A method and system for joint search and rescue of a seismic search and rescue robot based on swarm intelligence

Abstract

The present application relates to the technical field of intelligent earthquake rescue, in particular to a method and system for earthquake search and rescue robot joint search and rescue based on swarm intelligence, which comprises: constructing a distributed communication network and a role dynamic allocation mechanism of the earthquake search and rescue robot cluster; using an improved particle swarm optimization algorithm to realize global path planning and regional collaborative coverage of the cluster; completing life signal and dangerous area identification through multi-robot heterogeneous sensing data fusion; sharing search and rescue information and decision-making collaboration based on group consensus mechanism; introducing a feedback regulation mechanism to dynamically optimize the cluster search and rescue strategy. Compared with the technical problems of low single machine operation efficiency, limited coverage range and serious information island in the prior art, the present application realizes the collaborative closed loop of the earthquake search and rescue robot cluster through swarm intelligence, and improves the search and rescue efficiency, coverage integrity and target recognition accuracy in the earthquake debris environment.

Description

Technical Field

[0001] This invention relates to the field of intelligent earthquake rescue technology, specifically to a collaborative search and rescue method and system for earthquake search and rescue robots based on swarm intelligence. Background Technology

[0002] Following an earthquake, the rubble environment is complex, with numerous obstacles, unstable structures, and disrupted communications. Traditional manual search and rescue methods suffer from low efficiency, high risk, and limited coverage. Earthquake search and rescue robots have become an important auxiliary tool. However, existing technologies have significant shortcomings: The limitations of single-machine operation are significant: most search and rescue robots work as stand-alone machines, lacking a collaborative mechanism, making it difficult to cover large areas of rubble. Furthermore, the types of sensors in a single robot are limited, making it impossible to simultaneously meet the multi-dimensional needs of terrain exploration, life detection, and hazard identification. Lack of coordination in path planning: Existing path planning is mostly aimed at individual robots and does not consider the overall coverage efficiency of the cluster. This can easily lead to multiple robots repeatedly searching and rescuing or missing blind spots, and it is difficult to dynamically adapt to changes in the terrain of the ruins, such as secondary collapses. The problem of information silos is prominent: the perception data collected by each robot is stored and processed independently, and global sharing cannot be achieved. This leads to repeated judgment of life signals or failure to synchronize danger information in a timely manner in the same area, affecting search and rescue decisions. Poor dynamic adaptability: It lacks a feedback optimization mechanism for the search and rescue process. When the robot malfunctions, communication is interrupted, or the environment of the target area changes, it cannot adjust the division of roles and operation strategies in a timely manner, resulting in a decrease in search and rescue efficiency.

[0003] Therefore, there is an urgent need for a collaborative search and rescue method based on swarm intelligence, which can solve the problems of inefficiency, insufficient coverage, information isolation and poor adaptability of existing technologies by using the collaborative communication, planning, perception and decision-making of robot swarms, thereby improving the efficiency and reliability of earthquake rubble search and rescue. Summary of the Invention

[0004] To address the aforementioned technical shortcomings, this invention provides a collaborative search and rescue method and system for earthquake search and rescue robots based on swarm intelligence, thereby achieving high rescue efficiency, high coverage integrity, and high target recognition accuracy for earthquake search and rescue robots in earthquake rubble environments.

[0005] This invention is achieved through the following technical solution: A collaborative search and rescue method for earthquake search and rescue robots based on swarm intelligence is provided, the method comprising the following steps: Step S10: Build a distributed mesh communication network for the earthquake search and rescue robot cluster. Based on the requirements of earthquake rubble search and rescue missions, use fuzzy hierarchical analysis to dynamically allocate the roles of earthquake search and rescue robots and establish a basic framework for cluster collaborative operation. Step S20: Based on the basic framework of cluster collaborative operation, an improved particle swarm optimization algorithm is adopted, and terrain fitness weight and regional coverage constraints are introduced to perform global path planning and regional collaborative coverage of the earthquake search and rescue robot cluster, and generate the initial operation path of each earthquake search and rescue robot. Step S30: Each earthquake search and rescue robot collects data on the ruins environment and life signals through heterogeneous sensors. The heterogeneous sensors include exploratory type with lidar and panoramic camera, perception type with life detector and gas sensor, and rescue-assisted type with stress sensor. Federated learning is used for distributed data fusion to remove redundant and abnormal data and identify the location of survivors and dangerous areas, such as flammable gas leaks and secondary collapse risk areas. Step S40: Based on the identified survivor locations and danger zones, real-time sharing of search and rescue information within the cluster is achieved through a group consensus mechanism, and the operational priorities and regional assignments of each robot are dynamically adjusted. Step S50: Based on the dynamic adjustment results, provide feedback to optimize the path planning parameters and the role assignment of the earthquake search and rescue robot until the search and rescue of the entire rubble area is completed, and output a complete report including the location of survivors, the distribution of dangerous areas and the search and rescue trajectory.

[0006] Preferably, the step of building a distributed mesh communication network for the earthquake search and rescue robot cluster in step S10 includes: An industrial-grade wireless mesh module is used to build a self-organizing communication network. Each robot acts as a communication node, supporting multi-hop transmission. The communication distance is no less than 50m, and the anti-interference capability meets the GB / T30263-2013 standard. A fuzzy hierarchical analysis model was established, using exploration efficiency, perception accuracy, and load capacity as evaluation indicators. Exploration efficiency includes movement speed and terrain adaptability; perception accuracy includes sensor type and detection range; and load capacity includes the load-bearing capacity of the robotic arm and endurance. Role matching was performed on the earthquake search and rescue robot cluster. Among them, exploration robots, equipped with LiDAR and panoramic cameras, focus on rapid terrain mapping and path opening, emphasizing movement and terrain adaptability, accounting for 40%-50%. Perception robots integrate life detectors and gas sensors. The life detector is used to detect heart rate and respiration, and the body sensor is used to detect combustible / toxic gases, focusing on target recognition, accounting for 30%-40%. Rescue assistance robots are equipped with robotic arms and stress sensors, focusing on clearing dangerous areas and assisting survivor rescue, accounting for 10%-20%. A dynamic role adjustment mechanism allows nearby exploratory robots to automatically switch to the role of communication relay to ensure network connectivity when the communication signal in a certain area is substandard.

[0007] Preferably, step S20 employs an improved particle swarm optimization algorithm, introducing terrain fitness weights and regional coverage constraints, to perform global path planning and regional collaborative coverage for the earthquake search and rescue robot swarm. The rubble search and rescue area is gridded with a grid size of 0.5m × 0.5m. The particle position is defined as the working path node of the earthquake search and rescue robot, and the particle velocity is the path adjustment direction. The fitness function is improved as F = ω1 × C + ω2 × D + ω3 × S, where C is the area coverage, D is the path distance cost, S is the terrain safety factor, and ω1, ω2, and ω3 are weighting coefficients, with the sum of ω1, ω2, and ω3 being 1. A regional division-coordinated coverage strategy is introduced, dividing the ruins into multiple sub-regions. Each sub-region is assigned 3-5 earthquake search and rescue robot teams. Within each team, the local path is optimized using a particle swarm optimization algorithm, and the coverage progress is synchronized between teams through a communication network to avoid duplicate coverage and blind spots.

[0008] Preferably, step S30, which employs federated learning for distributed data fusion, removes redundant and outlier data, and identifies survivor locations and danger zones, includes: Each earthquake search and rescue robot preprocesses the collected data locally, uses voxel filtering to denoise the lidar data, uses wavelet transform to enhance the life detector signal, and calibrates the gas sensor data based on ambient temperature compensation. A federated learning model is constructed, in which each earthquake search and rescue robot acts as a local node to train a local model. A lightweight CNN network is used, and only the model parameters are uploaded to the cluster coordination node. The coordination node aggregates the parameters to generate a global model, which is then distributed to each node for updates. The criteria for target identification and judgment are as follows: survivors are identified by vital signs, which are set as heart rate of 10-200 beats / minute and respiratory rate of 10-30 breaths / minute; hazardous areas are identified by gas concentration or structural stress, and the criteria for hazardous areas include: gas concentration of combustible gas > 10% of the lower explosive limit, toxic gas > TLV-TWA value, and structural stress of structural stress > 80% of the material yield strength.

[0009] Preferably, step S40, which involves real-time sharing of search and rescue information within the cluster through a group consensus mechanism and dynamic adjustment of the operational priorities and regional divisions of each robot, includes: Using the BFT Byzantine Fault Tolerance algorithm, when more than 2 / 3 of the earthquake search and rescue robot nodes verify that a certain search and rescue information is consistent, it is determined to be valid information and written into the cluster shared database; Establish collaborative decision-making rules. When a life signal is detected in a certain area, the priority is set to the highest, and the nearest 3-4 earthquake search and rescue robots are dispatched to verify and prepare for rescue. At least one of them should be a perception type and one should be a rescue assistance type. When a collapse risk is detected on a certain path, it should be marked as prohibited and synchronized to all earthquake search and rescue robots to replan the detour path. Real-time operational status monitoring: The system collects the battery level and load of each earthquake search and rescue robot through the communication network. When the battery level is below 20%, it triggers a return to base for charging. When the robotic arm load is above 80%, it triggers an assistance request and dynamically adjusts task allocation.

[0010] Preferably, step S50, which involves feeding back and optimizing path planning parameters and assigning roles to the earthquake search and rescue robot based on the dynamic adjustment results, includes: Regularly monitor the progress of search and rescue coverage. When the coverage of a certain sub-area is less than 80%, analyze the reasons and determine whether it is due to obstacles or robot malfunction. If it is due to obstacles, dispatch additional rescue-assisted robots with robotic arms to clear them. If it is due to robot malfunction, dispatch backup earthquake search and rescue robots to fill the gap. Optimize path planning parameters. When the earthquake search and rescue robot frequently avoids obstacles on a certain path segment, increase the weight of the terrain safety factor in the fitness function from ω3 to 0.3. When the coverage efficiency does not meet the requirements, increase the weight of the regional coverage from ω1 to 0.6. The search and rescue operation is terminated when the rubble coverage reaches more than 95% and no new life signals are detected for 10 consecutive minutes. A report containing the coordinates of survivors, the distribution of dangerous areas, and the search and rescue trajectories of each earthquake search and rescue robot is generated and submitted to the rescue command center.

[0011] Furthermore, to achieve the above objectives, this invention also proposes a joint search and rescue system for earthquake search and rescue robots based on swarm intelligence, wherein the joint search and rescue system for earthquake search and rescue robots based on swarm intelligence includes: Cluster Communication and Earthquake Search and Rescue Robot Role Assignment Module: This module is used to build a distributed mesh communication network for earthquake search and rescue robot clusters. Based on the requirements of earthquake rubble search and rescue missions, it uses fuzzy hierarchical analysis to dynamically assign earthquake search and rescue robot roles and establish a basic framework for cluster collaborative operations. Global Path and Coverage Planning Module: This module is used as the basic framework for cluster-based collaborative operations. It employs an improved particle swarm optimization algorithm, introduces terrain fitness weights and regional coverage constraints, and performs global path planning and regional collaborative coverage for the earthquake search and rescue robot cluster, generating the initial operational paths for each earthquake search and rescue robot. Heterogeneous data fusion and identification module: This module is used by various earthquake search and rescue robots to collect data on the ruin environment and life signals through heterogeneous sensors. The heterogeneous sensors include exploratory type with lidar and panoramic camera, perception type with life detector and gas sensor, and rescue-assisted type with stress sensor. Federated learning is used for distributed data fusion to remove redundant and abnormal data and identify the location of survivors and dangerous areas, such as flammable gas leaks and secondary collapse risk areas. Group consensus and decision-making coordination module: Based on the identified survivor locations and danger zones, it is used to share search and rescue information within the cluster in real time through a group consensus mechanism, and dynamically adjust the operation priority and regional division of each robot. Feedback optimization and report generation module: Based on the dynamic adjustment results, it provides feedback to optimize path planning parameters and earthquake search and rescue robot role assignments until the search and rescue of the entire rubble area is completed, and outputs a complete report including survivor locations, distribution of dangerous areas and search and rescue trajectories.

[0012] Furthermore, to achieve the above objectives, this invention also proposes a swarm intelligence-based earthquake search and rescue robot joint search and rescue device. The device includes: a memory, a processor, and programs such as a swarm intelligence-based earthquake search and rescue robot joint search and rescue algorithm stored in the memory and executable on the processor. The swarm intelligence-based earthquake search and rescue robot joint search and rescue algorithm and other programs are steps for implementing the swarm intelligence-based earthquake search and rescue robot joint search and rescue method described above.

[0013] Furthermore, to achieve the above objectives, the present invention also provides a computer program product, which includes programs such as a joint search and rescue algorithm for earthquake search and rescue robots based on swarm intelligence. When the joint search and rescue algorithm for earthquake search and rescue robots based on swarm intelligence is executed by a processor, it implements the joint search and rescue method for earthquake search and rescue robots based on swarm intelligence as described above.

[0014] The advantages and effects of this invention are: This invention proposes a collaborative search and rescue method and system for earthquake search and rescue robots based on swarm intelligence. Through distributed communication and role division, the earthquake search and rescue robot swarm can simultaneously cover large areas of rubble, avoiding the limitations of single-machine operation. Compared with traditional single-machine search and rescue, it improves coverage efficiency and shortens the time to find survivors. At the same time, the improved particle swarm optimization algorithm combined with the regional collaborative coverage strategy effectively avoids repeated search and rescue and blind spots, improving rubble coverage. In addition, distributed fusion of heterogeneous data from multiple robots is achieved through federated learning, avoiding information silos, improving the accuracy of life signal recognition, reducing the misjudgment rate of dangerous areas, and the feedback adjustment mechanism can adjust role division and path planning in real time to cope with emergencies such as robot failure and terrain changes. The system's robustness is significantly improved, ensuring that the search and rescue process is not interrupted. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of a collaborative search and rescue method for earthquake search and rescue robots based on swarm intelligence, according to the present invention.

[0017] Figure 2 This is a schematic diagram of a collaborative search and rescue system for earthquake search and rescue robots based on swarm intelligence, according to the present invention.

[0018] Figure 3 This is a schematic block diagram of a collaborative search and rescue electronic device for earthquake search and rescue robots based on swarm intelligence, according to the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] like Figure 1 As shown, in one embodiment of the present invention, a collaborative search and rescue method for earthquake search and rescue robots based on swarm intelligence includes the following steps: Step S10: Build a distributed mesh communication network for the earthquake search and rescue robot cluster. Based on the requirements of earthquake rubble search and rescue missions, use fuzzy hierarchical analysis to dynamically allocate roles for earthquake search and rescue robots and establish a basic framework for cluster collaborative operations.

[0021] Specifically, step S10, which involves building a distributed mesh communication network for the earthquake search and rescue robot cluster, includes the following steps: An industrial-grade wireless mesh module is used to build a self-organizing communication network. Each robot acts as a communication node, supporting multi-hop transmission. The communication distance is no less than 50m, and the anti-interference capability meets the GB / T30263-2013 standard. A fuzzy hierarchical analysis model was established, using exploration efficiency, perception accuracy, and load capacity as evaluation indicators. Exploration efficiency includes movement speed and terrain adaptability; perception accuracy includes sensor type and detection range; and load capacity includes the load-bearing capacity of the robotic arm and endurance. Role matching was performed on the earthquake search and rescue robot cluster. Among them, exploration robots, equipped with LiDAR and panoramic cameras, focus on rapid terrain mapping and path opening, emphasizing movement and terrain adaptability, accounting for 40%-50%. Perception robots integrate life detectors and gas sensors. The life detector is used to detect heart rate and respiration, and the body sensor is used to detect combustible / toxic gases, focusing on target recognition, accounting for 30%-40%. Rescue assistance robots are equipped with robotic arms and stress sensors, focusing on clearing dangerous areas and assisting survivor rescue, accounting for 10%-20%. A dynamic role adjustment mechanism allows nearby exploratory robots to automatically switch to the role of communication relay to ensure network connectivity when the communication signal in a certain area is substandard.

[0022] For example, 10 search and rescue robots were selected, including 6 exploratory robots, 3 perception robots, and 1 rescue assistance robot. A distributed mesh communication network was built using Huawei CloudCampusMesh modules. The communication distance was tested to reach 60m. Role allocation was completed using fuzzy hierarchical analysis. Two exploratory robots were preset as backup communication relay nodes.

[0023] Step S20: Based on the basic framework of cluster collaborative operation, an improved particle swarm optimization algorithm is adopted, and terrain fitness weight and regional coverage constraints are introduced to perform global path planning and regional collaborative coverage of the earthquake search and rescue robot cluster, and generate the initial operation path of each earthquake search and rescue robot.

[0024] Specifically, step S20 employs an improved particle swarm optimization algorithm, introducing terrain fitness weights and regional coverage constraints, to perform global path planning and regional collaborative coverage for the earthquake search and rescue robot swarm. The rubble search and rescue area is gridded with a grid size of 0.5m × 0.5m. The particle position is defined as the working path node of the earthquake search and rescue robot, and the particle velocity is the path adjustment direction. The fitness function is improved as F = ω1 × C + ω2 × D + ω3 × S, where C is the area coverage, D is the path distance cost, S is the terrain safety factor, ω1, ω2, and ω3 are weight coefficients, and the sum of ω1, ω2, and ω3 is 1. For example, ω1 = 0.5, ω2 = 0.3, and ω3 = 0.2. The smaller D is, the better. S takes the value [0,1]. A regional division-coordinated coverage strategy is introduced, dividing the ruins into multiple sub-regions. Each sub-region is assigned 3-5 earthquake search and rescue robot teams. Within each team, the local path is optimized using a particle swarm optimization algorithm, and the coverage progress is synchronized between teams through a communication network to avoid duplicate coverage and blind spots.

[0025] For example, the simulated earthquake ruins area is 500m² 2 The terrain includes concrete obstacles and narrow passages, and is rasterized into 1000×1000 grids. The terrain data is collected by an exploratory robot and synchronized to the cluster. An improved particle swarm optimization algorithm is used to divide the ruins into 10 sub-regions with ω1=0.5, ω2=0.3, and ω3=0.2. Each sub-region is assigned 1 exploratory robot and 0-1 perception robots to generate the initial operation path.

[0026] Step S30: Each earthquake search and rescue robot collects environmental data and life signals from the ruins through heterogeneous sensors. The heterogeneous sensors include exploratory type with lidar and panoramic camera, perception type with life detector and gas sensor, and rescue-assisted type with stress sensor. Federated learning is used for distributed data fusion to remove redundant and abnormal data and identify the location of survivors and dangerous areas, such as areas with flammable gas leaks and secondary collapse risk.

[0027] Specifically, step S30, which employs federated learning for distributed data fusion, removes redundant and outlier data, and identifies survivor locations and danger zones, includes the following steps: Each earthquake search and rescue robot preprocesses the collected data locally. The lidar data is denoised using voxel filtering to remove point cloud noise caused by debris dust. The life detector signal is enhanced using wavelet transform to extract weak heart rate and respiratory signals. The gas sensor data is compensated and calibrated based on ambient temperature, which is collected from the robot's temperature sensor to avoid temperature affecting the concentration detection accuracy. The stress sensor data is filtered using moving average to eliminate errors caused by mechanical vibration. A federated learning model is constructed, with each robot acting as a local node. The local model is trained based on the lightweight CNN network MobileNetV3. The input is preprocessed sensor data, and the output is the target recognition result. Only the model parameters, not the raw data, are uploaded to the cluster coordination node, which is the best-performing exploratory robot. The coordination node uses a federated averaging algorithm to aggregate the parameters of each node, generate a global model, and then distribute it to each node for updates, realizing a fusion mode of data not being uploaded and model co-training. The target identification criteria are as follows: survivors are identified by vital signs. A heart rate of 10-200 beats / minute and a respiratory rate of 10-30 breaths / minute are considered survivors, and their coordinates are recorded with an accuracy of ±0.5m. If the gas concentration is greater than 10% of the lower explosive limit (e.g., methane > 0.5%), or greater than the TLV-TWA value (e.g., carbon monoxide > 25 ppm), or if the structural stress is greater than 80% of the material's yield strength (e.g., concrete structural stress > 20 MPa), the area is identified as a danger zone and marked as prohibited / warning.

[0028] For example, each earthquake search and rescue robot collects data along its path, the exploration robot draws terrain maps using lidar, and the perception robot detects two life signals using a life detector: a heart rate of 70 beats / minute and a respiratory rate of 18 breaths / minute, and a heart rate of 85 beats / minute and a respiratory rate of 20 breaths / minute. The gas sensor detects a methane concentration of 0.8%, with the lower explosive limit of 10% being 0.5%. By using federated learning to fuse the data, two survivors and one flammable gas hazard area are identified.

[0029] Step S40: Based on the identified survivor locations and danger zones, real-time sharing of search and rescue information within the cluster is achieved through a group consensus mechanism, and the operational priorities and regional assignments of each robot are dynamically adjusted.

[0030] Specifically, step S40, which involves real-time sharing of search and rescue information within the cluster through a group consensus mechanism and dynamic adjustment of the operational priorities and regional assignments of each robot, includes the following steps: The BFT (Byzantine Fault Tolerance) algorithm is adopted. When a robot detects a survivor or a dangerous area, it broadcasts the information to the cluster through the communication network, including coordinates, detection time, and sensor data. Other earthquake search and rescue robots verify the information. For example, a proximity-sensing robot verifies the life signal. When more than 2 / 3 of the earthquake search and rescue robot nodes verify the information, it is determined to be valid information and written into the cluster's shared database. The data is distributed and stored on each robot node to ensure data redundancy and backup, and to avoid decision-making errors caused by misjudgment by a single robot. Establish collaborative decision-making rules: when a life signal is detected in a certain area, the priority is set to the highest, and the nearest 3-4 earthquake search and rescue robots are dispatched to verify and prepare for rescue, while operations in other low-priority areas are suspended. The robots should include at least one perception type and one rescue assistance type. When more than 3 earthquake search and rescue robots have gathered in a certain area, subsequent robots are dispatched to areas that are not covered or have insufficient coverage to avoid wasting resources. When a collapse risk is detected on a certain path, it is immediately marked as prohibited and synchronized to all earthquake search and rescue robots, and detour routes are replanned. Real-time operational status monitoring: The system collects status data of each earthquake search and rescue robot in real time through the communication network, including battery level, load, and sensor operating status. When the battery level of an earthquake search and rescue robot is below 20%, a return-to-base charging command is triggered. When the load of the robotic arm of the rescue assistance robot is above 80%, an assistance request is triggered, and nearby rescue assistance robots are dispatched to provide support, dynamically adjusting task allocation.

[0031] For example, by verifying information using the Byzantine fault tolerance algorithm, if the verification is consistent across 8 robots, the survivor area is set to the highest priority. The two nearest perception robots are then dispatched to verify the signal, including one that was originally responsible for the area and one that was transferred from a neighboring area. One rescue assistance robot is sent to the danger zone to mark it as off-limits, and the path planning permissions of all robots in that area are simultaneously disabled.

[0032] Step S50: Based on the dynamic adjustment results, provide feedback to optimize the path planning parameters and the role assignment of the earthquake search and rescue robot until the search and rescue of the entire rubble area is completed, and output a complete report including the location of survivors, the distribution of dangerous areas and the search and rescue trajectory.

[0033] Specifically, step S50, which involves feeding back and optimizing path planning parameters and assigning roles to the earthquake search and rescue robot based on the dynamic adjustment results, includes: Regularly analyze the progress of search and rescue coverage. For example, every 5 minutes, when the coverage of a certain sub-area is less than 80%, analyze the reasons and determine whether it is due to obstacles or robot malfunction. If it is due to obstacles, dispatch additional rescue-assisted robots with robotic arms to clear them. If it is due to robot malfunction, dispatch backup earthquake search and rescue robots to fill the gap. Optimize path planning parameters. When the earthquake search and rescue robot frequently avoids obstacles on a certain path segment, increase the weight of the terrain safety factor in the fitness function from ω3 to 0.3. When the coverage efficiency does not meet the requirements, increase the weight of the regional coverage from ω1 to 0.6. Role allocation optimization: When there is a high density of life signals in a certain area, such as 100m 2 If more than three survivors are detected, it indicates that the area has a high demand for sensing. 2-3 sensing robots will be dispatched from other low-demand areas to provide support and improve the accuracy of signal verification. When dangerous areas are concentrated in a certain area, additional rescue assistance robots will be dispatched to mark the restricted areas to prevent other robots from entering by mistake. The search and rescue operation is terminated when the rubble coverage reaches over 95% and no new life signals are detected for 10 consecutive minutes. The operation is then considered complete, and a full report is output, including the location of survivors, the distribution of dangerous areas, and the search and rescue trajectory. The report includes the coordinates of the survivors' locations, the strength of life signals, a map of the dangerous areas, risk levels, the search and rescue trajectories of each robot, the time-coordinate correspondence, equipment status statistics, the number of participating robots, the number of malfunctions, and their remaining battery life. The report is submitted to the rescue command center via the communication network to provide precise guidance for manual rescue efforts.

[0034] For example, the coverage progress is tallied every 5 minutes. If a sub-area is found to have only 70% coverage due to obstacles, additional rescue-assist robots are dispatched to clear the obstacles, and the path planning ω3 for that area is adjusted to 0.3. After 1 hour, the coverage of the entire ruins area reaches 96%, and no new life signals are detected for 10 consecutive minutes. A report is output, showing the locations of 2 survivors at coordinates X1=25.3m, Y1=18.7m and X2=38.5m, Y2=42.1m, and 1 danger zone at X3=12.8m, Y3=30.5m. The trajectories of all earthquake search and rescue robots are complete and traceable.

[0035] In addition, such as Figure 2 As shown, in one embodiment of the present invention, a joint search and rescue system for earthquake search and rescue robots based on swarm intelligence is proposed. The system includes: Cluster Communication and Earthquake Search and Rescue Robot Role Assignment Module: This module is used to build a distributed mesh communication network for earthquake search and rescue robot clusters. Based on the requirements of earthquake rubble search and rescue missions, it uses fuzzy hierarchical analysis to dynamically assign earthquake search and rescue robot roles and establish a basic framework for cluster collaborative operations. Global Path and Coverage Planning Module: This module is used as the basic framework for cluster-based collaborative operations. It employs an improved particle swarm optimization algorithm, introduces terrain fitness weights and regional coverage constraints, and performs global path planning and regional collaborative coverage for the earthquake search and rescue robot cluster, generating the initial operational paths for each earthquake search and rescue robot. Heterogeneous data fusion and identification module: This module is used by various earthquake search and rescue robots to collect data on the ruin environment and life signals through heterogeneous sensors. The heterogeneous sensors include exploratory type with lidar and panoramic camera, perception type with life detector and gas sensor, and rescue-assisted type with stress sensor. Federated learning is used for distributed data fusion to remove redundant and abnormal data and identify the location of survivors and dangerous areas, such as flammable gas leaks and secondary collapse risk areas. Group consensus and decision-making coordination module: Based on the identified survivor locations and danger zones, it is used to share search and rescue information within the cluster in real time through a group consensus mechanism, and dynamically adjust the operation priority and regional division of each robot. Feedback optimization and report generation module: Based on the dynamic adjustment results, it provides feedback to optimize path planning parameters and earthquake search and rescue robot role assignments until the search and rescue of the entire rubble area is completed, and outputs a complete report including survivor locations, distribution of dangerous areas and search and rescue trajectories.

[0036] This application provides a swarm intelligence-based earthquake search and rescue robot joint search and rescue system, employing a swarm intelligence-based earthquake search and rescue robot joint search and rescue method described in the above embodiments. This system addresses the technical problems of low efficiency in single-machine operation, poor information sharing, and insufficient dynamic adaptability in existing technologies. Compared with existing technologies, the beneficial effects of the swarm intelligence-based earthquake search and rescue robot joint search and rescue system provided in this application are the same as those of the swarm intelligence-based earthquake search and rescue robot joint search and rescue method described in the above embodiments. Furthermore, other technical features of the swarm intelligence-based earthquake search and rescue robot joint search and rescue system are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0037] This application provides a swarm intelligence-based earthquake search and rescue robot joint search and rescue device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to execute the swarm intelligence-based earthquake search and rescue robot joint search and rescue method in Embodiment 1 above.

[0038] like Figure 3 As shown in the illustration, in one embodiment of the present invention, a structural schematic diagram of a swarm intelligence-based earthquake search and rescue robot joint search and rescue device suitable for implementing the embodiments of this application is presented. The swarm intelligence-based earthquake search and rescue robot joint search and rescue device in this application embodiment may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), etc., as well as fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The earthquake search and rescue robot joint search and rescue device based on swarm intelligence shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0039] Figure 3The illustrated swarm intelligence-based earthquake search and rescue robot joint search and rescue device may include a processor 1001 (e.g., a central processing unit, graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a machine-readable storage medium (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the swarm intelligence-based earthquake search and rescue robot joint search and rescue device. The processor 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and a communication unit 1009. Communication unit 1009 allows a swarm intelligence-based earthquake search and rescue robot joint search and rescue device to communicate wirelessly or wiredly with other devices to exchange data. Although the figure shows a swarm intelligence-based earthquake search and rescue robot joint search and rescue device with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0040] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication system, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processor 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0041] This application provides a swarm intelligence-based earthquake search and rescue robot joint search and rescue device, which employs a swarm intelligence-based earthquake search and rescue robot joint search and rescue method described in the above embodiments. This addresses the technical problems of low efficiency in single-machine operation, poor information sharing, and insufficient dynamic adaptability in existing technologies. Compared with existing technologies, the beneficial effects of the swarm intelligence-based earthquake search and rescue robot joint search and rescue device provided in this application are the same as those of the swarm intelligence-based earthquake search and rescue robot joint search and rescue method described in the above embodiments. Furthermore, other technical features of this swarm intelligence-based earthquake search and rescue robot joint search and rescue device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0042] The various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0043] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described earthquake search and rescue robot joint search and rescue method based on swarm intelligence.

[0044] The computer program product provided in this application can solve the technical problems of low efficiency of single-machine operation, poor information sharing, and insufficient dynamic adaptability in the prior art. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the earthquake search and rescue robot joint search and rescue method based on swarm intelligence provided in the above embodiments, and will not be repeated here.

[0045] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A swarm intelligence-based earthquake search and rescue robot joint search and rescue method, characterized in that, The method includes the following steps: Step S10: Build a distributed mesh communication network for the earthquake search and rescue robot cluster. Based on the requirements of earthquake rubble search and rescue missions, use fuzzy hierarchical analysis to dynamically allocate the roles of earthquake search and rescue robots and establish a basic framework for cluster collaborative operation. Step S20: Based on the basic framework of cluster collaborative operation, an improved particle swarm optimization algorithm is adopted, and terrain fitness weight and regional coverage constraints are introduced to perform global path planning and regional collaborative coverage of the earthquake search and rescue robot cluster, and generate the initial operation path of each earthquake search and rescue robot. Step S30: Each earthquake search and rescue robot collects environmental data and life signals from the ruins through heterogeneous sensors, uses federated learning to perform distributed data fusion, removes redundant and abnormal data, and identifies the location of survivors and dangerous areas. Step S40: Based on the identified survivor locations and danger zones, real-time sharing of search and rescue information within the cluster is achieved through a group consensus mechanism, and the operational priorities and regional assignments of each robot are dynamically adjusted. Step S50: Based on the dynamic adjustment results, provide feedback to optimize the path planning parameters and the role allocation of the earthquake search and rescue robot until the search and rescue of the entire rubble area is completed, and output a complete report including the location of survivors, the distribution of dangerous areas and the search and rescue trajectory; The steps in step S20, which employ an improved particle swarm optimization algorithm and introduce terrain fitness weights and regional coverage constraints, to perform global path planning and regional collaborative coverage for the earthquake search and rescue robot swarm, include: The rubble search and rescue area is rasterized, and the particle position is defined as the operation path node of the earthquake search and rescue robot, and the particle velocity is the path adjustment direction. The fitness function is improved as F = ω1 × C + ω2 × D + ω3 × S, where C is the area coverage, D is the path distance cost, S is the terrain safety factor, and ω1, ω2, and ω3 are weighting coefficients, with the sum of ω1, ω2, and ω3 being 1. A regional division-coordinated coverage strategy is introduced, dividing the ruins into multiple sub-regions. Each sub-region is assigned 3-5 earthquake search and rescue robot teams. Within each team, the local path is optimized using a particle swarm optimization algorithm, and the coverage progress is synchronized between teams through a communication network.

2. The swarm intelligence-based earthquake search and rescue robot joint search and rescue method according to claim 1, characterized in that, The steps in step S10 of building a distributed mesh communication network for the earthquake search and rescue robot cluster include: An industrial-grade wireless mesh module is used to build a self-organizing communication network, with each robot acting as a communication node, supporting multi-hop transmission and a communication distance of no less than 50m; A fuzzy hierarchical analysis model was established to perform role matching for an earthquake search and rescue robot cluster, using exploration efficiency, perception accuracy, and load capacity as evaluation indicators. A dynamic role adjustment mechanism allows nearby exploratory robots to automatically switch to the role of communication relay to ensure network connectivity when the communication signal in a certain area is substandard. 3.The earthquake search and rescue robot joint search and rescue method based on swarm intelligence according to claim 1, characterized in that, The steps in step S30, which employ federated learning for distributed data fusion, remove redundant and outlier data, and identify survivor locations and danger zones, include: Each earthquake search and rescue robot preprocesses the collected data locally, uses voxel filtering to denoise the lidar data, uses wavelet transform to enhance the life detector signal, and calibrates the gas sensor data based on ambient temperature compensation. A federated learning model is constructed, in which each earthquake search and rescue robot acts as a local node to train a local model. A lightweight CNN network is used, and only the model parameters are uploaded to the cluster coordination node. The coordination node aggregates the parameters to generate a global model, which is then distributed to each node for updates. The criteria for target identification and judgment include identifying survivors through life signals and identifying dangerous areas through gas concentration or structural stress.

4. The swarm intelligence-based earthquake search and rescue robot joint search and rescue method according to claim 1, characterized in that, The steps in step S40, which involve real-time sharing of search and rescue information within the cluster through a group consensus mechanism and dynamic adjustment of the operational priorities and regional assignments of each robot, include: Using the BFT Byzantine Fault Tolerance algorithm, when more than 2 / 3 of the earthquake search and rescue robot nodes verify that a certain search and rescue information is consistent, it is determined to be valid information and written into the cluster shared database; Establish collaborative decision-making rules. When a life signal is detected in a certain area, the priority is set to the highest, and the nearest 3-4 earthquake search and rescue robots are dispatched to verify and prepare for rescue. When a collapse risk is detected on a certain path, it is immediately marked as prohibited and synchronized to all earthquake search and rescue robots to replan the detour route. Real-time operational status monitoring is implemented, and the power and load of each earthquake search and rescue robot are collected through the communication network to dynamically adjust task allocation.

5. The swarm intelligence-based earthquake search and rescue robot joint search and rescue method according to claim 1, characterized in that, The step S50, which involves feeding back and optimizing path planning parameters and assigning roles to the earthquake search and rescue robot based on the dynamic adjustment results, includes: Regularly monitor the progress of search and rescue coverage. When the coverage of a certain sub-area is less than 80%, analyze the reasons and determine whether it is due to obstacles or robot malfunction. If it is due to obstacles, dispatch additional rescue-assisted robots with robotic arms to clear them. If it is due to robot malfunction, dispatch backup earthquake search and rescue robots to fill the gap. Optimize path planning parameters. When the earthquake search and rescue robot frequently avoids obstacles on a certain path segment, increase the weight of the terrain safety factor in the fitness function from ω3 to 0.

3. When the coverage efficiency does not meet the requirements, increase the weight of the regional coverage from ω1 to 0.

6. The search and rescue operation is terminated when the rubble coverage reaches more than 95% and no new life signals are detected for 10 consecutive minutes. A report containing the coordinates of survivors, the distribution of dangerous areas, and the search and rescue trajectories of each earthquake search and rescue robot is generated and submitted to the rescue command center.

6. A swarm intelligence based earthquake search and rescue robot joint search and rescue system, executing the method of any one of claims 1 to 5, characterized in that, include: Cluster Communication and Earthquake Search and Rescue Robot Role Assignment Module: This module is used to build a distributed mesh communication network for earthquake search and rescue robot clusters. Based on the requirements of earthquake rubble search and rescue missions, it uses fuzzy hierarchical analysis to dynamically assign earthquake search and rescue robot roles and establish a basic framework for cluster collaborative operations. Global Path and Coverage Planning Module: This module is used as the basic framework for cluster-based collaborative operations. It employs an improved particle swarm optimization algorithm, introduces terrain fitness weights and regional coverage constraints, and performs global path planning and regional collaborative coverage for the earthquake search and rescue robot cluster, generating the initial operational paths for each earthquake search and rescue robot. Heterogeneous data fusion and identification module: This module is used by various earthquake search and rescue robots to collect data on the ruin environment and life signals through heterogeneous sensors, and uses federated learning to perform distributed data fusion, remove redundant and abnormal data, and identify the location of survivors and dangerous areas. Group consensus and decision-making coordination module: Based on the identified survivor locations and danger zones, it is used to share search and rescue information within the cluster in real time through a group consensus mechanism, and dynamically adjust the operation priority and regional division of each robot. Feedback optimization and report generation module: Based on the dynamic adjustment results, it provides feedback to optimize path planning parameters and earthquake search and rescue robot role assignments until the search and rescue of the entire rubble area is completed, and outputs a complete report including survivor locations, distribution of dangerous areas and search and rescue trajectories.

7. A swarm intelligence based earthquake search and rescue robot joint search and rescue equipment, characterized in that, include: The system includes a memory, a processor, and a swarm intelligence-based earthquake search and rescue robot joint search and rescue program stored in the memory and executable on the processor. When the swarm intelligence-based earthquake search and rescue robot joint search and rescue program is executed by the processor, it implements a swarm intelligence-based earthquake search and rescue robot joint search and rescue method as described in any one of claims 1 to 5.

8. A computer program product, characterised in that, The computer program product includes a swarm intelligence-based earthquake search and rescue robot joint search and rescue program, which, when executed by a processor, implements a swarm intelligence-based earthquake search and rescue robot joint search and rescue method as described in any one of claims 1 to 5.

Images