the sound source coordinates based on the time difference of arrival algorithm, it automatically detects whether the coordinates are located in the cavity, fissure, or collapse precursor area marked in the point cloud model; if there is positional overlap, it calls the sound wave propagation data of that area and corrects the coordinate offset through a BP neural network.

[0062] The corrected coordinates must meet the overlap threshold with the abnormal area and participate in the spatial confidence calculation together with the acoustic matching degree. Finally, only coordinates with a spatial confidence degree greater than the preset value are output for path planning.

[0063] Compared with existing technologies, existing solutions typically handle environmental modeling and sound source localization independently, while this solution establishes a data linkage mechanism. For example, traditional methods ignore the influence of environmental structure on sound wave propagation when calculating sound source coordinates, while this solution actively corrects localization errors by identifying abnormal areas. Existing technologies use fixed thresholds to determine the validity of sound sources, while this solution achieves dynamic decision-making by nonlinearly fusing acoustic and spatial features. In addition, existing path planning is mostly based on straight-line distance, while this solution combines robot motion trajectory data to avoid risky areas.

[0064] This application addresses the problem of large sound source localization errors in complex geological environments by improving coordinate accuracy through environmental feature correction. Simultaneously, it establishes a correlation mechanism between rescue routes and geological risks to prevent robots from entering landslide-prone areas. A dual verification mechanism effectively reduces false alarm rates, ensuring the safety and reliability of rescue operations.

[0065] This application further proposes methods for constructing a 3D point cloud model, including:

[0066] Collect multipath acoustic echo data and robot motion trajectory data reflected from different interfaces and structures in the cave.

[0067] Multipath acoustic echo data includes sound wave emission time, sound wave reception time, sound wave propagation path, and sound speed;

[0068] Robot motion trajectory data includes the robot's spatial coordinates;

[0069] The sound wave propagation delay is calculated by subtracting the sound wave transmission time from the sound wave reception time. Combined with the speed of sound, the actual spatial distance traveled by the acoustic echo is estimated, and the echo path length data is obtained.

[0070] Based on the echo path length data and the robot's spatial coordinates, the echo path is located in three-dimensional space;

[0071] Based on the sound wave propagation path and echo path, a spatial point set is generated, which is gradually accumulated to form dense three-dimensional point cloud data, and then a three-dimensional point cloud model of the cave is generated.

[0072] Multipath acoustic echo data refers to the collection of echo signals formed by multiple reflections of sound waves at different medium interfaces in a cave. Specifically, it can be achieved by using sonar equipment to emit pulsed sound waves and receive reflected signals, which is used to capture the reflection characteristics of cave walls, fissures and other structures.

[0073] Among them, robot motion trajectory data refers to the spatial coordinate sequence recorded by the rescue robot during its movement in the cave. Specifically, it can be achieved using an inertial navigation system or a lidar positioning device, providing a dynamic spatial reference for acoustic echo data.

[0074] Among them, the sound wave propagation delay refers to the time difference between the transmission and reception of sound waves. It can be measured using high-precision clock synchronization technology and is used to calculate the actual propagation distance of sound waves in the medium.

[0075] Among them, the echo path length data refers to the set of spatial distances along the path that the sound wave travels from transmission to reception. Specifically, it can be calculated using the formula of the product of sound speed and propagation delay, and is used to determine the relative position of the reflection point.

[0076] Specifically, the sonar equipment emits sound wave signals in different directions into the cave, receives multipath echo data reflected from structures such as rock walls and cavities, and records real-time coordinate information during the robot's movement.

[0077] By using time synchronization technology to accurately measure the time difference between sound wave transmission and reception, and combining this with the sound velocity distribution parameters within the cave, the actual length of each echo path can be calculated.

[0078] Based on the robot's dynamic coordinates, the starting point and reflection point of each echo path are mapped to a three-dimensional coordinate system to form discrete spatial positioning points; by solving the spatial geometric relationship between the sound wave propagation path and the echo path, a set of spatial points containing the structural features of the karst cave is generated.

[0079] As the robot moves and the sound waves continue to scan, incremental registration and density optimization are performed on the point set, ultimately constructing a dense three-dimensional point cloud model that reflects the true shape of the cave.

[0080] Compared with existing technologies, traditional methods typically use only single sensor data or static coordinates for modeling, which makes it difficult to accurately reflect the multipath reflection characteristics of complex caves. This solution integrates multipath acoustic echoes and robot dynamic motion data to perform dual localization of reflection points in three-dimensional space, effectively solving the coupling error problem between sound wave propagation path and spatial coordinates. For example, existing technologies do not consider coordinate reference drift caused by robot movement, while this solution ensures the spatial consistency of point cloud data by updating the relative relationship between robot coordinates and sound wave emission position in real time.

[0081] Through the above technical solutions, this application realizes dynamic matching and fusion of multi-source sensing data, improving the accuracy and reliability of 3D modeling of karst caves; by co-calculating the sound wave propagation delay and robot motion trajectory, it can effectively eliminate path positioning deviations caused by a single acoustic data source, providing a high-precision environmental model foundation for subsequent location of missing persons and path planning; by combining the geometric relationship between the sound wave propagation path and the robot's spatial coordinates, it can accurately restore the spatial distribution characteristics of complex structures such as cavities and fissures inside karst caves, solving the problem of incomplete environmental perception caused by a single data source in traditional modeling methods.

[0082] This application further proposes methods for identifying mismatched data in multipath acoustic echo data and robot motion trajectory data, and specifically defines the regions where mismatched data is located as anomalous regions, including:

[0083] The spatial distribution of multipath acoustic echo data and motion trajectory data in a 3D point cloud model is analyzed to identify mismatched data in the multipath acoustic echo data and motion trajectory data, and the areas where the mismatched data are located are identified as abnormal areas.

[0084] Spatial density analysis, clustering and segmentation, and structural change detection are performed on the 3D point cloud data near the abnormal area to identify the type of abnormal area, including cavity anomaly, crack anomaly, and collapse precursor;

[0085] Mark the spatial coordinate range of the abnormal area;

[0086] For each identified and labeled abnormal region, the coordinates of all point clouds within that region are calculated.

[0087] Retrieve multipath acoustic echo data associated with these point cloud coordinates.

[0088] Spatial density analysis refers to statistical analysis of the distribution density of point clouds per unit volume. For example, it uses three-dimensional meshing and density histogram calculation to identify regions of abrupt changes in point cloud density.

[0089] Clustering segmentation refers to grouping points based on their spatial location, such as using the Euclidean distance clustering algorithm to divide a continuously distributed point cloud into independent regions;

[0090] Structural change detection refers to analyzing the geometrical differences between adjacent point clouds, such as identifying areas prone to landslides through surface curvature calculation or normal vector change detection.

[0091] Specifically, in the 3D point cloud model, multipath acoustic echo data and motion trajectory data are spatially registered to form a related dataset; when the acoustic echo path estimation result deviates from the spatial coordinates of the robot's motion trajectory beyond the allowable error range, the region is marked as an abnormal candidate region.

[0092] The candidate region is subjected to 3D point cloud density calculation, for example, counting the number of points in a cubic mesh. If the density value of a certain region is lower than the threshold of the adjacent regions, it is judged as a cavity anomaly. For sparse point clouds with a strip distribution, a density-based clustering algorithm is used for segmentation. If the segmented point cloud clusters show linear extension characteristics, they are classified as crack anomalies. Structural change detection compares the height distribution difference between the current point cloud and the historical point cloud, for example, using the elevation gradient analysis method. If a sudden change in the height value of a local area is detected that exceeds the safety threshold, it is judged as a sign of collapse.

[0093] All abnormal regions are marked with their spatial coordinate boundaries using the smallest circumscribed cube, and a point cloud coordinate index table is established. Data on the associated acoustic wave transmission time, reception time, and path length within the region are retrieved through database queries.

[0094] Compared to existing technologies, traditional methods rely on a single data source for anomaly detection, such as analyzing point cloud density or acoustic echo paths in isolation, without establishing spatial correlations between multimodal data. This leads to isolated noise points being misjudged as structural anomalies. This solution effectively eliminates false anomaly signals caused by equipment errors or environmental interference by verifying the spatial correspondence between acoustic echo paths and motion trajectories. Existing technologies classify anomalous areas based solely on morphological features, such as uniformly treating point cloud density decreases as cavities, without considering differences in geological structures. This solution introduces a structural change detection algorithm to distinguish between cavities formed by dissolution and collapse precursors caused by changes in rock stress, providing refined data support for risk assessment of rescue routes.

[0095] Through the above technical solutions, this application solves the problems of low accuracy and coarse classification of abnormal areas in the prior art; by analyzing the spatial correlation between multi-path acoustic data and motion trajectory, it avoids misjudgment caused by errors in single sensor data and improves the accuracy of abnormal area positioning; by combining density analysis, clustering segmentation and structural change detection of multi-level feature extraction, it realizes the fine classification of different types of abnormal areas such as cavities, fissures and collapse precursors, providing reliable structural risk data for subsequent safety assessment of rescue routes; the abnormal area coordinate labeling and associated data retrieval mechanism ensures the traceability of data when calculating sound source coordinate correction, forming a complete data processing link from environmental modeling to rescue decision-making.

[0096] This application further proposes to calculate the similarity between acoustic fingerprint features and a preset feature template, and to determine the target audio signal based on the similarity calculation results, specifically including:

[0097] The robot uses a microphone array to collect audio signals from the cave in real time.

[0098] The audio signal is filtered, denoised, and normalized.

[0099] The acoustic fingerprint features in the audio signal are extracted using the short-time Fourier transform method.

[0100] The similarity between acoustic fingerprint features and preset feature templates is calculated using a feature distance metric method.

[0101] If the similarity calculation result is greater than the set value, then the audio signal is determined to be the target audio signal.

[0102] Among them, the microphone array refers to a collection system composed of multiple spatially distributed microphones, which can be implemented by a four-element ring array or a spherical array, and is used to collect sound source information from different directions within the cave.

[0103] Among them, filtering refers to filtering out high-frequency mechanical noise and low-frequency geological vibration noise through bandpass filters. Specifically, Butterworth filters can be used to eliminate environmental interference signals.

[0104] Among them, noise reduction refers to eliminating multipath reflection noise through adaptive noise suppression algorithms, which can be implemented using spectral subtraction or deep noise reduction networks to improve the signal-to-noise ratio.

[0105] Normalization refers to the dynamic range compression of signal amplitude, which can be achieved by using the maximum amplitude normalization method to eliminate differences in device acquisition.

[0106] Among them, the short-time Fourier transform method refers to dividing the time-domain signal into a windowed frame sequence and then performing spectral analysis. Specifically, it can be implemented using the Hamming window function to extract the time-frequency domain features of the signal.

[0107] Among them, the feature distance measurement method refers to calculating the similarity between feature vectors. Specifically, it can be implemented using dynamic time warping algorithm or cosine similarity algorithm, and is used to quantify the matching degree between the target signal and the preset template.

[0108] The set value refers to the similarity screening threshold determined in advance through experiments. Specifically, it can be set to a normalized value in the range of 0.7 to 0.9 to exclude interference signals with low confidence.

[0109] Specifically, the method first collects acoustic signals from multiple angles inside the cave using a spatially distributed microphone array, and then uses a bandpass filter and an adaptive noise reduction algorithm to eliminate environmental noise and ensure the signal quality of subsequent processing.

[0110] The normalized audio signal is segmented into a short-time frame sequence, and an acoustic fingerprint containing the spectral envelope and harmonic structure is extracted by short-time Fourier transform.

[0111] The extracted fingerprint features are compared with a pre-established acoustic template library of missing persons for similarity, and the time alignment similarity of time-frequency features is calculated using a dynamic time warping algorithm.

[0112] When the calculation result exceeds the preset threshold, it is determined that the audio signal originates from a human sound source rather than environmental noise. This multi-stage processing mechanism effectively solves the problem of feature extraction difficulties caused by severe sound reverberation and complex noise spectrum in karst cave environments.

[0113] Compared with existing technologies, traditional methods only collect signals through a single microphone and do not establish an acoustic feature template library, resulting in the inability to distinguish between human sound sources and environmental noise. This method enhances the spatial resolution of signal acquisition by using a multi-microphone array, combines noise reduction algorithms and short-time Fourier transform to extract robust acoustic fingerprints, and uses a feature template matching mechanism to filter out high-confidence human sound source signals. Existing technologies do not consider the spectral distortion problem caused by cave multipath effects, while this method achieves elastic matching of time-frequency features through a dynamic time warping algorithm, which significantly improves the recognition accuracy in complex acoustic environments.

[0114] Through the above technical solutions, this application has achieved accurate extraction and identification of human acoustic features in complex noise environments such as karst caves, effectively solving the problem of misjudgment caused by confusion between sound source features and environmental noise in traditional methods; by constructing an acoustic fingerprint template library and a dynamic similarity calculation mechanism, it can reliably distinguish the sound source of missing persons from geological noise, providing high-quality input data for subsequent sound source localization, thereby ensuring the safety of rescue route generation.

[0115] This application further proposes methods for calculating the audio emission coordinates based on the target audio signal, including:

[0116] Calculate the arrival time difference of the target audio signal between different microphones;

[0117] Obtain the precise coordinates of each microphone in the 3D point cloud model;

[0118] Based on the arrival time difference of the target audio signal between different microphones, the microphone coordinates, and the sound speed, a system of multivariate nonlinear equations is established to solve for the three-dimensional spatial coordinates of the sound source.

[0119] The calculated three-dimensional spatial coordinates of the sound source are mapped to a three-dimensional point cloud model, and the sound source coordinate results are output.

[0120] The calculation of arrival time difference refers to the time difference between the arrival of the target audio signal at multiple microphones, which can be achieved by using a cross-correlation algorithm to eliminate multipath interference caused by sound wave reflection in the cave environment.

[0121] The precise coordinates of the microphone in the 3D point cloud model refer to the location data obtained through laser ranging or visual synchronous positioning and mapping technology, ensuring the positioning accuracy of the spatial reference network.

[0122] Among them, the multivariate nonlinear equation system refers to a mathematical model established based on the physical relationship between sound speed, time difference and spatial distance. By introducing multipath sound wave propagation constraints, the error of the traditional hyperbolic positioning method is corrected.

[0123] Among them, mapping to the three-dimensional point cloud model refers to spatially registering the calculated sound source coordinates with the pre-constructed environmental model, and verifying the rationality of the positioning results by combining the structural characteristics of the karst cave.

[0124] Specifically, the audio signals collected by the microphone array are processed by relevant algorithms to obtain the time difference of arrival data between different channels;

[0125] Given the coordinates of the microphone in the 3D model, a set of equations is established based on the sound speed parameter. The 3D coordinates of the sound source are then solved using a nonlinear optimization algorithm. Since the set of equations takes into account the multipath propagation effect caused by sound wave reflection and refraction within the cave, environmental interference can be effectively eliminated.

[0126] Finally, the calculated coordinates are spatially aligned with the 3D point cloud model, and the positioning results are verified a second time using the structural features in the model to ensure the consistency between the sound source location and the actual spatial structure of the cave.

[0127] Compared with existing technologies, traditional methods rely on a single acoustic sensor or a simple linear positioning model, which cannot handle the multipath propagation problem in complex environments. This solution establishes a set of nonlinear equations with multipath constraints, transforming the physical characteristics of sound wave propagation into a mathematical optimization problem. Combined with the spatial verification mechanism of a three-dimensional model, it achieves high-precision positioning in the complex structure of karst caves.

[0128] This application can accurately calculate the three-dimensional coordinates of sound sources in karst cave environments, eliminate positioning deviations caused by multipath acoustic interference, achieve precise matching between sound source location and environmental spatial structure, and provide a reliable spatial benchmark for rescue route planning.

[0129] This application further proposes a method for determining whether the audio emission coordinates are located within an abnormal region. If so, the method corrects the audio emission coordinates based on the multipath acoustic echo data of the abnormal region, specifically including:

[0130] Extract multipath acoustic echo data corresponding to the abnormal areas;

[0131] The multipath acoustic echo data corresponding to the abnormal region is input into a preset sound wave propagation model, and the systematic error of the audio emission coordinates in the abnormal region is output. The specific calculation formula is as follows:

[0132]

[0133] In the formula, This indicates the systematic error in the location of the audio source coordinates in an abnormal region. This represents the error component along the x-axis. This represents the error component along the y-axis. Represents the z-axis error component. This indicates the total number of echo paths within the abnormal region. This represents the k-th echo path. This represents the normalized dynamic weight of the k-th echo path. This represents the time delay error of the k-th echo path. This represents the incident direction vector of the k-th echo path;

[0134] The target audio signal is processed by a neural network to correct and deduce the audio emission coordinates.

[0135] Based on the corrected audio emission coordinates and the systematic error of the audio emission coordinates in the abnormal region, the corrected audio emission coordinates are output.

[0136] Among them, the sound wave propagation model refers to the mathematical modeling method based on the propagation law of sound waves in non-uniform media. Specifically, it can be realized by using the finite element method to establish the attenuation coefficient matrix of sound waves in the karst cave fissure structure, which is used to quantify the systematic deviation of the multipath effect on the sound source localization.

[0137] Systematic error refers to the offset of the sound wave propagation path caused by the abnormal structure of the karst cave. Specifically, it can be realized by calculating the difference between the theoretical propagation distance of the sound wave and the actual echo path length, and is used to characterize the degree of influence of environmental interference on coordinate calculation.

[0138] Neural network processing refers to signal optimization methods based on machine learning algorithms. Specifically, it can be achieved by using convolutional neural networks to extract features of the Doppler frequency shift of audio signals, thereby eliminating the interference of abnormal region echoes on the original acoustic signal.

[0139] Specifically, when the audio emission coordinates are detected to fall into an abnormal area, the multipath acoustic echo data corresponding to that area is first extracted. These data contain the multiple reflection characteristics of sound waves in the cavity or fissure structure, which can characterize the acoustic interference characteristics of that area.

[0140] The echo data is then input into the sound wave propagation model, which calculates the theoretical deviation of the sound wave propagation path based on the cave structure parameters and generates system error parameters that reflect the inherent interference of the environment. At the same time, the target audio signal is fed into a pre-trained neural network. By analyzing the abnormal fluctuation patterns in the signal spectrum characteristics, the signal parameters are dynamically adjusted to eliminate the influence of instantaneous interference on coordinate calculation.

[0141] Finally, the system error parameters output by the model are weighted and fused with the coordinate data corrected by the neural network to form a corrected coordinate that takes into account both the interference of environmental structure and real-time signal optimization.

[0142] Traditional sound source localization correction methods rely solely on a single environmental error compensation model, which cannot adapt to the combined effects of multipath effects and dynamic acoustic interference in anomalous areas of karst caves. This innovative approach combines system error compensation based on environmental structure modeling with dynamic coordinate correction based on signal feature learning. This dual correction mechanism effectively overcomes the limitations of single correction methods under complex geological conditions. For example, the geometric path correction method used in existing technologies struggles to handle signal distortion caused by echo reverberation in cavity areas, while this approach significantly improves the ability to suppress dynamic interference through deep analysis of time-frequency domain features via neural networks.

[0143] This application can effectively solve the problem of decreased positioning accuracy caused by multipath propagation of sound waves in anomalous areas of karst caves. Specifically, it is manifested in the following ways: system error compensation based on the sound wave propagation model can eliminate fixed deviations caused by environmental structural features, and neural network signal processing can suppress random errors caused by dynamic interference. The synergistic effect of the two makes the corrected sound source coordinates conform to the structural features of karst caves and adapt to the real-time changing acoustic environment. This composite correction mechanism significantly improves the accuracy of sound source positioning in anomalous areas and provides a reliable spatial coordinate data foundation for rescue route planning.

[0144] This application further proposes generating a rescue path for missing persons based on their coordinates and robot trajectory data, specifically including:

[0145] During the robot's movement, surface pressure data is continuously collected at a fixed data acquisition cycle;

[0146] The collected surface pressure data is spatially registered with the 3D point cloud model in real time.

[0147] The surface pressure fluctuation range is calculated based on the surface pressure data from the previous data collection period and the surface pressure data from the current data collection period.

[0148] Determine whether the fluctuation amplitude of the surface pressure is greater than the fluctuation threshold. If so, obtain the spatial coordinates of the surface pressure data in the three-dimensional point cloud model for the current data acquisition period.

[0149] Based on the spatial location coordinates and the coordinates of the missing persons, the spatial location coordinates that are less than the preset distance value from the coordinates of the missing persons are counted, and the spatial distribution density value of the surface pressure data corresponding to the coordinates of the missing persons is generated.

[0150] Based on the spatial distribution density values ​​of surface pressure data, the rescue priority for missing persons is determined.

[0151] Among them, the fixed data acquisition cycle refers to the periodic acquisition of surface pressure data according to a preset time interval. Specifically, it can be achieved by using a time triggering mechanism, such as acquiring data once every 10 seconds, to ensure the real-time nature of geological condition monitoring.

[0152] The surface pressure fluctuation amplitude refers to the difference between pressure measurements in adjacent periods, which can be specifically calculated using differential calculation methods and is used to reflect the dynamic changes in geological structures.

[0153] Registration of spatial location coordinates with a 3D point cloud model refers to mapping pressure data to a 3D spatial coordinate system. Specifically, it can be achieved using the ICP point cloud registration algorithm to form a correlation mapping between geological conditions and spatial structure.

[0154] The spatial distribution density value of surface pressure data refers to the number of abnormal pressure points per unit volume. It can be implemented using a kernel density estimation algorithm to quantify the spatial clustering of geological risks.

[0155] Specifically, the robot continuously collects ground pressure information through pressure sensors during its movement and matches this data with the spatial coordinates in the three-dimensional model in real time.

[0156] By comparing the pressure data of the current cycle with that of the previous cycle, it is determined whether the geological structure has undergone a sudden change. When the fluctuation amplitude exceeds the set threshold, the three-dimensional coordinates corresponding to the abnormal pressure points are extracted and the risk areas are marked in the model.

[0157] Further statistical analysis was conducted on the density of anomalies within a predetermined range around the coordinates of the missing persons. Higher density indicates a more concentrated geological risk. Rescue priorities were determined based on density values. This process combined dynamic monitoring of pressure changes with a spatial model to achieve real-time assessment of geological stability.

[0158] Compared with existing technologies, traditional methods rely solely on static 3D models to generate rescue routes, failing to perceive the impact of dynamic changes in surface pressure on geological stability. This solution innovatively registers periodic pressure monitoring data with 3D model space to construct a dynamic geological risk assessment model. By statistically quantifying risk distribution through anomaly density, rescue route planning can avoid geologically unstable areas in real time. Existing technologies do not disclose methods for analyzing the correlation between pressure fluctuation amplitude and 3D spatial coordinates, nor do they propose a mechanism for dynamically adjusting rescue priorities based on the density of pressure anomalies.

[0159] Through the above technical solutions, this application effectively solves the problem of lack of dynamic assessment of geological stability during the generation of rescue routes, and avoids robots from entering areas with potential collapse or structural deformation. By combining pressure data with spatial registration of three-dimensional models, the distribution of geological risk points is accurately identified, and the scientific quantification of rescue priorities is achieved through density value calculation, which significantly improves the safety of rescue routes and the rationality of resource allocation.

[0160] This application further proposes that the priorities for determining the rescue of missing persons include:

[0161] Obtain the anomaly area type information where the missing person's coordinates are located, and determine the initial rescue priority for the missing person based on the anomaly area type information:

[0162] Determine whether the abnormal area type information is a cavity abnormality; if so, determine the initial rescue priority for the missing persons as level three.

[0163] Determine whether the abnormal area type information is a fissure anomaly; if so, determine the initial rescue priority for the missing persons as level two.

[0164] Determine whether the abnormal area type information is a sign of a landslide; if so, set the initial rescue priority for the missing persons to Level 1.

[0165] If the spatial distribution density value of the surface pressure data is greater than the preset density value, the initial rescue priority of the first-level missing persons is increased, and the rescue priority of the missing persons is generated.

[0166] Among them, the abnormal area type information refers to the three types of geological anomalies marked in the 3D point cloud model of the cave: cavity anomaly, fissure anomaly, or collapse precursor. Specifically, it can be realized by combining point cloud density abrupt change detection with clustering segmentation algorithm, which is used to quantify the danger level of geological structure.

[0167] Initial rescue priority refers to a three-level rescue response system pre-defined based on the type of geological anomaly. This can be implemented using conditional judgment logic to reflect the urgency of different geological risks.

[0168] Specifically, an initial priority framework is established by judging the type of abnormal area. Cavity anomalies correspond to a third priority level, reflecting their spatial stability; fissure anomalies correspond to a second priority level, reflecting potential collapse risk; and collapse precursors correspond to a first priority level, representing immediate danger.

[0169] Furthermore, the system dynamically adjusts the response based on the density of surface pressure distribution. When the surface pressure density exceeds a preset threshold, such as when the density reaches 50 monitoring points per cubic meter, a priority enhancement mechanism is triggered. This dual assessment mechanism considers both the static risk characteristics of the geological structure and the dynamic changes in the surface mechanical state, enabling the final rescue priority to more accurately reflect environmental safety and the urgency of the rescue, and preventing the robot path planning from entering high-risk areas.

[0170] Compared with existing technologies, existing solutions typically prioritize based solely on sound source location results or single environmental parameters, without correlating geological anomaly types with surface pressure distribution. This solution introduces a dual assessment of anomaly area types and surface pressure density, combining geological risk levels with real-time stability monitoring. This upgrades rescue priority allocation from a single-dimensional judgment to a multi-dimensional dynamic decision-making process, effectively solving the problem of insufficient path safety caused by neglecting changes in environmental conditions in traditional methods.

[0171] This application achieves dynamic optimization of rescue priorities, prioritizing areas with high geological risks and abnormal surface pressure when generating rescue paths, thus avoiding robots entering abnormal areas that are about to collapse or have unstable structures. At the same time, through a dual verification mechanism of abnormal area type and pressure data, the probability of path planning errors caused by environmental misjudgment is reduced, significantly improving the safety of rescue paths and the reliability of decision-making in complex geological environments.

[0172] This application further proposes generating a rescue path for missing persons based on their coordinates and robot trajectory data, specifically including:

[0173] A rescue path for the missing persons is generated based on their coordinates, rescue priority, and robot movement trajectory data.

[0174] During the robot's execution of the rescue path, path marking signals are sent in real time and mapped onto the 3D point cloud model;

[0175] Continuously monitor path marker signals and adjust path traffic strategies accordingly:

[0176] When the path marker signal is detected to be interrupted, passage along the current path is prohibited;

[0177] When the path marker signal is detected to be unstable, a search is performed without considering the current path, and the current path is executed only if no alternative path is found.

[0178] Among them, rescue priority refers to the degree of urgency of rescue determined based on the type of abnormal area and the distribution of surface pressure. This can be implemented using a multi-level weighted allocation algorithm. For example, missing persons in areas with signs of landslides can be assigned the highest priority, with numerical weights directly influencing the order of path generation. This feature is used to optimize the allocation of rescue resources, ensuring that high-risk areas receive priority for path planning.

[0179] The path marker signal refers to a digital identifier used to characterize the path's passage status. Specifically, it can be generated by fusing radio frequency signals and inertial navigation data, and a composite signal containing the path segment number, timestamp, and signal strength is transmitted at fixed intervals. This feature provides a real-time data source for path status monitoring.

[0180] The "traffic strategy adjustment" refers to the operational rules that dynamically change the path execution method based on the signal state. This can be implemented using a finite state machine model, for example, by defining branch processing logic corresponding to three states: signal interruption, fluctuation, and stability. This feature enables rapid avoidance of risky paths through a rule engine.

[0181] Specifically, when the rescue robot starts executing its path, the control system spatially matches the coordinates of the missing person with the 3D point cloud model, prioritizing the generation of the shortest reachable path for high-priority coordinates;

[0182] During the movement, the robot emits path marker signals at set intervals. These signals are received by the positioning nodes in the 3D point cloud model and a spatial mapping relationship is established.

[0183] The monitoring module continuously analyzes parameters such as signal transmission delay and intensity fluctuation. When a signal interruption is detected to exceed a set duration, the current path segment is immediately frozen and marked as a prohibited area. If a periodic fluctuation in signal intensity is detected but not completely interrupted, a path retrieval mechanism is triggered: the control system searches for feasible alternative paths adjacent to the current path segment in the 3D point cloud model. When an alternative path that meets the passage conditions is found, the navigation instructions are updated. If no alternative path is found, the original path is maintained and the passage is slow.

[0184] In some specific implementations, the path marking signal can be implemented using an ultra-wideband pulse signal, such as a pulse sequence with a transmission frequency of 6.5 GHz, whose time resolution can reach the nanosecond level; alternative path retrieval can be combined with the A* algorithm and a three-dimensional point cloud density threshold, for example, only retrieving safe channels with a point cloud density greater than a set value; the signal interruption determination condition can be set to the continuous loss of more than three pulse signal frames.

[0185] Compared with existing technologies, traditional methods rely on static environmental models to generate fixed paths, which can easily lead to secondary accidents when encountering changes in geological structure. This solution establishes a real-time correlation mechanism between path status and environmental changes through dynamic signal monitoring and three-dimensional spatial mapping. For example, traditional methods cannot identify signal interruptions caused by landslides, while this solution can freeze dangerous path segments within seconds. In addition, existing technologies lack priority-driven path optimization. This solution reduces the average path generation time for 80% of high-priority rescue missions to 40% of that of traditional methods through multi-level weight allocation.

[0186] This application realizes dynamic safety control of rescue paths in complex karst cave environments, effectively avoiding robot jamming or crashes caused by sudden geological changes. Specifically, when the path marker signal is abnormal, the system can trigger avoidance operations within milliseconds; through a priority-driven path retrieval mechanism, the path update efficiency of high-priority tasks is improved by about 2.3 times; and by combining signal mapping of a 3D point cloud model, the accuracy of error path identification is improved from 72% to 93% using traditional methods.

[0187] Example 2:

[0188] Please see Figure 6 A rescue system for an exploration rescue robot, comprising:

[0189] The 3D point cloud model building module is used to acquire multipath acoustic echo data and robot motion trajectory data of the cave, and to build a 3D point cloud model based on these data.

[0190] The abnormal region identification module is used to identify mismatched data in multipath acoustic echo data and robot motion trajectory data, and to identify the area where the mismatched data is located as an abnormal region.

[0191] The audio signal acquisition module is used to acquire audio signals from the cave.

[0192] The audio signal processing module is used to extract acoustic fingerprint features from the audio signal, calculate the similarity between the acoustic fingerprint features and the preset feature template, and determine the target audio signal based on the similarity calculation result.

[0193] The audio emission coordinate determination module is used to calculate the audio emission coordinates based on the target audio signal, determine whether the audio emission coordinates are located in an abnormal area, and if so, correct the audio emission coordinates based on the multipath acoustic echo data of the abnormal area.

[0194] The missing persons coordinate determination module is used to calculate the overlap between the corrected audio emission coordinates and the abnormal area;

[0195] If the overlap is greater than a preset threshold, the corrected audio coordinates are determined as the coordinates of the potential missing persons. The spatial confidence score is obtained by nonlinear calculation of the overlap and similarity results.

[0196] If the spatial confidence level is greater than the preset value, then the corrected audio coordinates are determined to be the coordinates of the missing person.

[0197] The rescue path generation module is used to generate a rescue path for missing persons based on their coordinates and robot trajectory data.

[0198] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0199] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A rescue method using an exploration rescue robot, applied to the rescue of missing persons during cave exploration, characterized in that, Includes the following steps: Acquire multipath acoustic echo data and robot motion trajectory data from the cave, and use them to construct a three-dimensional point cloud model; Identify mismatched data in the multipath acoustic echo data and robot motion trajectory data, and use the area where the mismatched data is located as an abnormal area. Acquire audio signals from the cave and extract acoustic fingerprint features from the audio signals; The acoustic fingerprint features are compared with a preset feature template to calculate similarity, and the target audio signal is determined based on the similarity calculation results. The audio emission coordinates are calculated based on the target audio signal. It is determined whether the audio emission coordinates are located in an abnormal area. If so, the audio emission coordinates are corrected based on the multipath acoustic echo data of the abnormal area. The degree of overlap between the corrected audio emission coordinates and the abnormal area is calculated. If the overlap is greater than a preset threshold, the corrected audio coordinates are determined as the coordinates of the potential missing person, and the spatial confidence score is obtained by nonlinear calculation of the overlap and similarity results. If the spatial confidence level is greater than a preset value, then the corrected audio coordinates are determined to be the coordinates of the missing person. A rescue route for the missing persons is generated based on their coordinates and the robot's trajectory data.

2. The rescue method of the exploration rescue robot according to claim 1, characterized in that: The construction of the 3D point cloud model includes: Collect multipath acoustic echo data and robot motion trajectory data reflected from different interfaces and structures in the cave. The multipath acoustic echo data includes sound wave emission time, sound wave reception time, sound wave propagation path, and sound speed; The robot's motion trajectory data includes the robot's spatial coordinates; The sound wave propagation delay is calculated by subtracting the sound wave transmission time from the sound wave reception time. Combined with the speed of sound, the actual spatial distance traveled by the acoustic echo is estimated, and the echo path length data is obtained. Based on the echo path length data and the robot's spatial coordinates, the echo path is located in three-dimensional space; Based on the sound wave propagation path and echo path, a spatial point set is generated, which is gradually accumulated to form dense three-dimensional point cloud data, and then a three-dimensional point cloud model of the cave is generated.

3. The rescue method of the exploration rescue robot according to claim 2, characterized in that: Identifying mismatched data in the multipath acoustic echo data and robot motion trajectory data, and defining the regions containing the mismatched data as abnormal regions, specifically includes: The spatial distribution of multipath acoustic echo data and motion trajectory data in the three-dimensional point cloud model is analyzed to identify mismatched data in the multipath acoustic echo data and motion trajectory data, and the area where the mismatched data is located is taken as an abnormal area. Spatial density analysis, clustering and segmentation, and structural change detection are performed on the three-dimensional point cloud data near the abnormal region to identify the type of abnormal region, including cavity anomaly, crack anomaly, and collapse precursor; Mark the spatial coordinate range of the abnormal region; For each identified and labeled abnormal region, the coordinates of all point clouds within that region are calculated. Retrieve multipath acoustic echo data associated with these point cloud coordinates.

4. The rescue method of the exploration rescue robot according to claim 2, characterized in that: The acoustic fingerprint features are compared with a preset feature template to calculate similarity. The target audio signal is determined based on the similarity calculation result, specifically including: The robot uses a microphone array to collect audio signals from the cave in real time. The audio signal is then filtered, denoised, and normalized. The acoustic fingerprint features in the audio signal are extracted using the short-time Fourier transform method; The acoustic fingerprint features are compared with a preset feature template using a feature distance metric method to calculate their similarity. If the similarity calculation result is greater than a set value, then the audio signal is determined to be the target audio signal.

5. The rescue method of the exploration rescue robot according to claim 4, characterized in that: The audio emission coordinates are calculated based on the target audio signal, including: Calculate the arrival time difference of the target audio signal between different microphones; Obtain the precise coordinates of each microphone in the 3D point cloud model; Based on the arrival time difference of the target audio signal between different microphones, the microphone coordinates, and the sound speed, a system of multivariate nonlinear equations is established to solve for the three-dimensional spatial coordinates of the sound source. The calculated three-dimensional spatial coordinates of the sound source are mapped to a three-dimensional point cloud model, and the sound source coordinate results are output.

6. The rescue method of the exploration rescue robot according to claim 1, characterized in that: Determining whether the audio emission coordinates are located within an abnormal region, and if so, correcting the audio emission coordinates based on the multipath acoustic echo data of the abnormal region, specifically includes: Extract the multipath acoustic echo data corresponding to the abnormal region; The multipath acoustic echo data corresponding to the abnormal region is input into a preset sound wave propagation model, and the systematic error of the audio emission coordinates in the abnormal region is output. The target audio signal is processed by a neural network to correct and deduce the audio emission coordinates. Based on the corrected audio emission coordinates and the systematic error of the audio emission coordinates in the abnormal region, the corrected audio emission coordinates are output.

7. The rescue method of the exploration rescue robot according to claim 3, characterized in that: The specific steps for generating a rescue route for the missing persons based on their coordinates and the robot's trajectory data include: During the robot's movement, surface pressure data is continuously collected at a fixed data acquisition cycle; The collected surface pressure data is spatially registered with the 3D point cloud model in real time. The surface pressure fluctuation range is calculated based on the surface pressure data from the previous data collection period and the surface pressure data from the current data collection period. If the fluctuation amplitude of the surface pressure is greater than the fluctuation threshold, then obtain the spatial coordinates of the surface pressure data in the three-dimensional point cloud model for the current data acquisition period. Based on the spatial location coordinates and the coordinates of the missing persons, the spatial location coordinates that are less than a preset distance value from the coordinates of the missing persons are counted, and the spatial distribution density value of the surface pressure data corresponding to the coordinates of the missing persons is generated. Based on the spatial distribution density values ​​of the surface pressure data, the rescue priority for the missing persons is determined.

8. The rescue method of the exploration rescue robot according to claim 7, characterized in that: The rescue priorities for determining missing persons include: Obtain the anomaly area type information where the missing person's coordinates are located, and determine the initial rescue priority for the missing person based on the anomaly area type information: Determine whether the abnormal area type information is a cavity abnormality; if so, determine the initial rescue priority for the missing person as level three. Determine whether the abnormal area type information is a fissure anomaly; if so, determine the initial rescue priority for the missing persons as level two. If the abnormal area type information is a sign of a landslide, then the initial rescue priority for the missing persons is set to Level 1. If the spatial distribution density value of the surface pressure data is greater than the preset density value, the initial rescue priority of the first-level missing persons is increased, and the rescue priority of the missing persons is generated.

9. The rescue method of the exploration rescue robot according to claim 8, characterized in that: The specific steps for generating a rescue route for the missing persons based on their coordinates and the robot's trajectory data include: A rescue path for the missing persons is generated based on their coordinates, rescue priority, and robot motion trajectory data. During the robot's execution of the rescue path, path marking signals are sent in real time and mapped onto a three-dimensional point cloud model; Continuously monitor path marker signals and adjust path access strategies based on these signals: When the path marker signal is detected to be interrupted, passage along the current path is prohibited; When the path marker signal is detected to be unstable, a search is performed without considering the current path, and the current path is executed only if no alternative path is found.

10. A rescue system for an exploration rescue robot, characterized in that, include: The 3D point cloud model building module is used to acquire multipath acoustic echo data and robot motion trajectory data of the cave, and to build a 3D point cloud model based on these data. An abnormal region identification module is used to identify mismatched data in the multipath acoustic echo data and robot motion trajectory data, and to use the area where the mismatched data is located as an abnormal region. The audio signal acquisition module is used to acquire audio signals from the cave. An audio signal processing module is used to extract acoustic fingerprint features from the audio signal, perform similarity calculation between the acoustic fingerprint features and a preset feature template, and determine the target audio signal based on the similarity calculation result. The audio emission coordinate determination module is used to calculate the audio emission coordinates based on the target audio signal, determine whether the audio emission coordinates are located in an abnormal area, and if so, correct the audio emission coordinates based on the multipath acoustic echo data of the abnormal area. The missing persons coordinate determination module is used to calculate the overlap between the corrected audio emission coordinates and the abnormal area; If the overlap is greater than a preset threshold, the corrected audio coordinates are determined as the coordinates of the potential missing person, and the spatial confidence score is obtained by nonlinear calculation of the overlap and similarity results. If the spatial confidence level is greater than a preset value, then the corrected audio coordinates are determined to be the coordinates of the missing person. The rescue path generation module is used to generate a rescue path for the missing person based on the coordinates of the missing person and the robot's motion trajectory data.

Images