Disaster victim positioning method based on disaster scenes such as earthquake and landslide

Abstract

The invention relates to the technical field of disaster rescue, and particularly discloses a disaster victim positioning method based on disaster scenes such as earthquake, landslide and the like. The method comprises the following steps: collecting time-varying data of wireless signals of a post-disaster area, extracting a signal fluctuation coefficient, and identifying an accompanying agent node; a relative medium barrier index is calculated for the adjoint agent node, the burying depth difference between the trapped target and the adjoint node is evaluated, and a longitudinal layering strategy is formulated; controlling the search and rescue terminal to approach the adjoint agent node, monitoring signal response sensitivity to judge a blocking condition and determine a rescue channel; scanning by taking the adjoint proxy node as a reference, and locking a relative azimuth vector of a trapped target based on an energy gradient divergence feature; and generating a rescue instruction in combination with the longitudinal layering strategy and the relative azimuth vector. According to the method, the homologous interference effect and the energy gradient characteristics are utilized, and the relative positioning and safe excavation guiding functions in the non-line-of-sight environment are achieved.

Description

Technical Field

[0001] This invention relates to the field of disaster relief technology, specifically to a method for locating disaster victims in disaster scenarios such as earthquakes and landslides. Background Technology

[0002] With the widespread use of mobile smart terminals, passive positioning using wireless signals (such as Wi-Fi, Bluetooth, and cellular signals) emitted by the mobile phones carried by trapped people has become an important technical means for post-disaster search and rescue.

[0003] However, existing wireless signal positioning technologies (such as RSSI triangulation and fingerprint database matching) face technical challenges in extreme disaster scenarios such as earthquakes and landslides, mainly in the following aspects: The unsteady nature of the post-disaster physical environment causes traditional reference frames to fail. Existing indoor and outdoor positioning technologies typically assume that the positions of reference nodes (such as base stations and routers) are fixed. However, after a landslide or building collapse, these reference nodes often undergo unknown physical displacement along with the earthwork or building structure. If the pre-disaster static coordinate information is continued to be used as anchor points, it will lead to huge systematic biases in the positioning algorithm, resulting in a significant "drift error" between the calculated location of the trapped person and the actual location.

[0004] Therefore, this invention provides a method for locating disaster victims in disaster scenarios such as earthquakes and landslides. Summary of the Invention

[0005] The purpose of this invention is to provide a method for locating disaster victims in disaster scenarios such as earthquakes and landslides, so as to solve the aforementioned background problems.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] The method for locating disaster victims in scenarios such as earthquakes and landslides includes the following steps:

[0008] Collect time-varying data of wireless signals in the disaster area, extract the signal fluctuation coefficient of the trapped target signal based on the time-varying data, and identify the accompanying agent node with the same signal change trend of the trapped target based on the signal fluctuation coefficient;

[0009] For the accompanying proxy node, signal blocking analysis is performed in combination with the signal strength of the trapped target to obtain the medium blocking index; based on the medium blocking index, the difference in burial depth between the trapped target and the accompanying node is evaluated, and a vertical stratification strategy for excavation operations is formulated.

[0010] Implement a vertical stratification strategy to monitor the response sensitivity of the trapped target's signal strength as a function of distance; determine whether the response sensitivity conforms to the free space attenuation law, and establish a rescue channel based on the determination results;

[0011] Based on the rescue channel strategy, a circumferential micro-motion scan is performed with the accompanying agent node as the reference axis to extract the energy divergence of the trapped target signal relative to the accompanying agent node signal; based on the energy divergence, the azimuth sector with the strongest signal energy coupling is identified, and a relative azimuth vector is established.

[0012] Furthermore, the method for identifying the accompanying agent node is as follows:

[0013] The accompanying judgment logic is constructed based on the signal fluctuation coefficient, and high-confidence candidate object nodes are selected from all scanned environmental nodes, and a candidate accompanying set is established.

[0014] Perform an adjoint validity check on the candidate object nodes in the candidate adjoint set to determine the adjoint proxy nodes of the environment.

[0015] Furthermore, the signal fluctuation coefficient is obtained as follows:

[0016] Extract the instantaneous value of the received signal strength indication from the transmitter, continuously record the instantaneous values ​​of the received signal strength indication corresponding to the same media access control address, and construct a trapped target signal sequence of length T and N environmental node signal sequences;

[0017] The DC component is removed from any environmental node signal sequence of the acquired trapped target signal sequence;

[0018] Correlation analysis is performed on the signal sequence of any environmental node after removing the DC component from the trapped target signal sequence to obtain the signal fluctuation coefficient.

[0019] Furthermore, the signal blocking analysis is performed as follows:

[0020] Extract the signal strength sequence of the accompanying proxy node and the signal strength sequence of the trapped target;

[0021] The two signal sequences are time-aligned, the signal intensity difference at each sampling time is calculated, and a signal difference sequence is generated.

[0022] Calculate the mean of the signal difference sequence within the static observation window. ;

[0023] Will The absolute value and the system's preset reference strength Normalization is performed to obtain the normalized relative decay rate;

[0024] Calculate the standard deviation of the signal difference sequence to obtain the time-domain drift;

[0025] The relative attenuation rate and time-domain drift are normalized, and the two are linearly weighted and summed using preset weighting coefficients to obtain the relative dielectric barrier index.

[0026] Furthermore, the response sensitivity is obtained as follows:

[0027] A vertical hierarchical strategy is implemented, and the fastest moving trajectory accompanied by the signal enhancement of the agent node is planned to obtain the accompanying target path;

[0028] The signal strength of the trapped target and the signal strength of the accompanying agent node in the following path are collected, and the two signal strengths are dynamically tracked and a distance-signal strength curve is established.

[0029] Differential processing is performed on the distance-signal strength curve to obtain the two signal growth slopes of the trapped target signal strength and the accompanying agent node signal strength on the accompanying follow path. Response sensitivity analysis is performed on the two signal growth slopes to obtain the response sensitivity.

[0030] Furthermore, the response sensitivity analysis is performed as follows:

[0031] Differentiate the distance-signal strength curves and calculate the signal growth slopes of both along the movement path, namely the trapped growth slope and the accompanying growth slope.

[0032] The response sensitivity is obtained by calculating the ratio of the trapped growth slope to the accompanying growth slope.

[0033] Furthermore, the energy divergence is extracted as follows:

[0034] The scanning range was set according to the rescue channel strategy, and an adaptive circumferential micro-motion scan was performed:

[0035] During the scanning process, the signals of the trapped target and the accompanying agent node are simultaneously acquired, and the energy gradient is obtained by energy gradient analysis.

[0036] Based on the energy gradient, the energy gradient divergence between the trapped target and the accompanying agent node at the same azimuth angle is extracted.

[0037] Furthermore, based on the phase azimuth vector combined with the vertical layering strategy, rescue instructions are formulated and the spatiotemporal uncertainty analysis of the trapped target signal is performed to obtain the signal uncertainty and divide the excavation operation into regions; at the same time, the convergence degree of the signal uncertainty is monitored and the boundary of the operation area is dynamically reduced.

[0038] Furthermore, the spatiotemporal uncertainty analysis is performed as follows:

[0039] Establish a follow-up observation window to simultaneously collect the fluctuation amplitude of the trapped target signal in the time dimension and the angular drift in the spatial dimension;

[0040] Second-order statistical operations are performed on the fluctuation amplitude to obtain the discrete variance in the time domain.

[0041] Gradient analysis of the angle drift is performed to obtain the spatial divergence entropy;

[0042] The signal uncertainty is obtained by summing the discrete variance in the time domain and the divergent entropy in the spatial domain after making them dimensionless.

[0043] Furthermore, the convergence degree is monitored by calculating the gradient convergence rate of the signal uncertainty over time, which characterizes the convergence degree of the signal uncertainty.

[0044] The beneficial effects of this invention are as follows:

[0045] 1. By quantifying the signal resonance between the trapped target and environmental nodes under external physical disturbances, accompanying proxy nodes located in the same physical collapse unit as the trapped individual are identified. This constructs a relatively correlated positioning reference anchor point with high physical correlation in the electromagnetic environment, improving the ability to identify the physical attribution of signal sources under low signal-to-noise ratio conditions. By comparing the differences in signal penetration loss and multipath dynamic stability between the accompanying node and the trapped target, the relative burial level and medium properties of the two in vertical space are calculated. This is beneficial for identifying physical scenarios of co-existence, deep burial, and shallow shielding, reducing the risk of structural instability and secondary damage to the trapped individual due to blind deep excavation or excessive stripping.

[0046] 2. By utilizing the difference in signal growth slope during the approach process of the search and rescue terminal, the common effects of free-space path loss can be eliminated, and the presence of high-density hard obstructions or multipath reflection virtual images on the propagation path can be identified. By distinguishing channel characteristics such as direct low-impedance, high-impedance hard obstructions, and non-line-of-sight multipath, the physical demolition difficulty of different paths can be predicted before excavation operations. This is beneficial for optimizing the spatial planning efficiency of rescue paths and reducing the time lost on ineffective or high-risk paths.

[0047] 3. Implement circumferential micro-motion scanning and energy gradient divergence analysis centered on the accompanying node. Utilize the anisotropic isomorphic characteristics of the trapped target and the accompanying node within the same physical cavity regarding the external radiation field, and locate the relative azimuth vector by finding highly synchronized sectors of the energy gradient change rate. This helps reduce the shortcomings of traditional amplitude positioning, which is susceptible to false peaks caused by multipath reflections.

[0048] 4. Map the random disturbance characteristics of wireless signals to the safety boundaries of excavation operations in physical space, and establish a regional dynamic adjustment mechanism based on gradient convergence rate. During the rescue operation, construct a dynamic feedback loop between signal quality and the scope of engineering operations. Adaptively adjust the spatial proportion of mechanical stripping and manual fine operations based on the real-time evolution trend of signal transmission characteristics. While maintaining the speed of rescue progress, reduce the possibility of secondary mechanical injuries to trapped personnel through flexible control of physical operation boundaries. Attached Figure Description

[0049] The invention will now be further described with reference to the accompanying drawings.

[0050] Figure 1 This is a flowchart of the method for locating disaster victims in disaster scenarios such as earthquakes and landslides, based on the present invention;

[0051] Figure 2 This is a flowchart of the process for obtaining response sensitivity according to the present invention;

[0052] Figure 3 This is a flowchart illustrating the various possibilities for determining rescue routes in this invention. Detailed Implementation

[0053] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0054] Example 1:

[0055] like Figure 1 As shown, the method for locating disaster victims in scenarios such as earthquakes and landslides includes the following steps:

[0056] S10. Collect time-varying data of wireless signals in the disaster area, extract the signal fluctuation coefficient of the trapped target signal based on the time-varying data; identify environmental nodes with consistent signal change trends of the trapped target based on the signal fluctuation coefficient, and mark the environmental nodes as the accompanying agent nodes of the trapped person.

[0057] The method for collecting time-varying data of wireless signals in the post-disaster area and extracting the signal fluctuation coefficient of the trapped target's signal based on the time-varying data is as follows:

[0058] In some embodiments, a drone equipped with a Wi-Fi probe or a search and rescue terminal controlled by a handheld search and rescue device is used to maintain a stationary spatial position at a predetermined detection point to establish a temporary static observation window;

[0059] The radio frequency scanning module is activated based on the static observation window, and beacon frames of all transmitting sources within the coverage area are captured by polling in the full frequency band (such as 2.4GHz / 5GHz).

[0060] Identify and parse the MAC (Media Access Control) address in each beacon frame as a device identifier, and extract the instantaneous value of Received Signal Strength Indicator (RSSI);

[0061] According to the preset high-frequency sampling frequency (e.g., 10Hz), the instantaneous value of the received signal strength indication corresponding to the same MAC address is continuously recorded to construct a trapped target signal sequence of length T and N environmental node signal sequences;

[0062] It should be noted that the trapped target signal sequence is a specific signal stream locked by knowing the trapped person's mobile phone information (such as IMSI or MAC), while the environmental node signal sequence is the signal stream of non-mobile facilities such as routers and smart home devices scanned on site.

[0063] The DC component of any environmental node signal sequence of the acquired trapped target signal sequence is removed, that is, the mean of each sequence within the observation window is subtracted to extract the AC component that reflects the signal jitter characteristics.

[0064] Correlation analysis is performed on the signal sequence of any environmental node after removing the DC component from the trapped target signal sequence to obtain the signal fluctuation coefficient, which is used to quantify the degree of linear correlation between the two over time.

[0065] Preferably, the Pearson correlation coefficient algorithm is used for correlation analysis; if the data exhibits nonlinear characteristics, the Spearman rank correlation coefficient is used.

[0066] It is understandable that the physical meaning of calculating the signal fluctuation coefficient is that: by utilizing the co-source interference effect, when a trapped target (such as a mobile phone) and an environmental node (such as a router) are buried under the same precast concrete slab or collapsed structure, the small physical disturbances caused by the settlement and vibration of the external environmental ruins and the movement of rescue personnel above will produce almost identical modulation effects on the signal transmission paths of the trapped target and the environmental node, resulting in the RSSI signals of the two exhibiting resonant fluctuations in the same frequency and direction.

[0067] Among them, the method of identifying environmental nodes with consistent signal change trends of the trapped target based on signal fluctuation coefficients and marking these environmental nodes as accompanying agent nodes of the trapped target is as follows:

[0068] S101. Construct an adjoint determination logic based on the signal fluctuation coefficient, and select high-confidence candidate object nodes from all scanned environmental nodes, and establish a candidate adjoint set;

[0069] Preferably, the method for constructing the accompanying decision logic is as follows:

[0070] The distribution of signal fluctuation coefficients of all environmental nodes within the current observation window is statistically analyzed, and a normal distribution fit test is performed.

[0071] If the data distribution conforms to a normal distribution, select the region located on the right side of the distribution curve with values ​​greater than [value missing]. The environment nodes are used as candidate object nodes, and a candidate companion set is established;

[0072] in, These are the mean and standard deviation of the signal fluctuation coefficients for all environmental nodes, respectively.

[0073] If the data distribution does not conform to the characteristics of a normal distribution, that is, the data is sparse or discrete, then the direct threshold truncation method is adopted, which directly traverses all nodes and selects nodes with signal fluctuation coefficients greater than the preset strong correlation limit as candidate object nodes and includes them in the candidate adjoint set.

[0074] If no node satisfies the boundary, the K nodes ranked first in the volatility coefficient sorting are selected as candidate nodes and marked as weak adjoint states.

[0075] Preferably, K=10;

[0076] S102. Perform an adjoint validity check on the candidate object nodes in the candidate adjoint set to determine the adjoint proxy nodes of the environment.

[0077] Preferably, the method for performing the adjoint validity test is as follows:

[0078] Find the absolute value of the received signal strength indication of the candidate object node in the candidate companion set;

[0079] It should be noted that, since signals are generally attenuated in deeply buried environments, signal strength is not the only criterion for elimination. Instead, nodes with RSSI values ​​close to the receiver sensitivity noise floor and fluctuation amplitude exhibiting disordered random noise characteristics are eliminated to reduce the artificially high random correlation caused by extremely low signal-to-noise ratio, i.e., pseudo-accompaniment nodes.

[0080] Among the nodes that pass the screening, the environmental node with the largest signal fluctuation coefficient and the highest data capture rate is selected and marked as the accompanying agent node;

[0081] As will be understood by those skilled in the art, the data capture rate is defined as the ratio of the number of beacon frames actually captured to the theoretically required number of beacon frames to be sent within a static observation window.

[0082] The signal fluctuation coefficient calculated by the environmental node is considered valid only when the ratio is greater than the preset integrity threshold (e.g., 0.95), and is used to eliminate the interference of random packet loss caused by channel congestion or multipath fading on the extraction of fluctuation features.

[0083] A logical binding relationship is established between the MAC address of the accompanying agent node and the ID of the trapped target, and the baseline difference between the signal strength of the two at the current moment is recorded as the baseline zero point for subsequent judgment of the burial depth of the trapped target relative to the accompanying node.

[0084] S20. For the accompanying agent node, perform signal blocking analysis based on the signal strength of the trapped target to obtain the medium blocking index; evaluate the difference in burial depth between the trapped target and the accompanying node based on the medium blocking index, and formulate a vertical stratification strategy for excavation operations.

[0085] Specifically, for the accompanying proxy node, the method for obtaining the medium barrier index by combining the signal barrier analysis of the trapped target signal strength is as follows:

[0086] Maintain the static observation window of S10 and simultaneously extract the signal strength sequence of the accompanying proxy node and the signal strength sequence of the trapped target;

[0087] The two signal sequences are time-aligned, the signal intensity difference at each sampling time is calculated, and a signal difference sequence is generated.

[0088] Calculate the mean of the signal difference sequence within the static observation window. ;

[0089] Will The absolute value and the system's preset reference strength Normalization is performed to obtain the normalized relative decay rate;

[0090] The standard deviation of the signal difference sequence is calculated to reflect the dynamic instability of the signal difference between the accompanying agent node and the trapped target within the observation window, i.e., the time domain drift.

[0091] The relative attenuation rate and time-domain drift are normalized, and the two are linearly weighted and summed using preset weighting coefficients to obtain the relative dielectric barrier index.

[0092] It should be noted that the relative medium barrier index is positively correlated with the magnitude of the difference ratio and the value of the time domain drift; that is, the greater the difference in signal strength between the trapped target and the accompanying proxy node, and the more drastic the fluctuation of this difference over time, the higher the calculated relative medium barrier index, which indicates that the physical barrier effect between the two is stronger; conversely, if the difference approaches zero and remains highly stable, the lower the relative medium barrier index, which indicates that the two are in an unobstructed homogeneous space.

[0093] The preset weighting coefficients include the first weight. Second weight satisfy );

[0094] For earthquake ruin scenarios, since the inhomogeneity of the medium has a greater impact on signal fluctuations than static attenuation, the preferred setting is... Used for weighted difference ratio (static indicator) setting Used for weighted time-domain drift;

[0095] Among them, the method of assessing the difference in burial depth between the trapped target and the accompanying node based on the medium barrier index, and the vertical stratification strategy for excavation operations is as follows:

[0096] By constructing a hierarchical judgment logic, the numerical range of the relative medium barrier index is mapped to the corresponding physical scenario.

[0097] Preferably, the methods for constructing hierarchical determination logic to determine the vertical hierarchical strategy for mining operations include:

[0098] Same-layer accompanying scenario: If the relative medium barrier index is less than the preset first threshold and the difference ratio approaches zero, it indicates that the trapped target and the accompanying agent node are on the same physical horizontal plane or located in the same unobstructed cavity.

[0099] The strategy for accompanying scenarios on the same layer is as follows:

[0100] The accompanying agent node is marked as a direct symbiotic target, and the search and rescue command is set to dig omnidirectionally with equal depth centered on the accompanying agent node, with the expectation of finding the trapped person at the same time as finding the accompanying agent node;

[0101] In deep burial scenarios: If the relative medium barrier index is greater than the preset second threshold, and the signal strength of the accompanying proxy node is significantly higher than that of the trapped target, it indicates that the accompanying proxy node is located above the trapped target, and there is a strong absorbing medium (such as floor slab or brick wall) between the two.

[0102] The strategy for deep burial scenarios is as follows: mark the accompanying agent node as the upper-level road sign, and set the search and rescue command to guided layered excavation; that is, first quickly excavate and expose the accompanying agent node, and use it as a physical reference point to carry out small-scale demolition operations downward to penetrate the medium layer in order to search for the trapped person below.

[0103] Shallow occlusion scenario: If the relative medium barrier index is greater than the preset second threshold, but the signal strength of the trapped target is significantly higher than that of the accompanying agent node, it indicates that the trapped target is located above the accompanying agent node, or the signal overflow channel at the location of the trapped target is better than that of the accompanying node.

[0104] The strategy for shallow concealment scenarios is as follows: mark the accompanying agent node as the lower reference, set the search and rescue command to be a cautious stripping operation; remind rescuers that before reaching the depth of the accompanying agent node, they should conduct a detailed search in the shallow area in advance to reduce the risk of secondary injury to the trapped people on the upper layer due to the instability of the ruins structure below caused by excessive excavation.

[0105] It should be noted that the relative medium barrier index (Value range [0,1]) Set the following empirical threshold:

[0106] The first threshold is set to 0.2, which corresponds to the upper limit of the system measurement noise floor under free space or slight dust obstruction.

[0107] The second threshold is set to 0.6, which corresponds to the signal attenuation characteristic limit when there is at least one layer of precast concrete slab or brick-concrete structure blocking the signal.

[0108] Example 2:

[0109] Please see Figure 1 As shown, the method for locating disaster victims in scenarios such as earthquakes and landslides includes the following steps:

[0110] S30. Implement a vertical layering strategy to monitor the response sensitivity of the trapped target's signal strength as a function of distance in real time; determine whether the response sensitivity conforms to the free space attenuation law, and establish a rescue channel based on the determination result;

[0111] Among them, such as Figure 2 As shown, the method for implementing a vertical layering strategy and monitoring the response sensitivity of the trapped target's signal strength as a function of distance in real time is as follows:

[0112] In some embodiments, based on the vertical layering strategy of S20, the drone is used as a search and rescue terminal and transformed into a space hovering state;

[0113] S301. Execute the vertical layering strategy and plan the fastest moving trajectory accompanied by the signal enhancement of the agent node to obtain the accompanying target path;

[0114] Preferably, the method for obtaining the accompanying path is as follows:

[0115] Based on the spatial hovering search and rescue terminal, the search and rescue terminal locks the signal of the accompanying agent node. The gradient ascent algorithm is used to calculate the spatial change rate of the accompanying agent node signal strength in real time and automatically plan a moving trajectory that can make the accompanying agent node signal strengthen the fastest, which is defined as the accompanying target path.

[0116] S302. Collect the signal strength of the trapped target and the signal strength of the accompanying proxy node in the accompanying following path, dynamically track the two signal strengths and establish a distance-signal strength curve;

[0117] The preferred method for dynamically tracking and establishing the distance-signal strength curve is as follows:

[0118] During the approach movement, the instantaneous values ​​of the signal strength indication of the trapped target, i.e., the signal strength of the trapped target, and the instantaneous values ​​of the signal strength indication of the accompanying proxy node, i.e., the signal strength of the accompanying proxy node, are collected simultaneously.

[0119] The changes in signal strength of the accompanying agent node and the trapped target were calculated separately, and the displacement data read by the displacement sensor of the rescue terminal were combined to establish a distance-signal strength curve.

[0120] S303. Differentiate the distance-signal strength curve to obtain the two signal growth slopes of the trapped target signal strength and the accompanying agent node signal strength in the accompanying follow path. Perform response sensitivity analysis on the two signal growth slopes to obtain the response sensitivity.

[0121] Preferably, the distance-signal strength curve is differentiated to calculate the signal growth slope of both on the moving path, namely the trapped growth slope and the accompanying growth slope.

[0122] The ratio of the trapped growth slope to the accompanying growth slope is calculated to obtain the response sensitivity, which is then labeled as... ;

[0123] It is understandable that the physical meaning of calculating signal response sensitivity lies in utilizing the Fresnel zone blocking effect of radio wave propagation. When the search and rescue terminal moves towards the target, if there are no high-density obstacles (such as huge concrete beams) in the signal propagation path, the signal of the trapped target should increase significantly logarithmically as the distance decreases (high sensitivity, response sensitivity close to 1); if there are high-density obstacles in the path, most of the electromagnetic energy is absorbed or reflected, and even if the distance is reduced, the signal enhancement will be offset by the blocking loss, resulting in sluggish signal growth or even fluctuations.

[0124] The method for determining whether the response sensitivity conforms to the free space decay law and establishing a rescue channel based on the determination result is as follows:

[0125] By constructing channel impedance evaluation logic, the dynamic characteristics of signal response sensitivity are classified to determine the final rescue and demolition path.

[0126] Preferred, such as Figure 3 As shown, the various ways to determine the possibilities of rescue routes are as follows:

[0127] Set the standard range of signal response sensitivity. Mark the two endpoints of the standard range interval as follows: (Minimum endpoint value) and (Maximum endpoint value)

[0128] when The current accompanying target path is defined as a direct low-resistance channel, and the growth curve of the trapped target signal is highly fitted with the accompanying agent node, which conforms to the free space attenuation law. It is determined that there is no high-density absorber on the connection between the search and rescue terminal and the trapped target, and the current rubble is mainly composed of loose soil and rock or wooden structures with good wave transmission.

[0129] This instructs the rescue team to dig directly along the normal direction of the current approach path, where the operational resistance is minimal and the efficiency is highest.

[0130] when The current accompanying target path is defined as a high-resistance hard obstruction channel. If the accompanying agent node signal increases rapidly, but the trapped target signal increases extremely slowly, and a significant slope hysteresis phenomenon occurs, it is determined that there is a high-density, strong-attenuation hard obstruction (such as a triangular structure formed by the collapse of a load-bearing wall) between the search and rescue terminal and the trapped target. Direct excavation may encounter obstacles that are difficult to break down.

[0131] The rescue team was instructed to stop digging in the current direction and instead circle the accompanying agent node laterally to find an azimuth angle where the response sensitivity would recover as a new entry point, avoiding direct hard obstruction.

[0132] When Negative values ​​or violent oscillations are defined as non-line-of-sight multipath channels. If, during the approach process, the signal strength of the trapped target exhibits a rapid fading phenomenon with fluctuating magnitudes, resulting in violent oscillations or even negative growth in the sensitivity curve, it is determined that the trapped target is located inside a complex metal ruin or steel cage structure, and the signal mainly relies on multiple reflections (multipath effect) to overflow. The current path is not the actual physical shortest path, but rather a virtual path reflected by the signal.

[0133] The rescue team was instructed to use an endoscope or snake-eye life detector to perform insertion verification in the gaps near the accompanying proxy node, avoiding large-scale demolition to prevent the collapse of the ruins from injuring the people inside.

[0134] It should be noted that, if The path was determined to be blocked by general media, indicating the presence of loose debris (such as wooden structures or soil) along the route. The rescue team was instructed to maintain the current direction but reduce the power of mechanical operations. The signal source is determined to be a non-line-of-sight multipath channel, indicating that it is a reflected virtual image. Large-scale demolition is prohibited, and an endoscope must be used to verify the gap.

[0135] S40. Based on the rescue channel strategy, perform circumferential micro-motion scanning with the accompanying agent node as the reference axis to extract the energy divergence of the trapped target signal relative to the accompanying agent node signal; identify the azimuth sector with the strongest signal energy coupling based on the energy divergence and establish a relative azimuth vector.

[0136] Among them, based on the rescue channel strategy, the method of performing circumferential micro-motion scanning with the accompanying agent node as the reference axis to extract the energy divergence of the trapped target signal relative to the accompanying agent node signal is as follows:

[0137] The search and rescue terminal maintains the optimal approach distance or lateral detour point determined by S30 and switches to attitude lock-around scanning mode;

[0138] S401. Set the scanning range according to the rescue channel strategy and perform adaptive circumferential micro-motion scanning:

[0139] Preferably, the adaptive circumferential micro-motion scanning method is as follows: based on the rescue channel strategy, the scanning range is set: if S30 determines that it is a direct low-resistance channel, then a narrow sector fine scan is performed (e.g., ), and swing left and right with high density, centered on the current direction of the agent node;

[0140] If the channel is a high-resistivity, hard-blocked channel, then perform a wide-area, large-angle scan (e.g.) It then performs a lateral orbiting flight around the accompanying agent node;

[0141] S402. During the scanning process, the signals of the trapped target and the accompanying proxy node are simultaneously acquired based on the changes in azimuth angle, and the energy gradient is obtained by energy gradient analysis:

[0142] Preferably, during the scanning process, the signal of the trapped target is acquired simultaneously. and accompanying agent node signals With azimuth Changes in data;

[0143] By performing angular domain differentiation on the two signal intensities, the energy gradients of the two signals at different azimuth angles are calculated to obtain the energy gradient of the trapped target signal. and the energy gradient of the accompanying agent node signal :

[0144] For example, the energy gradient is calculated as follows:

[0145] Through the gradient equation system: Obtain the energy gradient of the trapped target signal and the energy gradient of the accompanying agent node signal ;

[0146] S403. Based on the energy gradient, extract the energy gradient divergence between the trapped target and the accompanying agent node at the same azimuth angle:

[0147] The difference in energy gradient between the trapped target and the accompanying agent node at the same azimuth angle is calculated and defined as the energy gradient divergence.

[0148] in, The preset normalized coupling coefficient is determined by the formula: Obtain;

[0149] The formula is the ratio of the maximum magnitude of the energy gradient of the trapped target to the maximum magnitude of the energy gradient of the accompanying surrogate node. By normalizing the coupling coefficient, the gradient magnitude of the accompanying surrogate node is projected onto the energy level of the trapped target to ensure the calculation of divergence. When both have the same weight contribution, the weak signal characteristics are reduced from being masked due to the large difference in signal strength.

[0150] It should be noted that the physical meaning of energy gradient divergence lies in characterizing the anisotropic differences between the trapped target and its accompanying proxy node in the spatial radiation field. If the two are closely adjacent (in the same room), the signal is highly isomorphic due to the obstruction and reflection from the surrounding ruins. Therefore, at the correct azimuth angle, the trend of the signal energy of the two should be consistent with the angle (i.e., gradient synchronization, divergence). (Tends to 0); however, at incorrect azimuth angles, due to the discreteness of the multipath effect, the rates of change of the two will show a significant divergence (large divergence).

[0151] Among them, the method of identifying the azimuth sector with the strongest signal energy coupling based on energy divergence and establishing the relative azimuth vector is as follows:

[0152] Minimize the energy gradient divergence over the entire scan range;

[0153] Valley detection is performed on the energy gradient divergence curve across the entire scan range;

[0154] Set adaptive convergence threshold (For example: That is, taking the bottom 20% of the dynamic range). and These are the minimum and maximum values ​​of the energy gradient divergence in the energy gradient divergence curve, respectively.

[0155] Selecting the energy gradient divergence value Continuously below the adaptive convergence threshold And the angular span is greater than the minimum effective sector width (e.g. The continuous angular intervals of ) are marked as the sectors with the strongest signal energy coupling;

[0156] Within the sector with the strongest signal energy coupling, calculate the peak value of the second derivative of the trapped target's signal intensity (i.e., the signal energy ridge), and then determine the angle corresponding to the peak value. Locked to relative azimuth vector.

[0157] Example 3:

[0158] Please see Figure 1 As shown, the method for locating disaster victims in scenarios such as earthquakes and landslides includes the following steps:

[0159] S50. Based on the phase azimuth vector combined with the vertical layering strategy, formulate rescue instructions and perform spatiotemporal uncertainty analysis of the trapped target signal to obtain the signal uncertainty and divide the excavation operation area; at the same time, monitor the convergence of the signal uncertainty and dynamically shrink the boundary of the operation area.

[0160] Among them, the method for formulating rescue instructions based on the phase azimuth vector combined with the vertical layering strategy is as follows:

[0161] By combining the vertical layering strategy and the phase displacement vector, a rescue command is output;

[0162] For example, the way to output rescue instructions is as follows: if the vertical layering strategy outputs a dig 1 meter downwards, the relative azimuth vector outputs the azimuth angle. Using the accompanying agent node as the reference point, along the azimuth angle Direction, in vertical depth Demolition work was carried out at the site.

[0163] The spatiotemporal uncertainty analysis of the trapped target signal is performed to obtain the signal uncertainty in the following way:

[0164] During the excavation operation to carry out the rescue order, a follow-up observation window was established to synchronously collect the fluctuation amplitude of the trapped target signal in the time dimension and the angular drift in the spatial dimension.

[0165] Second-order statistical operations are performed on the fluctuation amplitude to obtain the time-domain discrete variance, which is used to characterize the degree of random disturbance of the signal caused by excavation vibration and medium inhomogeneity.

[0166] Gradient analysis of the angular drift is performed to obtain the spatial divergence entropy, which is used to characterize the multipath scattering complexity of the signal transmission path in the current ruin structure.

[0167] The time-domain discrete variance and spatial-domain divergence entropy are made dimensionless, and the dimensionless time-domain discrete variance and spatial-domain divergence entropy are summed to obtain the signal uncertainty.

[0168] The method for analyzing the angle drift is as follows: perform a first-order difference operation on the angle drift within the tracking observation window to obtain an angle gradient sequence reflecting the degree of angle change. Statistically analyze the probability distribution of each numerical interval in this gradient sequence. Calculate the spatial divergence entropy based on Shannon's entropy theorem. If the entropy value is high, it indicates that the signal wave direction is chaotic and disordered (strong multipath effect); if the entropy value is low, it indicates that the signal direction is singular and stable (a direct path exists).

[0169] It should be noted that the physical meaning of signal uncertainty is: to quantify the unreliability of wireless signals into the error radius of physical space. The greater the signal uncertainty, the more severe the signal distortion in the current environment, the greater the potential deviation between the calculated rescue target point and the actual trapped location, and the higher the risk of accidentally injuring the trapped person during the excavation operation.

[0170] The method of simultaneously monitoring the convergence of signal uncertainty and dynamically shrinking the boundary of the working area is as follows:

[0171] S501. Determine the physical radius based on the signal uncertainty, and divide the area into hierarchical regions based on the physical radius to obtain the work areas for manual and mechanical construction:

[0172] Using the target point determined by the rescue order as the geometric center, the signal uncertainty is mapped to the physical radius;

[0173] It should be noted that the signal uncertainty U (range of values) The specific method for mapping to the physical radius R is to use a linear interpolation model, as shown in the following formula: in: The inherent error radius of the device (e.g., 0.5 meters) represents the minimum positioning accuracy of the search and rescue terminal in an ideal, interference-free environment. The maximum safety warning radius (e.g., 5 meters) represents the maximum restricted area boundary for mechanical operations that must be set to prevent accidental injury in extremely chaotic signal environments.

[0174] Preferably, the method of dividing the tiered areas is as follows: the area outside the physical radius is designated as the mechanical operation zone, allowing large rescue machinery to carry out efficient stripping operations;

[0175] The area within the physical radius is designated as a manual precision operation zone, where only manual tools are allowed for low-disturbance excavation to prevent the robotic arm from blindly operating within the error range and causing secondary damage.

[0176] S502. Based on the work area, monitor the convergence of signal uncertainty and dynamically adjust the boundary of the work area:

[0177] Preferably, the method for monitoring the convergence of signal uncertainty and dynamically adjusting the boundary of the work area is as follows:

[0178] As the excavation depth increases and the obstructing medium is removed, the gradient convergence rate of the signal uncertainty over time is calculated in real time.

[0179] If the gradient convergence rate is negative and continuously increases (i.e., the uncertainty decreases rapidly), it indicates that the obstacles are gradually being cleared, the signal path tends to be more direct, the positioning accuracy is significantly improved, the boundary dynamic contraction operation is performed, the physical radius is reduced proportionally, the range of the mechanical operation area is gradually expanded, and the rescue efficiency is improved.

[0180] If the gradient convergence rate is positive or turns from negative to positive (i.e. uncertainty bounce), it indicates that the excavation may have disturbed the unstable metal structure or caused the multipath effect to intensify. In this case, a fusion-type boundary expansion operation is performed to immediately increase the value of the physical radius, move the mechanical operation area back, expand the buffer range of the manual fine operation area, and improve the safety of operation in extreme environments.

[0181] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A method for locating disaster victims in scenarios such as earthquakes and landslides, characterized in that, Includes the following steps: Collect time-varying data of wireless signals in the disaster area, extract the signal fluctuation coefficient of the trapped target signal based on the time-varying data, and identify the accompanying agent node with the same signal change trend of the trapped target based on the signal fluctuation coefficient; For the accompanying agent node, signal blocking analysis is performed in conjunction with the signal strength of the trapped target to obtain the medium blocking index. The difference in burial depth between the trapped target and the accompanying node is assessed based on the medium barrier index, and a vertical stratification strategy for excavation operations is formulated. Implement a vertical stratification strategy to monitor the response sensitivity of the trapped target's signal strength as a function of distance; determine whether the response sensitivity conforms to the free space attenuation law, and establish a rescue channel based on the determination results; Based on the rescue channel strategy, a circumferential micro-motion scan is performed with the accompanying agent node as the reference axis to extract the energy divergence of the trapped target signal relative to the accompanying agent node signal. Based on energy divergence, the azimuth sector with the strongest signal energy coupling is identified, and a relative azimuth vector is established.

2. The method for locating disaster victims based on disaster scenarios such as earthquakes and landslides as described in claim 1, characterized in that: The method for identifying the accompanying agent node is as follows: The accompanying judgment logic is constructed based on the signal fluctuation coefficient, and high-confidence candidate object nodes are selected from all scanned environmental nodes, and a candidate accompanying set is established. Perform an adjoint validity check on the candidate object nodes in the candidate adjoint set to determine the adjoint proxy nodes of the environment.

3. The method for locating disaster victims based on disaster scenarios such as earthquakes and landslides as described in claim 2, characterized in that: The method for obtaining the signal fluctuation coefficient is as follows: Extract the instantaneous value of the received signal strength indication from the transmitter, continuously record the instantaneous values ​​of the received signal strength indication corresponding to the same media access control address, and construct a trapped target signal sequence of length T and N environmental node signal sequences; The DC component is removed from any environmental node signal sequence of the acquired trapped target signal sequence; Correlation analysis is performed on the signal sequence of any environmental node after removing the DC component from the trapped target signal sequence to obtain the signal fluctuation coefficient.

4. The method for locating disaster victims based on disaster scenarios such as earthquakes and landslides as described in claim 1, characterized in that: The method for performing the signal blocking analysis is as follows: Extract the signal strength sequence of the accompanying proxy node and the signal strength sequence of the trapped target; The two signal sequences are time-aligned, the signal intensity difference at each sampling time is calculated, and a signal difference sequence is generated. Calculate the mean of the signal difference sequence within the static observation window. ; Will The absolute value and the system's preset reference strength Normalization is performed to obtain the normalized relative decay rate; Calculate the standard deviation of the signal difference sequence to obtain the time-domain drift; The relative attenuation rate and time-domain drift are normalized, and the two are linearly weighted and summed using preset weighting coefficients to obtain the relative dielectric barrier index.

5. The method for locating disaster victims based on disaster scenarios such as earthquakes and landslides as described in claim 1, characterized in that: The method for obtaining the response sensitivity is as follows: A vertical hierarchical strategy is implemented, and the fastest moving trajectory accompanied by the signal enhancement of the agent node is planned to obtain the accompanying target path; The signal strength of the trapped target and the signal strength of the accompanying agent node in the following path are collected, and the two signal strengths are dynamically tracked and a distance-signal strength curve is established. Differential processing is performed on the distance-signal strength curve to obtain the two signal growth slopes of the trapped target signal strength and the accompanying agent node signal strength on the accompanying follow path. Response sensitivity analysis is performed on the two signal growth slopes to obtain the response sensitivity.

6. The method for locating disaster victims based on disaster scenarios such as earthquakes and landslides as described in claim 5, characterized in that: The response sensitivity analysis is performed as follows: Differentiate the distance-signal strength curves and calculate the signal growth slopes of both along the movement path, namely the trapped growth slope and the accompanying growth slope. The response sensitivity is obtained by calculating the ratio of the trapped growth slope to the accompanying growth slope.

7. The method for locating disaster victims based on disaster scenarios such as earthquakes and landslides as described in claim 1, characterized in that: The energy divergence is extracted as follows: The scanning range was set according to the rescue channel strategy, and an adaptive circumferential micro-motion scan was performed: During the scanning process, the signals of the trapped target and the accompanying agent node are simultaneously acquired, and the energy gradient is obtained by energy gradient analysis. Based on the energy gradient, the energy gradient divergence between the trapped target and the accompanying agent node at the same azimuth angle is extracted.

8. The method for locating disaster victims based on disaster scenarios such as earthquakes and landslides as described in claim 1, characterized in that: Based on the phase azimuth vector combined with the vertical layering strategy, rescue instructions are formulated and the spatiotemporal uncertainty analysis of the trapped target signal is performed to obtain the signal uncertainty and divide the excavation operation area; at the same time, the convergence degree of the signal uncertainty is monitored and the boundary of the operation area is dynamically shrunken.

9. The method for locating disaster victims based on disaster scenarios such as earthquakes and landslides as described in claim 8, characterized in that: The spatiotemporal uncertainty analysis is performed as follows: Establish a follow-up observation window to simultaneously collect the fluctuation amplitude of the trapped target signal in the time dimension and the angular drift in the spatial dimension; Second-order statistical operations are performed on the fluctuation amplitude to obtain the discrete variance in the time domain. Gradient analysis of the angle drift is performed to obtain the spatial divergence entropy; The signal uncertainty is obtained by summing the discrete variance in the time domain and the divergent entropy in the spatial domain after making them dimensionless.

10. The method for locating disaster victims based on disaster scenarios such as earthquakes and landslides as described in claim 8, characterized in that: The convergence degree is monitored by calculating the gradient convergence rate of the signal uncertainty over time, which characterizes the convergence degree of the signal uncertainty.

Images