Multi-load air-ground cooperative rescue system and communication networking method thereof

Abstract

This invention relates to the field of rescue system technology, specifically to a multi-payload air-ground collaborative rescue system and its communication networking method, comprising: a drone, an unmanned vehicle, a command center, and a data sharing platform deployed within the command center; the drone includes: a first frame, a first flight control module, a first power module, a first positioning module, a first communication module, a first mission payload interface, and a panoramic camera, an infrared thermal imager, and a material grabbing mechanism detachably mounted on the first frame via the first mission payload interface; this invention effectively solves the problems of limited visibility and poor terrain adaptability of traditional single rescue equipment in complex environments such as mountainous areas, ruins, and underground spaces by constructing an air-ground collaborative rescue system composed of drones, unmanned vehicles, a command center, and a data sharing platform, and by establishing a communication networking method that integrates a local communication link composed of a 433MHz data transmission module and a wide-area backhaul link composed of a 4G public network;

Description

Technical Field

[0001] This invention relates to the field of rescue system technology, specifically to a multi-load air-ground collaborative rescue system and its communication networking method. Background Technology

[0002] With the intensification of global climate change and the increasing frequency of natural disasters and emergencies, traditional rescue methods suffer from problems such as slow response, limited visibility, poor terrain adaptability, and weak coordination. Existing rescue equipment mostly relies on single drones or unmanned vehicles, making it difficult to achieve efficient, continuous, and precise rescue in complex environments (such as mountains, ruins, and underground spaces). The industry generally recognizes the golden rescue time as 72 hours after a disaster. How to achieve rapid reconnaissance, accurate positioning, material delivery, and casualty transfer in extreme environments has become a core pain point in the field of emergency rescue.

[0003] Therefore, we have made improvements to this by proposing a multi-load air-to-ground collaborative rescue system and its communication networking method. Summary of the Invention

[0004] The purpose of this invention is to provide a multi-load air-ground collaborative rescue system and its communication networking method to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: The system includes: a drone, an unmanned vehicle, a command center, and a data sharing platform deployed within the command center; the drone includes: a first frame, a first flight control module, a first power module, a first positioning module, a first communication module, a first payload interface, a panoramic camera, an infrared thermal imager, and a material grabbing mechanism detachably mounted on the first frame via the first payload interface; the first positioning module includes a first GPS receiver and a first RTK receiver; the first communication module includes a first data transmission module operating at a frequency of 433MHz and a first 4G communication module; the first flight control module includes a first microcontroller, a first inertial measurement unit, a first electronic speed controller, and a first rotor motor electrically connected to the first electronic speed controller; the material grabbing mechanism includes a pod, a winch installed inside the pod, a rope wound around the winch, and a hook connected to the rope; the drone includes: a first frame, a first flight control module, a first power module, a first positioning module, a first positioning module, a first communication module, a first mission payload interface ... The vehicle includes: a second frame, a second motion control module, a second power module, a second positioning module, a second communication module, a second task load interface, a robotic arm detachably mounted on the second frame via the second task load interface, a material storage compartment, a drive wheel assembly mounted on the bottom of the second frame, and an obstacle avoidance sensor; the second positioning module includes a second GPS receiver and a second RTK receiver; the second communication module includes a second data transmission module operating at a frequency of 433MHz and a second 4G communication module; the second motion control module includes a second microcontroller and a motor driver; the drive wheel assembly includes a DC motor electrically connected to the motor driver and omnidirectional wheels mechanically connected to the output shaft of the DC motor; the robotic arm includes a base, joints connected in series and mechanically connected at their ends to the base, an end effector mechanically connected to the end joint, and joint motors respectively disposed at each joint.

[0006] A communication networking method for a multi-load air-ground collaborative rescue system includes the following steps: the end effector includes a demolition tool or a gripper; the drone also includes a 3D modeling device and a servo mechanism mechanically connected to the material grabbing mechanism; the unmanned vehicle also includes radar and a panoramic camera.

[0007] As a preferred technical solution of this application, a first communication link is established between the first data transmission module and the second data transmission module; a second communication link is established between the first 4G communication module and the command center; a third communication link is established between the second 4G communication module and the command center; the first data acquired by the UAV is sent to the unmanned vehicle through the first communication link and simultaneously to the command center through the second communication link; the second data acquired by the unmanned vehicle is sent to the UAV through the first communication link and simultaneously to the command center through the third communication link; the command center aggregates the first data and the second data to the data sharing platform; when the second communication link is interrupted, the UAV sends the first data to the unmanned vehicle through the first communication link, and the unmanned vehicle forwards it to the command center through the third communication link; when the third communication link is interrupted, the unmanned vehicle sends the second data to the UAV through the first communication link, and the UAV forwards it to the command center through the second communication link; the first data includes first positioning data, first image data collected by a panoramic camera, and thermal imaging data collected by an infrared thermal imager; the second data includes second positioning data and obstacle avoidance data collected by an obstacle avoidance sensor.

[0008] As a preferred technical solution of this application, the method includes the following steps: After the UAV takes off, the panoramic camera acquires environmental images, the infrared thermal imager acquires thermal images, and these images are sent to the command center and stored in the data sharing platform via the second communication link; the second motion control module acquires second positioning data and sends it to the command center and stored in the data sharing platform via the third communication link; the command center generates first path planning data based on the environmental images and the thermal images, and sends it to the UAV via the third communication link; the second motion control module controls the movement of the drive wheel assembly based on the first path planning data; during the movement of the UAV, the obstacle avoidance sensor acquires obstacle distance data, the second motion control module generates obstacle avoidance path data based on the obstacle distance data and controls the movement of the drive wheel assembly; the second motion control module sends the obstacle avoidance path data to the UAV via the first communication link, and the first flight control module adjusts the hovering position of the UAV based on the obstacle avoidance path data; when the UAV hovers above the target position, the material grasping mechanism releases materials; when the UAV travels to the target position, the robotic arm performs a grasping action or a demolition action.

[0009] As a preferred technical solution of this application, the UAV obtains first positioning data through the first positioning module, and the unmanned vehicle obtains second positioning data through the second positioning module; the first flight control module receives the second positioning data through the first communication link, calculates relative distance data based on the first positioning data and the second positioning data, and adjusts the flight position of the UAV based on the relative distance data; the first flight control module obtains the flight attitude data of the first inertial measurement unit, and controls the rotational speed of each of the first rotor motors through the first electronic speed controller; the command center sends a first flight command to the UAV through the second communication link, and sends a first driving command to the unmanned vehicle through the third communication link; the unmanned vehicle sends a second flight command to the UAV through the first communication link, and the UAV sends a second driving command to the unmanned vehicle through the first communication link.

[0010] As a preferred technical solution of this application, the method further includes the following steps: In the hovering state of the UAV, the 3D modeling device collects 3D environmental data and sends it to the command center; the command center generates a 3D environment model based on the 3D environmental data, and the 3D environment model is constructed through the following steps: converting the 3D environmental data into point cloud data; performing voxel filtering on the point cloud data to divide the point cloud space into voxel grids, retaining the centroid point within each voxel grid; performing triangulation on the filtered point cloud data to generate a 3D mesh model; storing the 3D environment model in the data sharing platform; and updating the first path planning data based on the 3D environment model and sending it to the unmanned vehicle through the third communication link.

[0011] As a preferred technical solution of this application, the generation of obstacle avoidance path data adopts the dynamic window method, including the following steps: establishing the kinematic model of the unmanned vehicle; sampling multiple sets of speed combinations in the speed space, wherein the speed space is jointly defined by linear velocity and angular velocity constraints, motor acceleration and deceleration performance constraints, and safe braking distance constraints; performing trajectory prediction on each sampled speed combination to generate a predicted trajectory; evaluating the heading angle, safe distance, and speed of each predicted trajectory; selecting the optimal speed combination as the optimal speed command to control the movement of the drive wheel set.

[0012] As a preferred technical solution of this application, the cooperative localization of the UAV and the unmanned vehicle is achieved through data fusion via extended Kalman filtering, comprising the following steps: establishing a system state vector, the system state vector including the first positioning data, the second positioning data, the velocity component of the UAV, and the velocity component of the unmanned vehicle; establishing a state prediction equation and an observation equation, the observation vector including the positioning data output by the first positioning module, the positioning data output by the second positioning module, and the relative distance data measured by the first communication link; performing Kalman filter updates, calculating the Kalman gain, updating the state estimate, and updating the covariance matrix.

[0013] As a preferred technical solution of this application, the thermal imaging image acquired by the infrared thermal imager is used to identify trapped personnel through an AI algorithm. The AI ​​algorithm employs a convolutional neural network and includes the following steps: inputting the thermal imaging image into the convolutional neural network, which includes an input layer, multiple convolutional layers, multiple pooling layers, a fully connected layer, and an output layer; the convolutional layers perform convolution operations, and the convolution operation formula is: ; in For the input feature map, For convolution kernel, For bias terms, and The convolution kernel size; the pooling layer performs max pooling operation, and the pooling operation formula is: ; in The pooling window size; the fully connected layer performs a linear transformation: ;in For the input vector, This is the weight matrix. The bias vector is used; the output layer outputs the classification probability through the Softmax function: ;in For the first Output values ​​for each category, The total number of categories, For the first The predicted probability of each category is calculated; detection results with a predicted probability greater than a preset threshold are marked as trapped personnel, and the location coordinates of the trapped personnel are generated.

[0014] As a preferred technical solution of this application, the data sharing platform adopts a distributed database architecture, including a master node deployed at the command center and slave nodes deployed at the UAV and the unmanned vehicle; the master node and the slave nodes synchronize data through the second communication link and the third communication link; the data synchronization adopts a timestamp-based incremental synchronization mechanism, including the following steps: the slave node records the timestamp of local data changes, generates an incremental data packet and sends it to the master node; the master node compares the local timestamp with the timestamp of the incremental data packet, and if the timestamp of the incremental data packet is later than the local timestamp, it performs a data update; the master node broadcasts the merged data to all slave nodes.

[0015] Compared with the prior art, the beneficial effects of the present invention are: By constructing an air-ground collaborative rescue system consisting of drones, unmanned vehicles, a command center, and a data sharing platform, and establishing a communication network method that integrates a local communication link composed of 433MHz data transmission modules with a wide-area backhaul link composed of 4G public network, the system effectively solves the problems of limited field of vision, poor terrain adaptability, and weak collaborative capabilities of traditional single rescue equipment in complex environments such as mountainous areas, ruins, and underground spaces. When the public network is interrupted, the collaborative operation between drones and unmanned vehicles can still be maintained through the 433MHz link, and data forwarding can be completed using the uninterrupted link, ensuring communication reliability in extreme environments. Furthermore, by equipping drones with panoramic cameras, infrared thermal imagers, 3D modeling equipment, and material grasping mechanisms, aerial... Global perception and precise delivery, combined with unmanned vehicles equipped with robotic arms, material storage bins, obstacle avoidance sensors, and drive wheel sets, enable rapid ground mobility and precise operations, forming a heterogeneous collaborative mode of "air-guided ground and ground-supported air," which significantly improves rescue response speed and environmental adaptability. By aggregating positioning data, environmental images, thermal imaging data, and 3D models from drones and unmanned vehicles through a data sharing platform, and combining AI visual recognition algorithms with extended Kalman filtering for fusion positioning, accurate identification of trapped personnel and relative positioning of air and ground equipment are achieved. Furthermore, by combining obstacle avoidance paths generated by the dynamic window method with global paths updated by the A× algorithm, the accuracy of path planning and operational safety in complex terrain are significantly improved. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall structure of the present invention. Detailed Implementation

[0017] 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.

[0018] This invention provides a technical solution: such as Figure 1 The system, illustrating a multi-payload air-to-ground collaborative rescue system and its communication networking method, includes: a drone, an unmanned vehicle, a command center, and a data sharing platform deployed within the command center. The drone comprises: a first frame, a first flight control module, a first power module, a first positioning module, a first communication module, a first mission payload interface, a panoramic camera, an infrared thermal imager, and a material grabbing mechanism detachably mounted on the first frame via the first mission payload interface. The first positioning module includes a first GPS receiver and a first RTK receiver. The first communication module includes a first data transmission module operating at a frequency of 433MHz and a first 4G communication module. The first flight control module includes a first microcontroller, a first inertial measurement unit, a first electronic speed controller, and a first rotor motor electrically connected to the first electronic speed controller. The material grabbing mechanism includes a pod, a winch installed inside the pod, a rope wound around the winch, and a connecting... The unmanned vehicle includes: a second frame, a second motion control module, a second power module, a second positioning module, a second communication module, a second payload interface, a robotic arm detachably mounted on the second frame via the second payload interface, a material storage compartment, a drive wheel assembly mounted on the bottom of the second frame, and an obstacle avoidance sensor; the second positioning module includes a second GPS receiver and a second RTK receiver; the second communication module includes a second data transmission module with a working frequency band of 433MHz and a second 4G communication module; the second motion control module includes a second microcontroller and a motor driver; the drive wheel assembly includes a DC motor electrically connected to the motor driver and a universal wheel mechanically connected to the output shaft of the DC motor; the robotic arm includes a base, joints connected in series and whose first end is mechanically connected to the base, an end effector mechanically connected to the end joint, and joint motors respectively disposed at each joint.

[0019] A communication networking method for a multi-load air-ground collaborative rescue system includes the following steps: the end effector includes a demolition tool or a gripper; the drone also includes a 3D modeling device and a servo mechanism mechanically connected to the material grabbing mechanism; the unmanned vehicle also includes radar and a panoramic camera.

[0020] Furthermore, a first communication link is established between the first data transmission module and the second data transmission module; a second communication link is established between the first 4G communication module and the command center; a third communication link is established between the second 4G communication module and the command center; the first data acquired by the UAV is sent to the unmanned vehicle via the first communication link and simultaneously to the command center via the second communication link; the second data acquired by the unmanned vehicle is sent to the UAV via the first communication link and simultaneously to the command center via the third communication link; the command center aggregates the first data and the second data to the data sharing platform; when the second communication link is interrupted, the UAV sends the first data to the unmanned vehicle via the first communication link, and the unmanned vehicle forwards it to the command center via the third communication link; when the third communication link is interrupted, the unmanned vehicle sends the second data to the UAV via the first communication link, and the UAV forwards it to the command center via the second communication link; the first data includes first positioning data, first image data collected by the panoramic camera, and thermal imaging data collected by the infrared thermal imager; the second data includes second positioning data and obstacle avoidance data collected by the obstacle avoidance sensor.

[0021] Further, the process includes the following steps: After the UAV takes off, the panoramic camera acquires environmental images, and the infrared thermal imager acquires thermal images, which are then sent to the command center and stored in the data sharing platform via the second communication link; the second motion control module acquires second positioning data and sends it to the command center and stored in the data sharing platform via the third communication link; the command center generates first path planning data based on the environmental images and the thermal images, and sends it to the UAV via the third communication link; the second motion control module controls the movement of the drive wheel assembly based on the first path planning data; during the movement of the UAV, the obstacle avoidance sensor acquires obstacle distance data, and the second motion control module generates obstacle avoidance path data based on the obstacle distance data and controls the movement of the drive wheel assembly; the second motion control module sends the obstacle avoidance path data to the UAV via the first communication link, and the first flight control module adjusts the hovering position of the UAV based on the obstacle avoidance path data; when the UAV hovers above the target position, the material grasping mechanism releases the material; when the UAV travels to the target position, the robotic arm performs a grasping or dismantling action.

[0022] Furthermore, the UAV acquires first positioning data through the first positioning module, and the unmanned vehicle acquires second positioning data through the second positioning module; the first flight control module receives the second positioning data through the first communication link, calculates relative distance data based on the first positioning data and the second positioning data, and adjusts the flight position of the UAV based on the relative distance data; the first flight control module acquires the flight attitude data of the first inertial measurement unit, and controls the rotational speed of each of the first rotor motors through the first electronic speed controller; the command center sends a first flight command to the UAV through the second communication link, and sends a first driving command to the unmanned vehicle through the third communication link; the unmanned vehicle sends a second flight command to the UAV through the first communication link, and the UAV sends a second driving command to the unmanned vehicle through the first communication link.

[0023] Furthermore, the method includes the following steps: While the UAV is hovering, the 3D modeling device collects 3D environmental data and sends it to the command center; the command center generates a 3D environment model based on the 3D environmental data, and the 3D environment model is constructed through the following steps: converting the 3D environmental data into point cloud data; performing voxel filtering on the point cloud data to divide the point cloud space into voxel grids, retaining the centroid point within each voxel grid; performing triangulation on the filtered point cloud data to generate a 3D mesh model; storing the 3D environment model on the data sharing platform; and the command center updates the first path planning data based on the 3D environment model and sends it to the unmanned vehicle via the third communication link.

[0024] Furthermore, the generation of the obstacle avoidance path data adopts the dynamic window method, including the following steps: establishing the kinematic model of the unmanned vehicle; sampling multiple sets of speed combinations in the speed space, which is jointly defined by linear velocity and angular velocity constraints, motor acceleration and deceleration performance constraints, and safe braking distance constraints; performing trajectory prediction on each sampled speed combination to generate a predicted trajectory; evaluating the heading angle, safe distance, and speed of each predicted trajectory; selecting the optimal speed combination as the optimal speed command to control the movement of the drive wheel set.

[0025] Furthermore, the cooperative localization of the UAV and the unmanned vehicle is achieved through data fusion using extended Kalman filtering, including the following steps: establishing a system state vector, which includes the first positioning data, the second positioning data, the velocity component of the UAV, and the velocity component of the unmanned vehicle; establishing a state prediction equation and an observation equation, where the observation vector includes the positioning data output by the first positioning module, the positioning data output by the second positioning module, and the relative distance data measured by the first communication link; performing Kalman filter updates, calculating the Kalman gain, updating the state estimate, and updating the covariance matrix.

[0026] Furthermore, the thermal imaging images acquired by the infrared thermal imager are used to identify trapped personnel through an AI algorithm. This AI algorithm employs a convolutional neural network and includes the following steps: inputting the thermal imaging image into the convolutional neural network, which includes an input layer, multiple convolutional layers, multiple pooling layers, a fully connected layer, and an output layer; the convolutional layers perform convolution operations, with the convolution operation formula being: ; in For the input feature map, For convolution kernel, For bias terms, and The convolution kernel size; the pooling layer performs max pooling operation, and the pooling operation formula is: ; in The pooling window size; the fully connected layer performs a linear transformation: ;in For the input vector, This is the weight matrix. The bias vector is used; the output layer outputs the classification probability through the Softmax function: ;in For the first Output values ​​for each category, The total number of categories, For the first The predicted probability of each category is calculated; detection results with a predicted probability greater than a preset threshold are marked as trapped personnel, and the location coordinates of the trapped personnel are generated.

[0027] Furthermore, the data sharing platform adopts a distributed database architecture, including a master node deployed at the command center and slave nodes deployed at the UAVs and unmanned vehicles; the master node and the slave nodes synchronize data through the second communication link and the third communication link; the data synchronization adopts a timestamp-based incremental synchronization mechanism, including the following steps: the slave node records the timestamp of local data changes, generates an incremental data packet and sends it to the master node; the master node compares the local timestamp with the timestamp of the incremental data packet, and if the timestamp of the incremental data packet is later than the local timestamp, it performs a data update; the master node broadcasts the merged data to all slave nodes.

[0028] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0029] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-load air-ground collaborative rescue system, characterized in that: include: Unmanned aerial vehicles, unmanned vehicles, command center, and data sharing platform deployed within the command center; The drone includes: a first frame, a first flight control module, a first power module, a first positioning module, a first communication module, a first payload interface, a panoramic camera, an infrared thermal imager, and a material grasping mechanism detachably mounted on the first frame via the first payload interface; the first positioning module includes a first GPS receiver and a first RTK receiver; the first communication module includes a first data transmission module operating at a frequency of 433MHz and a first 4G communication module; the first flight control module includes a first microcontroller, a first inertial measurement unit, a first electronic speed controller, and a first rotor motor electrically connected to the first electronic speed controller; the material grasping mechanism includes a pod, a winch installed inside the pod, a rope wound around the winch, and a hook connected to the rope; The unmanned vehicle includes: a second frame, a second motion control module, a second power module, a second positioning module, a second communication module, a second payload interface, a robotic arm detachably mounted on the second frame via the second payload interface, a material storage compartment, a drive wheel assembly mounted on the bottom of the second frame, and obstacle avoidance sensors; the second positioning module includes a second GPS receiver and a second RTK receiver; the second communication module includes a second data transmission module operating at a frequency of 433MHz and a second 4G communication module; the second motion control module includes a second microcontroller and a motor driver; the drive wheel assembly includes a DC motor electrically connected to the motor driver and omnidirectional wheels mechanically connected to the output shaft of the DC motor; the robotic arm includes an end effector mechanically connected to an end joint.

2. A communication networking method for a multi-load air-to-ground collaborative rescue system. A multi-load air-ground collaborative rescue system according to claim 1, characterized in that: Includes the following steps: The end effector includes a demolition tool or a gripper; the drone also includes a 3D modeling device and a servo mechanism mechanically connected to the material grasping mechanism; the unmanned vehicle also includes radar and a panoramic camera.

3. The communication networking method for a multi-load air-to-ground collaborative rescue system according to claim 1, characterized in that: A first communication link is established between the first data transmission module and the second data transmission module; a second communication link is established between the first 4G communication module and the command center; a third communication link is established between the second 4G communication module and the command center; the first data acquired by the UAV is sent to the unmanned vehicle via the first communication link and simultaneously to the command center via the second communication link; the second data acquired by the unmanned vehicle is sent to the UAV via the first communication link and simultaneously to the command center via the third communication link; the command center aggregates the first data and the second data to the data sharing platform; when the second communication link is interrupted, the UAV sends the first data to the unmanned vehicle via the first communication link, and the unmanned vehicle forwards it to the command center via the third communication link; when the third communication link is interrupted, the unmanned vehicle sends the second data to the UAV via the first communication link, and the UAV forwards it to the command center via the second communication link; the first data includes first positioning data, first image data collected by a panoramic camera, and thermal imaging data collected by an infrared thermal imager; the second data includes second positioning data and obstacle avoidance data collected by an obstacle avoidance sensor.

4. The communication networking method for a multi-load air-to-ground collaborative rescue system according to claim 1, characterized in that: Includes the following steps: After the drone takes off, the panoramic camera captures environmental images, and the infrared thermal imager captures thermal images, which are then transmitted to the command center and stored in the data sharing platform via the second communication link. The second motion control module acquires second positioning data and transmits it to the command center and stores it in the data sharing platform via the third communication link. The command center generates first path planning data based on the environmental images and the thermal images, and transmits it to the unmanned vehicle via the third communication link. The second motion control module controls the movement of the drive wheel assembly based on the first path planning data; During the movement of the unmanned vehicle, the obstacle avoidance sensor collects obstacle distance data, and the second motion control module generates obstacle avoidance path data based on the obstacle distance data and controls the movement of the drive wheel assembly; the second motion control module sends the obstacle avoidance path data to the drone through the first communication link, and the first flight control module adjusts the hovering position of the drone based on the obstacle avoidance path data; when the drone hovers above the target position, the material grasping mechanism releases the material; when the unmanned vehicle travels to the target position, the robotic arm performs a grasping action or a demolition action.

5. The communication networking method for a multi-load air-to-ground collaborative rescue system according to claim 4, characterized in that: The UAV acquires first positioning data through the first positioning module, and the unmanned vehicle acquires second positioning data through the second positioning module; the first flight control module receives the second positioning data through the first communication link, calculates relative distance data based on the first and second positioning data, and adjusts the UAV's flight position based on the relative distance data; the first flight control module acquires flight attitude data from the first inertial measurement unit and controls the rotational speed of each of the first rotor motors through the first electronic speed controller; the command center sends a first flight command to the UAV through the second communication link and a first driving command to the unmanned vehicle through the third communication link; the unmanned vehicle sends a second flight command to the UAV through the first communication link, and the UAV sends a second driving command to the unmanned vehicle through the first communication link.

6. The communication networking method for a multi-load air-to-ground collaborative rescue system according to claim 4, characterized in that: It also includes the following steps: While the drone is hovering, the 3D modeling device collects 3D environmental data and sends it to the command center. The command center generates a 3D environment model based on the 3D environmental data. The 3D environment model is constructed through the following steps: converting the 3D environmental data into point cloud data; performing voxel filtering on the point cloud data to divide the point cloud space into voxel grids, retaining the centroid point within each voxel grid. The filtered point cloud data is triangulated to generate a three-dimensional mesh model; the three-dimensional environment model is stored in the data sharing platform; the command center updates the first path planning data according to the three-dimensional environment model and sends it to the unmanned vehicle through the third communication link.

7. The communication networking method for a multi-load air-to-ground collaborative rescue system according to claim 4, characterized in that: The obstacle avoidance path data is generated using a dynamic window method, including the following steps: establishing a kinematic model of the unmanned vehicle; sampling multiple speed combinations in a speed space, which is jointly defined by linear velocity and angular velocity constraints, motor acceleration and deceleration performance constraints, and safe braking distance constraints; predicting the trajectory for each sampled speed combination to generate a predicted trajectory; evaluating the heading angle, safe distance, and speed for each predicted trajectory; and selecting the optimal speed combination as the optimal speed command to control the movement of the drive wheel set.

8. The communication networking method for a multi-load air-to-ground collaborative rescue system according to claim 4, characterized in that: The cooperative localization of the UAV and the unmanned vehicle is achieved through data fusion using extended Kalman filtering, comprising the following steps: establishing a system state vector, which includes the first positioning data, the second positioning data, the velocity component of the UAV, and the velocity component of the unmanned vehicle; establishing a state prediction equation and an observation equation, wherein the observation vector includes the positioning data output by the first positioning module, the positioning data output by the second positioning module, and the relative distance data measured by the first communication link; performing Kalman filter updates, calculating the Kalman gain, updating the state estimate, and updating the covariance matrix.

9. The communication networking method for a multi-load air-to-ground collaborative rescue system according to claim 4, characterized in that: The thermal imaging images acquired by the infrared thermal imager are used to identify trapped personnel through an AI algorithm. The AI ​​algorithm employs a convolutional neural network and includes the following steps: inputting the thermal imaging images into the convolutional neural network, which includes an input layer, multiple convolutional layers, multiple pooling layers, a fully connected layer, and an output layer; the convolutional layers perform convolution operations, with the convolution operation formula being: ； in For the input feature map, For convolution kernel, For bias terms, and The convolution kernel size; the pooling layer performs max pooling operation, and the pooling operation formula is: ； in The pooling window size; the fully connected layer performs a linear transformation: ;in For the input vector, This is the weight matrix. The bias vector is used; the output layer outputs the classification probability through the Softmax function: ;in For the first Output values ​​for each category, The total number of categories, For the first The predicted probability of each category is calculated; detection results with a predicted probability greater than a preset threshold are marked as trapped personnel, and the location coordinates of the trapped personnel are generated.

10. The communication networking method for a multi-load air-to-ground collaborative rescue system according to claim 4, characterized in that: The data sharing platform adopts a distributed database architecture, including a master node deployed at the command center and slave nodes deployed at the UAV and the unmanned vehicle; the master node and the slave nodes synchronize data through the second communication link and the third communication link; The data synchronization adopts a timestamp-based incremental synchronization mechanism, which includes the following steps: the slave node records the timestamp of local data changes, generates an incremental data packet, and sends it to the master node; The master node compares the local timestamp with the incremental data packet timestamp. If the incremental data packet timestamp is later than the local timestamp, the master node performs a data update. The master node then broadcasts the merged data to all slave nodes.

Images