Quadruped robot system for indoor fire rescue

Abstract

The invention discloses a quadruped robot system for indoor fire rescue. The quadruped robot system comprises a quadruped robot body, a main control unit carried by the robot, a sensing unit, a navigation unit, an execution unit and a motion control unit. The main control unit comprises two high-performance industrial personal computers which respectively process a motion control algorithm and a visual perception algorithm. The industrial personal computer communicates with other sensors through an IP and a serial port, and wirelessly communicates with the rear end of the fire-fighting command center through a TCP. After the fire-fighting command center issues a rescue task to the quadruped robot, the robot constructs a global map through the laser radar, and searches a fire source and trapped persons through the depth camera. When the robot recognizes the target, the robot sends the target position to the rear end of the fire-fighting command center, automatically avoids obstacles, and navigates to the target position to execute fire extinguishing and rescue work. The firefighter can be helped to quickly find a fire source and trapped persons, determine the positions of the trapped persons in the map, and guide the firefighter to reach the area to carry out rescue.

Description

Technical Field

[0001] The invention belongs to the technical field of fire-fighting robots, and in particular relates to a quadruped robot system for indoor firefighting and rescue. Background Art

[0002] With the rapid development of industrialization and urbanization, disasters such as fires occur frequently, especially in complex environments. The increase in super buildings in the process of urbanization has brought increasing pressure on firefighting. These buildings have complex structures, numerous functions, and large areas, making fire rescue difficult and inefficient, and posing a great threat to the personal safety of firefighters. In order to improve the efficiency of firefighting and rescue, it is particularly important to develop a fire reconnaissance robot that can assist firefighters in fire reconnaissance.

[0003] Existing reconnaissance robots face several major problems in fire scenes: First, the reliability of positioning in buildings is low, because existing indoor positioning technology relies on facilities such as base stations, and the power supply to the base stations cannot be guaranteed during fire scenes. In addition, the high temperature environment affects the reliability of the base stations, resulting in most reconnaissance robots being suitable for outdoor environments. Second, existing robots lack the ability to build real-time maps. High-temperature combustion in fire scenes causes large changes in building structures, and the maps loaded on existing robots are outdated, making it impossible to accurately position and navigate the robots. Finally, the robot controller is bulky and inconvenient to carry. Current fire scene robot controllers are mainly large industrial computers, which need to be placed in a fixed position during operation. They lack flexibility, increase the burden on firefighters, occupy their hands, and affect other rescue operations.

[0004] Therefore, developing a fire reconnaissance robot that can realize autonomous navigation and map construction, detect the location of the fire source, and search and rescue trapped people in the fire scene is of great significance to improving rescue efficiency and protecting the lives of firefighters. This robot needs to have a high degree of autonomy, flexibility and adaptability, be able to work stably in a complex and changeable fire environment, and provide firefighters with key reconnaissance information and rescue support. To this end, the present invention proposes a quadruped robot system for indoor fire rescue. Summary of the invention

[0005] The purpose of the present invention is to provide a quadruped robot system for indoor firefighting and rescue, which can realize autonomous navigation and map construction, detect the location of fire source, search and rescue trapped persons in the fire scene, has a high degree of autonomy, flexibility and adaptability, can work stably in complex and changeable fire environment, and provide key reconnaissance information and rescue support for firefighters.

[0006] The technical solution adopted by the present invention is as follows:

[0007] A quadruped robot system for indoor firefighting and rescue, comprising a robot device and a control system for controlling the robot device;

[0008] The robot device includes a quadruped robot body, an industrial computer, an edge computer, a laser radar, a depth camera, a fire extinguishing device, a toxic gas detector and a microphone.

[0009] The control system includes a main control unit, a sensing unit, a navigation unit, a motion control unit, an execution unit and a communication unit;

[0010] The main control unit is responsible for the motion control algorithm through the industrial computer, and the main control unit is responsible for the visual perception algorithm through the edge computer. The industrial computer and the edge computer communicate in a ROS master-slave manner to share topics and data. The main control unit communicates with the perception unit, the navigation unit, the execution unit, the motion control unit and the communication unit through IP and serial port;

[0011] The perception unit runs a target recognition algorithm, and the perception unit provides scanning point cloud data to the navigation unit through the laser radar, and the perception unit provides image and depth information to the target recognition algorithm through the depth camera; the target recognition algorithm uses the YOLOv8 algorithm, and the YOLOv8 algorithm is used to identify the fire source and trapped persons, and obtain (x, y) coordinates and distance information; the laser radar communicates with the industrial computer through IP, and the depth camera is connected to the industrial computer through USB;

[0012] The navigation unit first obtains the scanning point cloud data provided by the perception unit through the laser radar, and then uses the SLAM mapping algorithm, relocation algorithm, and path planning algorithm to complete intelligent navigation; the SLAM algorithm adopts FAST-LIO, which is a radar-inertial odometer mapping algorithm with high computational efficiency and robustness; the relocation algorithm adopts FAST-LIO-LOCALIZATION, and the relocation algorithm uses a tightly coupled iterative extended Kalman filter to fuse radar feature points with IMU data to achieve robust positioning in fast-moving, noisy or chaotic environments; the path planning algorithm receives the scanning point cloud data provided by the laser radar, constructs a three-dimensional point cloud map, and uses octomap to convert the three-dimensional point cloud map into a two-dimensional grid map as a global map, move_base uses the global map to generate a global cost map for global path planning, and uses the laser radar scanning data to plan a real-time path to achieve obstacle avoidance and navigation;

[0013] The motion control unit includes an NMPC+WBC control framework and an MPC adaptive algorithm based on DOB; the NMPC+WBC control framework is built based on OCS2 and ros-control and is used to calculate the optimal system state and input in real time; the MPC algorithm based on DOB uses Lyapunov analysis, introduces two DOBs to offset the influence of uncertainty, and incorporates friction constraints into the ground reaction force; the adaptive algorithm is applied to the Aliengo robot, which significantly improves the tracking accuracy of the center of mass position and reference direction;

[0014] The execution unit is used to control the fire extinguishing device, toxic gas detector and microphone; the execution unit controls the orientation and pitch angle of the robot quadruped robot body to adjust the angle of the fire extinguishing device. Ensure that the nozzle of the fire extinguisher is always aimed at the fire source, and turn on the fire extinguisher to achieve accurate and efficient fire extinguishing; the toxic gas detection device monitors the concentration of CO, SO, SO2 toxic gases in the environment in real time and reports to the command center; the microphone allows firefighters outside the fire scene to communicate with trapped people in the fire scene to calm the trapped people;

[0015] The communication unit uses a TCP module to communicate and transmit with the fire command center; the control system uses TCP to transmit field data to the command center for command personnel to view, and the command center also uses TCP to assign tasks to the robot device and control the robot's actions through the control system.

[0016] The technical effects achieved by the present invention are:

[0017] The fire-fighting quadruped robot includes a quadruped robot body, a main control unit, a perception unit, a navigation unit, an execution unit and a motion control unit carried by the robot. The main control units are two high-performance industrial computers, which process motion control algorithms and visual perception algorithms respectively, and communicate with each other through the ROS master-slave mode. The industrial computer communicates with other sensors through IP and serial ports, and communicates with the back end of the fire command center through TCP wireless communication. After the fire command center issues a rescue mission to the quadruped robot, the robot builds a global map through a laser radar, and searches for the source of the fire and trapped people through a depth camera. When the robot recognizes the target, it will send the target position to the back end of the fire command center and autonomously avoid obstacles and navigate to the target position to perform fire extinguishing and rescue work. The present invention can help firefighters quickly find the source of the fire and trapped people, determine their location in the map, and guide firefighters to the area to implement rescue. The robot enhances the perception ability of firefighters in the fire environment, ensures the life safety of firefighters, and improves the rescue efficiency. It aims to solve the problems of existing fire robots in the fire scene, such as weak ability to cross terrain, low positioning reliability, lack of autonomous navigation, fire extinguishing and rescue functions. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a three-dimensional structural diagram of a fire-fighting quadruped robot;

[0019] Figure 2 This is the control system block diagram of the fire-fighting quadruped robot;

[0020] Figure 3 It is the control flow chart of the fire-fighting quadruped robot. DETAILED DESCRIPTION

[0021] In order to make the purpose and advantages of the present invention more clearly understood, the present invention is specifically described below in conjunction with embodiments. It should be understood that the following text is only used to describe one or several specific embodiments of the present invention, and does not strictly limit the scope of protection of the specific claims of the present invention.

[0022] like Figure 1 - Figure 3 As shown, a quadruped robot system for indoor firefighting and rescue, including a robot device and a control system for controlling the robot device;

[0023] The robot device includes a quadruped robot body, an industrial computer, an edge computer, a laser radar, a depth camera, a fire extinguishing device, a toxic gas detector and a pickup; in this embodiment, the quadruped robot body selects the Aliengo robot, which has 12 high-performance servo motors that can output large torque. Under its high-level algorithm, it can carry a maximum of 13kg and has good impact resistance. The industrial computer selects the i9-14900 industrial computer, the edge computer selects the Nvidia Jetson AGX Orin, the laser radar selects the Livox mid360, and the depth camera selects the Intel Realsense d456.

[0024] The support plate material of the quadruped robot body of the present invention is made of carbon material, which has the characteristics of high strength and low mass, and the output voltages are 5V, 12V, 19V and BAT. With an adjustable voltage stabilizing module, the power supply requirements of all devices can be met.

[0025] The rated power of the industrial computer and edge computer of the fire-fighting quadruped robot of the present invention is relatively large, so it is considered to power them separately; because both are powered by 19V, a 6S, 8000mah aircraft model power supply is selected, with a rated working voltage of 22.8V, which is then converted to 19V voltage through a DCDC voltage regulator module to power the industrial computer.

[0026] The depth camera of the laser radar controlled by the perception unit of the fire-fighting quadruped robot of the present invention, the fire extinguishing device controlled by the execution unit, the toxic gas detector and the microphone and other equipment are all powered by the quadruped robot. The body output voltage of the Aliengo quadruped robot is 5V, 12V, 19V and BAT, and it is matched with an adjustable voltage stabilizing module to meet the power supply requirements of all equipment.

[0027] The Livox mid360 laser radar in this embodiment is small and light, weighing 265 grams, and is easy to install and carry. The Intel Realsense d456 depth camera in this embodiment is a long-distance, wide-angle, global shutter depth camera launched by Intel. It has an IP65 protection grade and can adapt to outdoor environments and various extreme conditions.

[0028] The control system includes a main control unit, a sensing unit, a navigation unit, a motion control unit, an execution unit and a communication unit;

[0029] The main control unit is responsible for the motion control algorithm through the industrial computer, and the main control unit is responsible for the visual perception algorithm through the edge computer. The industrial computer and the edge computer communicate in ROS master-slave mode to share topics and data; the i9-14900 processor equipped with the industrial computer is designed to provide powerful CPU computing power for the motion control algorithm to ensure the stability of the quadruped robot's main body movement; the Nvidia Jetson AGX Orin edge computer has 64G video memory and is designed for advanced robots and edge AI applications. It is used to process visual perception algorithms. The main control unit communicates with the perception unit, navigation unit, execution unit, motion control unit and communication unit through IP and serial ports;

[0030] The perception unit runs a target recognition algorithm. The perception unit provides scanning point cloud data to the navigation unit through the laser radar, and the perception unit provides image and depth information to the target recognition algorithm through the depth camera. The target recognition algorithm uses the YOLOv8 algorithm, which is used to identify the fire source and trapped people, and obtain (x, y) coordinates and distance information. The laser radar communicates with the industrial computer through IP, and the depth camera is connected to the industrial computer through USB.

[0031] The navigation unit first obtains the scanning point cloud data provided by the perception unit through the lidar. The navigation unit then uses the FAST-LIO SLAM mapping algorithm to build a two-dimensional grid map, the FAST-LIO-LOCALIZATION relocation algorithm to achieve relocation, and the move_base path planning algorithm to complete obstacle avoidance and navigation, and finally realize intelligent navigation;

[0032] The SLAM algorithm uses FAST-LIO, which is a computationally efficient and robust radar-inertial odometer mapping algorithm. The relocalization algorithm uses FAST-LIO-LOCALIZATION, which uses a tightly coupled iterative extended Kalman filter to fuse radar feature points with IMU data to achieve robust positioning in fast-moving, noisy or cluttered environments. The path planning algorithm receives the scanning point cloud data provided by the lidar, constructs a three-dimensional point cloud map, and uses octomap to convert the three-dimensional point cloud map into a two-dimensional grid map as a global map. Move_base uses the global map to generate a global cost map for global path planning, and uses the lidar scanning data to plan real-time paths to achieve obstacle avoidance and navigation.

[0033] In this embodiment, the navigation unit uses the FAST-LIO_LOCALIZATION relocalization algorithm to determine the robot's position in the map. You only need to set the robot's initial position in the command line or specify the initial position in rviz. The global path planner of move_base in this embodiment will combine the global cost map to plan a global path, and the local path planner will combine the robot's odometer information and the local cost map to plan the local path in real time.

[0034] The motion control unit includes the NMPC+WBC control framework and the DOB-based MPC adaptive algorithm; the NMPC+WBC control framework is built on OCS2 and ros-control to calculate the optimal system state and input in real time; the DOB-based MPC algorithm uses Lyapunov analysis, introduces two DOBs to offset the influence of uncertainty, and incorporates friction constraints into the ground reaction force; the adaptive algorithm is applied to the Aliengo robot, which significantly improves the tracking accuracy of the center of mass position and reference direction;

[0035] In this embodiment, the legged_control control framework of the fire-fighting quadruped robot is a high-performance nonlinear model predictive control (NMPC) and whole body control (WBC) framework designed specifically for legged robots, which is built based on OCS2 and ros-control.

[0036] The fire-fighting quadruped robot of this embodiment adds an MPC algorithm based on a disturbance observer (DOB) under the legged_control control framework. This method is based on a single rigid body dynamics model of the quadruped robot and uses Lyapunov analysis to introduce two DOBs to offset the influence of uncertainty.

[0037] The execution unit is used to control the fire extinguishing device, toxic gas detector and microphone. The execution unit controls the orientation and pitch angle of the robot's quadruped body to adjust the angle of the fire extinguishing device. It ensures that the nozzle of the fire extinguisher is always aimed at the fire source, and the fire extinguisher is turned on to achieve accurate and efficient fire extinguishing. The toxic gas detection device monitors the concentration of CO, SO, and SO2 toxic gases in the environment in real time and reports to the command center. The microphone allows firefighters outside the fire scene to communicate with trapped people inside the fire scene to comfort the trapped people.

[0038] The communication unit uses the TCP module to communicate and transmit with the fire command center; the control system uses TCP to transmit field data to the command center for command personnel to view, and the command center also uses TCP to assign tasks to the robot device and control the robot's actions through the control system.

[0039] The working principle of the present invention. In the specific implementation process of the present invention, the robot body used is the Unitree Aliengo quadruped robot. Since the quadruped robot needs to be equipped with hardware equipment such as a fire extinguisher, a power supply, an industrial computer, and a sensor, and the total mass exceeds 6kg, it is necessary to select a quadruped robot with a large load capacity. At the same time, since the application scenario of the present invention is an indoor fire scene, there may be uneven ground, many obstacles, and a complex environment, so the robot is required to have good terrain crossing capabilities and impact resistance. The Unitree Aliengo quadruped robot has 12 high-performance servo motors that can output large torque. Under its high-level algorithm, it can carry a maximum load of 13kg. Aliengo has good impact resistance. When running, jumping, colliding with external objects, or falling, the fuselage can resist the impact load well.

[0040] In order to realize the function of autonomous firefighting and rescue at the fire scene and adapt to different fire scene environments and task requirements, the fire-fighting quadruped robot of the present invention adopts a modular design, which is mainly divided into a main control unit, a sensing unit, a navigation unit, an execution unit and a motion control unit. The hardware equipment on the back of the robot is divided into two layers, and the support plate material is made of carbon material, which has the characteristics of high strength and low quality. The main control unit is the brain of the robot, and the present invention places it on the lower layer, while the sensing unit and the execution unit need to interact with the environment, so they are placed on the upper layer. The units mainly communicate through IP and serial ports. In terms of power supply, the two industrial computers of the main control unit have a large rated power, so separate power supply is considered. Since both are powered by 19V, we use a 6S, 8000mah aircraft model power supply with a rated working voltage of 22.8V, and then convert it to 19V voltage through a DCDC voltage regulator module to power the industrial computer, which can supply two industrial computers to work simultaneously for about half an hour. The perception unit, execution unit and other devices are all powered by the quadruped robot. Aliengo's body output voltages include 5V, 12V, 19V and BAT, and with an adjustable voltage regulator module, it can meet the power supply needs of all devices.

[0041] The main control unit is connected through an industrial computer and an edge computer. One of the industrial computers is a microcomputer equipped with i9-14900, and the edge computer uses Nvidia Jetson AGX Orin. The former is responsible for the motion control algorithm, because the motion control algorithms such as MPC used in the present invention require strong CPU computing power support to ensure the stability of the quadruped robot's movement; the latter is responsible for the visual perception algorithm. Nvidia Jetson AGX Orin has 64G video memory and is designed for advanced robots and edge AI applications. The two industrial computers communicate through the ROS master-slave mode, so that the two industrial computers can share topics and data without worrying about communication problems when running motion control algorithms and visual perception algorithms.

[0042] The hardware of the perception unit includes Livox MID360 laser radar and Intel RealSense D456 depth camera, and the algorithm includes target recognition algorithm. MID360 laser radar communicates with the industrial computer via IP, while the depth camera is connected to the industrial computer via USB. MID360 laser radar is small and light, weighing 265 grams, easy to install and carry. It adopts advanced optical-mechanical system, high point cloud coverage, non-repetitive scanning technology can detect richer details, and the official ROS driver is provided, which is easy to use. D456 depth camera is a long-distance, wide-angle, global shutter depth camera launched by Intel. It has an IP65 protection level and can adapt to outdoor environments and various extreme conditions. Laser radar mainly provides scanning point cloud data for navigation algorithm, and depth camera mainly provides image and depth information for target recognition algorithm. The target recognition algorithm uses YOLOv8, which can accurately identify the fire source and trapped personnel, and obtain the x, y coordinates and distance information of the target. After TF transformation to the world coordinate system, it can be used as the target point for navigation.

[0043] The navigation unit includes SLAM mapping algorithm, relocalization algorithm, and path planning algorithm. The SLAM algorithm uses FAST-LIO, which is a computationally efficient and robust radar-inertial odometer mapping algorithm. It uses a tightly coupled iterative extended Kalman filter to fuse radar feature points with IMU data to achieve robust navigation in fast-moving, noisy, or cluttered environments. It receives point cloud data from the MID360 lidar, builds a three-dimensional point cloud map, and then uses octomap to convert the three-dimensional point cloud map into a two-dimensional grid map as a global map for move_base navigation. After pre-building the map, the FAST_LIO_LOCALIZATION relocalization algorithm is used to determine the robot's position in the map. You only need to set the robot's initial position in the command line or specify the initial position in rviz. It will feature match the real-time point cloud scanned by the radar with the pre-built three-dimensional point cloud map to determine the robot's current true position in the map. The global path planner of move_base will combine the global cost map to plan a global path, and the local path planner will combine the robot's odometer information and the local cost map to plan the local path in real time. The quadruped robot will move along the global path to the target point. When there are obstacles on the path, move_base will plan the local path in real time to bypass the obstacles and realize the navigation and obstacle avoidance function.

[0044] The execution unit includes a fire extinguishing device, a poison gas detection device, and a microphone. When the quadruped robot approaches the fire source, the quadruped robot adjusts its orientation and pitch angle to ensure that the fire extinguisher nozzle is always aimed at the fire source, and turns on the fire extinguisher through the fire extinguisher switch module, thereby achieving accurate and efficient fire extinguishing. The poison gas detection device can monitor the concentration of toxic gases such as CO, SO, SO2 in the current environment in real time and report to the command center. When the robot approaches trapped people, the microphone allows firefighters outside the fire scene to communicate with trapped people inside the fire scene, soothe the trapped people and provide rescue guidance.

[0045] The motion control unit includes a NMPC+WBC control framework and an MPC algorithm based on DOB. The legged_control control framework open sourced by Liao Qiayuan is a high-performance nonlinear model predictive control (NMPC) and whole body control (WBC) framework designed for legged robots, built on OCS2 and ros-control. Legged_control can calculate the optimal system state and input in real time to ensure that the quadruped robot moves in a coordinated and efficient manner, so that the quadruped robot has high motion stability and flexibility. This fire-fighting quadruped robot needs to have a large load capacity, but the model of legged_control does not consider the load and does not have adaptive load capacity. Therefore, the present invention adds an MPC algorithm based on a disturbance observer (DOB) under the legged_control control framework. This method is based on a single rigid body dynamics model of a quadruped robot. Using Lyapunov analysis, we introduce two DOBs to offset the influence of uncertainty, and then we develop a DOB-based MPC framework to incorporate friction constraints into the ground reaction force. Experiments have shown that this algorithm is applied to the tracking accuracy of the center of mass position and reference direction on the Aliengo robot significantly improved.

[0046] The communication unit is a TCP communication module. The quadruped robot uses TCP to conduct long-distance wireless communication and data transmission with the fire command center. The robot can transmit images, videos, point clouds, and on-site conditions of the fire scene to the command center, which can monitor the situation on the scene in real time. The command center can also assign tasks to the robot and control the robot's actions.

[0047] The above is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention shall be implemented according to the conventional means in the art unless otherwise specified and limited.

Claims

1. A quadruped robot system for indoor firefighting and rescue, characterized in that: comprising a robotic device and a control system for controlling the robotic device; The robot device includes a quadruped robot body, an industrial computer, an edge computer, a laser radar, a depth camera, a fire extinguishing device, a toxic gas detector and a pickup; The control system includes a main control unit, a sensing unit, a navigation unit, a motion control unit, an execution unit and a communication unit; The main control unit is responsible for the motion control algorithm through the industrial computer, and the main control unit is responsible for the visual perception algorithm through the edge computer; A target recognition algorithm is run on the perception unit, the perception unit provides scanning point cloud data to the navigation unit through the laser radar, and the perception unit provides image and depth information to the target recognition algorithm through the depth camera; the target recognition algorithm uses the YOLOv8 algorithm; The navigation unit first obtains the scanning point cloud data provided by the perception unit through the laser radar, and then uses the SLAM mapping algorithm, the relocation algorithm, and the path planning algorithm to complete the intelligent navigation; The motion control unit includes an NMPC+WBC control framework and an MPC adaptive algorithm based on DOB; the NMPC+WBC control framework is built based on OCS2 and ros-control and is used to calculate the optimal system state and input in real time; the MPC algorithm based on DOB uses Lyapunov analysis, introduces two DOBs to offset the influence of uncertainty, and incorporates friction constraints into ground reaction forces; The execution unit is used to control the fire extinguishing device, the toxic gas detector and the pickup; The communication unit uses a TCP module to communicate and transmit with the fire command center; the control system uses TCP to transmit field data to the command center for command personnel to view, and the command center also uses TCP to assign tasks to the robot device and control the robot's actions through the control system.

2. A quadruped robot system for indoor firefighting and rescue according to claim 1, characterized in that: The industrial computer and the edge computer communicate in a ROS master-slave mode to share topics and data, and the main control unit communicates with the perception unit, the navigation unit, the execution unit, the motion control unit and the communication unit through IP and serial port.

3. A quadruped robot system for indoor firefighting and rescue according to claim 1, characterized in that: The YOLOv8 algorithm is used to identify the fire source and trapped persons, and obtain (x, y) coordinates and distance information.

4. A quadruped robot system for indoor firefighting and rescue according to claim 1, characterized in that: The laser radar communicates with the industrial computer via IP, and the depth camera is connected to the industrial computer via USB.

5. A quadruped robot system for indoor firefighting and rescue according to claim 1, characterized in that: The SLAM algorithm adopts the FAST-LIO radar-inertial odometer mapping algorithm; the relocalization algorithm adopts FAST-LIO-LOCALIZATION, and the relocalization algorithm uses a tightly coupled iterative extended Kalman filter to fuse radar feature points with IMU data to achieve positioning in fast-moving, noisy or chaotic environments; the path planning algorithm receives scanning point cloud data provided by the lidar, constructs a three-dimensional point cloud map, and uses octomap to convert the three-dimensional point cloud map into a two-dimensional grid map as a global map. Move_base uses the global map to generate a global cost map for global path planning, and uses the lidar scanning data to plan a real-time path to achieve obstacle avoidance and navigation.

6. A quadruped robot system for indoor firefighting and rescue according to claim 1, characterized in that: The execution unit controls and adjusts the orientation and pitch angle of the robot quadruped robot body to adjust the angle of the fire extinguishing device.

7. A quadruped robot system for indoor firefighting and rescue according to claim 1, characterized in that: The poison gas detection device monitors the concentration of CO, SO, and SO2 poisonous gases in the environment in real time and reports to the command center.

8. A quadruped robot system for indoor firefighting and rescue according to claim 1, characterized in that: The microphone can allow firefighters outside the fire scene to communicate with trapped persons in the fire scene.

Images