Indoor intelligent fire-fighting warrior

Abstract

The application discloses an indoor intelligent fire-fighting warrior, which is characterized in that the indoor intelligent fire-fighting warrior comprises a fire-fighting warrior main body, a multi-head monitoring system which is rotatable and has adjustable head pitch angles and is arranged on an indoor roof, a sensor network which is formed by at least one independent sensor system and is arranged at optional positions in other places in the indoor, and a room intelligent door lock. Through intercommunication of the sensor network, the fire-fighting warrior main body and a remote intelligent mobile phone, flexible path planning or self-adaptive path planning in the indoor is realized, movement is realized, the intelligent lock is controlled to be opened in the case that a door is closed, a door is entered to carry out fire extinguishing activities, when there are on-site personnel, the personnel can be directly followed to walk or the personnel can be followed to move to a target position in a tunnel to carry out fire extinguishing activities, and the indoor intelligent fire-fighting warrior can additionally execute remote fire-fighting help when receiving a distress voice information.

Description

Technical Field

[0001] This invention relates to a firefighter, and more particularly to an indoor intelligent firefighter. It belongs to the field of intelligent fire protection. Background Technology

[0002] Existing technologies include several solutions for fire suppression via indoor ceiling spraying or outdoor drone spraying. The industry's goal has always been to detect fires as soon as they occur and quickly reach them using on-site fire suppression devices to extinguish them in their initial stages. In addition, fire protection also involves emergency situations indoors, such as when someone is partially trapped by an object or unable to break free, and there is a lack of necessary means at the scene. This includes alarms for non-fire-related situations involving personal safety, especially when someone is indoors and unable to call for help. In such cases, having indoor fire suppression systems is crucial.

[0003] Using sensors to inform the operation of the roof sprinkler system is a conventional approach. However, existing sensors are installed inside the wall, which can achieve three-dimensional scanning, but because the wall is cubic, there are still areas that cannot be scanned. Therefore, how to expand the scanning area has become an urgent technical problem to be solved. Summary of the Invention

[0004] 1. Core Technology First, the indoor sprinkler system (which can be retained if conditions permit) was modified into a multi-sensor system, and at least one independent sensor was placed in other locations indoors. This enabled full-airspace smoke image recognition and temperature anomaly scanning. The concept of multi-directional sprinkler system was shifted to multi-directional sensing system, and the system was modified into a mobile sensing device with selectable placement. The overhead sensor and the independent sensor were combined to form a sensing network.

[0005] Secondly, by setting up indoor firefighters who communicate with multiple spatial coordinates formed by the sensor network, indoor path planning can be achieved based on the current location, or the path can be manually drawn on the display screen via a remote intelligent mobile device, thus serving the dual purpose of notification and path planning.

[0006] Third, for the solution of automatically waking up and opening the lock of the closed indoor door in an emergency, entry into the room is achieved.

[0007] 2. Detailed Technical Specifications This invention provides an indoor intelligent firefighter, comprising a firefighter body, a multi-head monitoring system mounted on the indoor ceiling that is rotatable and whose tilt angle is adjustable for each head, a sensor network consisting of at least one independent sensor system that can be arbitrarily installed in other locations within the room, and a smart door lock for the room. The multi-monitoring system and at least one independent sensor system in the sensor network form a network node, whose projected coordinates on the ground communicate with the firefighter's main body for path planning. When a fire alarm signal is generated in the room where the projection coordinates of at least one of the network nodes are located, the network node communicating with the smart lock corresponding to the room determines the current lock status. If and only if the lock status is locked, the smart lock is opened so that the firefighter body communicating with the network node corresponding to the projection coordinates can push open the door and enter the room where the projection coordinates are located during the movement. Otherwise, the firefighter body is triggered to start according to the distress signal sent by the network node corresponding to the projection coordinates to the firefighter body, and moves to a position at a preset distance from the projection coordinates to carry out fire rescue according to the path.

[0008] Optionally, the firefighters can be multiple (i.e., two or more), each of which communicates only with the preset ground projection coordinates (to achieve zoned action), and can be any shape such as a robot dog, humanoid, cylindrical, or cubic, capable of moving manually or automatically, or following humans (e.g., through image recognition perception technology) to the indoor fire location to carry out firefighting activities.

[0009] Optionally, the multi-head monitoring system and at least one independent sensing system each include at least one composite probe composed of a visible light camera and an infrared thermometer and / or an infrared camera, and a composite probe power supply system. The composite probe contains a chip, and the chip stores an indoor floor plan view set. The multi-head monitoring system comprises a bearing-type base detachably connected to the wall and ceiling, a hollow sleeve rod, a rack rod, a composite probe bracket fixedly connected to the hollow sleeve rod, a rotary motor, and a lifting motor. The bearing-type base includes an inner ring and an outer ring. The inner ring is detachably and fixedly connected to the top of the wall. The hollow sleeve rod extends detachably from the bottom of the outer ring. The rack rod is vertically and flexibly fitted inside the hollow sleeve rod, and a hanging spring is fixedly connected between the top of the rack rod and the inner wall of the hollow sleeve rod. The composite probe bracket includes a ring and a composite probe seat that can be tilted and rotated on the ring. The ring is fixedly connected to the hollow sleeve rod by a spring or connecting rod (which can be detachable). The outer surface of the rack rod is provided with one more straight rack than the number of composite probes. Each composite probe holder has an arc-shaped rack near the rotation axis of the hollow sleeve rod. The arc-shaped rack meshes with the corresponding straight rack, and the remaining straight rack meshes with the second drive gear at the output end of the lifting motor. The rack rod is driven to move up and down by the lifting motor, thereby adjusting the pitch angle of all the arc-shaped racks as a whole during the movement. The lifting motor bracket is detachably mounted on the hollow sleeve rod, and the lifting motor is detachably and fixedly placed on the lifting motor bracket. The lifting motor can switch between forward and reverse rotation, thereby enabling all the composite probes to move up and down. The composite probe mounted on the probe holder can swing back and forth around a set pitch angle. The rotary motor is detachably fixed by a rotary motor bracket that is detachably connected to the wall top. The side of the outer ring has a circumferential gear that meshes with the first drive gear on the output end of the rotary motor. Thus, the rotary motor drives the outer ring to rotate, causing the hollow sleeve rod, rack rod, composite probe bracket, lifting motor bracket, and lifting motor to rotate as a whole. Rotary encoders are installed at the output ends of the rotary motor and the lifting motor. The horizontal coordinate position of each composite probe holder is obtained by acquiring the rotary encoder signals through the control and analysis circuit.

[0010] It's easy to understand that the spring is not only a suspension and fixing component of the rack and pinion, but also a component that compresses and extends due to the switching between forward and reverse rotation. Without the swing mechanism, the lifting motor would need to continuously output static torque to overcome the spring's elasticity and maintain a stable pitch angle, which could lead to overheating. Switching between forward and reverse rotation keeps the motor running, preventing overheating and expanding the pitch monitoring range. Because multiple composite probes are used, large-angle rotation is unnecessary, and the cables connecting the composite probes to the power system do not suffer excessive twisting or tangling due to large-angle rotation.

[0011] Preferably, both the lifting motor and the rotating motor are servo motors that are wirelessly remotely controlled by the control and analysis circuit to switch between rotation and reversal, and both the lifting motor and the rotating motor operate alternately in a preset periodic manner.

[0012] Optionally, the cycle is 5 min to 1 h.

[0013] Optionally, the chip processes and analyzes visible light images and temperature and / or thermal imaging images in real time to identify real-time fire situations.

[0014] Optionally, the independent sensing system can be placed anywhere indoors and / or installed in indoor furnishings (e.g., furniture, such as beds, cabinets, cupboards, tables, etc.; office supplies, such as desks, filing cabinets, etc.), or it can be placed directly inside the firefighter's room, possessing the ability to interact with fixed sensors configured in each room. More preferably, the independent sensing system can also be equipped with precise three-dimensional positioning capabilities, achieved through communication with a satellite positioning system or a remote server.

[0015] Optionally, the identification method includes inputting real-time visible light images and thermal images into a pre-trained convolutional neural network (CNN), and / or inputting real-time visible light images into a pre-trained convolutional neural network and referencing real-time temperature measurements from an infrared thermometer to identify the fire.

[0016] Optionally, if either the real-time visible light image or the thermal image is input into the pre-trained convolutional neural network and a fire is identified, a suspicious situation is reported to the firefighter main unit; if both are identified, the firefighter main unit is immediately activated; and / or, If a pre-trained convolutional neural network detects a fire when a real-time visible light image is input, but the real-time temperature measurement does not exceed a threshold, it reports a suspicious situation to the main firefighter unit. If the pre-trained convolutional neural network does not detect a fire when a visible light image is input, but the real-time temperature measurement exceeds a threshold, or if the pre-trained convolutional neural network detects a fire when a visible light image is input, and the real-time temperature measurement exceeds a threshold, the main firefighter unit is immediately activated. When activating the firefighter body, the internal processor of the firefighter performs path planning based on the indoor floor plan transmitted by the chip of the network node in the room that generates the fire signal.

[0017] Optionally, the path planning includes the Floyd algorithm, Dijkstra's algorithm, and A... Algorithm, D Algorithm, any one of the ant colony algorithms.

[0018] Optionally, when multiple network nodes enter the main step of activating the firefighter, the corresponding multiple ground projection coordinates are transmitted to the firefighter. The internal processor of the firefighter performs path planning, specifically including linking the multiple coordinates end to end to form line segments, triangles, or polygons. In the case of line segments, the path planning is performed with the midpoint coordinates as the target coordinates. For the triangle case, the coordinates of its perpendicular point are taken as the target coordinates for path planning; For polygons, the polygon is divided into a minimum number of triangles, and the orthocenter of each triangle is found. If all orthocenters coincide, the coordinates of the orthocenter are the target coordinates. If there are non-coincident orthocenters, the orthocenters are connected sequentially. If a line segment is formed, the midpoint is taken. If a triangle is formed, the orthocenter is taken. If a polygon is formed, the polygon is divided into a minimum number of triangles until the orthocenter of the triangle formed by the midpoints of the line segments or the orthocenters is obtained. The coordinates of the midpoint or orthocenter are then used as the target coordinates for path planning.

[0019] Optionally, the firefighter is equipped with a high-definition camera to observe the real-time situation in the direction of travel. The firefighter carries extinguishing agents, such as dry powder or gaseous extinguishing agents. Upon activation of the firefighter and / or in response to any suspicious situations, the corresponding network nodes communicate with a remote smart mobile device. Users can observe the real-time situation on this remote smart mobile device and can reversibly switch the display to show an indoor floor plan marked with the locations of the corresponding network nodes. Users can also manually draw paths on the indoor floor plan or have the firefighter plan their own paths for movement. The movement method is as follows: S1 is based on the firefighter's current orientation as the reference direction, forming a reference ray in the diagram. S2 calculates the acute angle between the reference ray and the manually drawn path or the path planned by the firefighter at fixed intervals in the current moving position. If the acute angle is within the preset angle range, it continues to walk in the current direction. Otherwise, it changes the direction so that the acute angle is within the preset angle range, and continues to walk in the changed direction.

[0020] Optionally, the path can be drawn manually or planned by the firefighters themselves as a broken line formed by connecting at least one straight segment. The preset angle range is ±0.5°. When the distance between the current position coordinates and the target coordinates is the preset distance, the target position is considered to have been reached and the movement stops.

[0021] Optionally, the fixed time period is 0.2-0.5s, and the preset distance is 0.5-1m.

[0022] Optionally, the firefighter has a microphone for recognizing distress calls from people inside the building and for communicating with a remote fire station.

[0023] 3. Beneficial effects First, install multi-head sensors on the indoor ceiling, in conjunction with an independent sensing system, to create a sensor network that can identify the three-dimensional temperature field and fire situation in the entire space. Second, by setting up the main body of the indoor firefighter and communicating with the sensor network, it can realize the activation of the main body of the firefighter in case of suspicious situations, and reach the target location and the surrounding area according to the planned path to put out the fire. Third, it can also remotely monitor the environment around firefighters via smart mobile devices, and draw paths directly on the smart mobile devices so that firefighters can adaptively correct their direction and walk. The network node corresponding to the fire room controls the smart lock to open the door of the fire room, so that the firefighters can push open the door and enter to extinguish the fire. Fourth, by obtaining indoor distress call information through voiceprint training, firefighters can be activated to remotely request fire assistance in non-fire situations. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the main configuration of an indoor intelligent firefighter according to Embodiment 1 of the present invention. Figure 2 This is a schematic diagram of the multi-head monitoring system according to Embodiment 2 of the present invention. a is a front view, b is a top view, and c is a front view of the composite probe. Figure 3 A flowchart for fire identification and firefighting strategies. Figure 4a This is an illustration of a live feed of the firefighters' main camera, displayed on a remote user's smartphone. Figure 4b This diagram illustrates how a remote user's smartphone can reversibly switch to an indoor floor plan view, and includes an example of a path drawn manually with a finger. Figure 5 A schematic diagram of an adaptive correction scheme for the movement direction of the firefighters along the path is manually drawn. Figure 6 This diagram illustrates four different target point setting methods when multiple network nodes enter and activate the main body of the firefighter: a represents a line segment, b represents a triangle, c represents a quadrilateral, and d represents a pentagon. Detailed Implementation

[0025] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. Those skilled in the art will understand that this description is exemplary and that the present invention is not limited to the specific embodiments described herein.

[0026] Example 1 This embodiment will describe an indoor intelligent firefighter, such as... Figure 1 In a bedroom, it includes a cylindrical firefighter body equipped with a high-definition camera, a multi-head monitoring system that can rotate and has an adjustable tilt angle for each head mounted on the ceiling, a sensor network consisting of an independent sensor system placed on each of the bedside tables on both sides of the room and two independent sensors placed on the wardrobe, and a smart door lock for the room.

[0027] The multi-monitoring system and two independent sensing systems in the sensor network form a network node, which in Figure 1 The projected coordinates of the bedroom floor communicate with the main body of the firefighter for path planning. When a fire alarm signal is generated in the room corresponding to the projection coordinates of at least one of the network nodes, the network node communicating with the smart lock corresponding to that room determines the current lock status. If and only if the lock status is locked, the smart lock is controlled to open, allowing the firefighter body communicating with the network node corresponding to the projection coordinates to push open the door and enter the room where the projection coordinates are located during movement. Otherwise, the firefighter body is triggered to start according to the distress signal sent by the network node corresponding to the projection coordinates to the firefighter body, and moves to a position 0.5m away from the projection coordinates to extinguish the fire.

[0028] In other bedrooms, the living room, and the kitchen, multi-monitoring systems can be installed as network nodes, along with firefighters. Figure 1 Any one of the firefighters in the room that is only connected to the corresponding multi-monitoring system (i.e. Figure 1 The system communicates the projected coordinates of the floors in the bedroom, other bedrooms, living room, and kitchen to enable zoned movement. It can take the form of a robot dog, humanoid, cylindrical, or cubic shape, and can be moved manually or automatically, or follow humans to the location of an indoor fire to extinguish it. Each firefighter is equipped with a microphone to recognize cries for help from people inside the room (specifically, this can be achieved by training a recognition model using collected voiceprint data) and communicates with a remote fire station. This is particularly useful for fire rescue operations outside of fire emergencies.

[0029] Example 2 This embodiment describes the multi-camera monitoring system and the independent sensing system in Embodiment 1. Both the multi-camera monitoring system and at least one independent sensing system include at least one composite probe consisting of a visible light camera and an infrared thermometer, and a power supply system for the composite probe. The composite probe contains a chip, and the chip stores an indoor floor plan view set.

[0030] The system specifically includes: 6 ( ) installed on the top plate Figure 2 a) An example of a composite probe among composite probes with equally spaced distribution.

[0031] The composite probe has a power supply system and also includes a bearing-type base detachably connected to the wall and ceiling, a hollow sleeve rod, a rack rod, a composite probe bracket fixedly connected to the hollow sleeve rod, a rotary motor, and a lifting motor. The composite probe contains a chip (…). Figure 2 a and Figure 2(not shown in b), it communicates with a remote server, and the chip stores an indoor floor plan atlas. The composite probe is equipped with a remote control to remotely power off the composite probe when not needed, and has a timed automatic power-on function, such as automatically powering on at midnight.

[0032] The bearing-type base includes an inner ring and an outer ring. The inner ring is detachably and fixedly connected to the top of the wall. The hollow sleeve rod extends detachably from the bottom of the outer ring. The rack rod is vertically and flexibly fitted inside the hollow sleeve rod, and a hanging spring is fixedly connected between the top of the rack rod and the inner wall of the hollow sleeve rod. The specific fixing method is as follows... Figure 2 As shown in diagram a, two straight rods are used. The circled area in the diagram can be hooked to the top coil of the spring, and the same method can be used to hook it onto the inner wall of the hollow sleeve rod.

[0033] The composite probe holder includes a ring ( Figure 2 Figure a shows a partial cross-section near one of the composite probe mounts and a small section in the frontal view, as well as the pitch-rotatable composite probe mount on the ring, which is fixedly connected to the hollow sleeve rod via a connecting rod (specifically, fixedly connected to...). Figure 2 (On the slope of the conical funnel of a hollow sleeve rod).

[0034] The outer surface of the rack rod is provided with seven straight racks. Each composite probe holder has an arc-shaped rack near the rotation axis of the hollow sleeve rod. The arc-shaped rack meshes with the corresponding straight rack. The remaining straight rack meshes with the second drive gear at the output end of the lifting motor. The rack rod is thus driven by the lifting motor. Figure 2 The lifting motion indicated by the double straight arrows in section a drives all the arc-shaped racks to adjust their pitch angles during the movement. A lifting motor bracket is detachably mounted on the hollow sleeve rod, and the lifting motor is detachably and fixedly placed on the bracket. The lifting motor can switch between forward and reverse rotation, allowing all the composite probes mounted on the composite probe holders to move around the set pitch angle. Figure 2 As shown in figure a, the double-arrow indicates the back-and-forth swinging motion. The rotary motor is detachably fixed by a rotary motor bracket that is detachably connected to the wall top. The outer ring has a circumferential gear on its side that meshes with the first driving gear on the output end of the rotary motor. Thus, the rotary motor drives the outer ring to rotate, causing the hollow sleeve rod, rack rod, composite probe bracket, lifting motor bracket, and lifting motor to rotate as a whole. Rotary encoders are installed at the output ends of both the rotary motor and the lifting motor. Figure 2 (a not shown) The horizontal coordinates of each composite probe mount are obtained by acquiring rotary encoder signals through control and analysis circuits.

[0035] like Figure 2The electrode signal lines of the rotary motor and lifting motor shown in Figure a, and the signal lines of the rotary encoder ( Figure 2 (a not shown) are all connected to the wireless / battery module in the base to enable power supply and wireless acquisition of rotary encoder signals between the control and analysis circuits.

[0036] like Figure 2 As shown in a and 2b, the composite probes are connected by cables. Therefore, the six composite probes are connected to six batteries in the power supply box of the composite probe power supply system via six cables to provide power to the signal lines, enabling the composite probes to operate and move controllably. In this embodiment, the rotary motor only needs to rotate the composite probe holder 60° (once every 30 minutes). Therefore, the cables do not need to be too long, and the cables will not be twisted or overstretched due to larger rotation angles (exceeding 60°).

[0037] Example 3 This embodiment illustrates the firefighters' identification of a fire and route planning in Embodiment 1. In such... Figure 3 The chamfered dashed box shows the process of the chip identifying fire situations. For fire situations, there are two classification levels: entering manual decision-making and suspicious cases. Each set of composite probes is equipped with a chip. When multiple sets of visible light cameras and infrared thermometers are identified, not identified, and infrared temperature exceeding or not exceeding the threshold respectively through pre-trained CNN, a suspicious case is identified, and the manual decision-making step is initiated. Figure 3 The criteria for identification exceeding the threshold, non-identification exceeding the threshold, and identification not exceeding the threshold are respectively classified into three cases: α, β, and γ. Among them, α and β proceed to the manual decision-making step, while γ is a suspicious case.

[0038] See also Figure 3 ,when Figure 2 a, Figure 2 When the multi-monitoring system shown in b detects a fire indoors, it categorizes the situation into two levels: activating the main firefighting unit and handling suspicious cases. The firefighting unit's internal processor then uses the indoor floor plan transmitted by the chip corresponding to the network node identifying the fire to perform actions based on A. The algorithm's path planning and reporting to the firefighters.

[0039] The corresponding network nodes that report suspicious situations to the firefighters all communicate with remote users' smartphones, allowing users to observe the real-time situation on their smartphones. Figure 4a (Smoke is seen coming out of the doorway at point D), and the display can be reversibly switched to show the indoor floor plan marked with the location of the corresponding network node, and paths can be manually drawn on the indoor floor plan. Figure 4b (As shown), for the firefighters to move.

[0040] Figure 4b The path is given by a person's finger, manually drawn from the entrance door to the target displayed along the way (marked by the chip of the corresponding room's multi-monitoring system). Figure 1 The method for moving the cylindrical firefighter is as follows: Figure 5 As shown, on the manually drawn path, S1 is based on the firefighter's current orientation as the reference direction, forming a reference ray in the diagram. S2 calculates the acute angle γ between the reference ray and the manually drawn path every 0.2s in the current moving position. If the acute angle is within ±0.3°, it continues to walk in the current orientation; otherwise, it changes the orientation so that the acute angle is within ±0.3°, and continues to walk in the changed orientation.

[0041] The algorithm described in the above steps makes Figure 5 When the firefighter moves to point A or point B, because the reference ray is not aligned with the direction of line segments AB and BC by more than ±0.3°, it will adaptively rotate so that the reference ray stops rotating when it is within ±0.3° of the direction of line segments AB and BC, and then continue moving.

[0042] When the aforementioned network nodes initiate the firefighter's main action, they transmit the corresponding ground projection coordinates to the firefighter. The firefighter's internal processor then performs path planning, such as... Figure 6 As shown, specifically, when multiple coordinates are linked end-to-end to form a line segment, the coordinates of the midpoint are taken as the coordinates of the target point. Figure 6 a).

[0043] For the triangle case, the coordinates of its perpendicular point are taken as the target coordinates for path planning. Figure 6 b) For polygonal cases, such as Figure 6 Taking quadrilateral c as an example, the polygon is divided into the minimum number of two triangles, and the orthocenter of each triangle is found. If all orthocenters coincide, the coordinates of the orthocenter are the target coordinates. If there are non-coincident orthocenters ( Figure 6 c), then connect the orthocenters in sequence. If a line segment is formed, take the midpoint. For a pentagon, divide it into at least three triangles. Connect the orthocenters in sequence. If a triangle is formed, continue to take its orthocenter. That is, use the coordinates of the midpoint and the orthocenter as the target coordinates to perform path planning.

[0044] The present invention has been described in detail above through specific embodiments. However, this description is exemplary, and those skilled in the art can make various modifications and changes to it. As long as they do not depart from the spirit and purpose of the present invention, all such modifications and changes should fall within the protection scope of the present invention, which is defined by the appended claims.

Claims

1. An indoor intelligent firefighter, characterized in that, The system includes a firefighter's main body, a multi-head monitoring system mounted on the indoor ceiling that is rotatable and has adjustable tilt angles for each head, a sensor network consisting of at least one independent sensor system that can be optionally installed in other locations throughout the room, and a smart door lock for the room. The multi-monitoring system and at least one independent sensor system in the sensor network form a network node, whose projected coordinates on the ground communicate with the firefighter's main body for path planning. When a fire alarm signal is generated in the room where the projection coordinates of at least one of the network nodes are located, the network node communicating with the smart lock corresponding to the room determines the current lock status. If and only if the lock status is locked, the smart lock is opened so that the firefighter body communicating with the network node corresponding to the projection coordinates can push open the door and enter the room where the projection coordinates are located during the movement. Otherwise, the firefighter body is triggered to start according to the distress signal sent by the network node corresponding to the projection coordinates to the firefighter body, and moves to a position at a preset distance from the projection coordinates to carry out fire rescue according to the path.

2. The indoor intelligent firefighter according to claim 1, characterized in that, The firefighters consist of multiple entities, each of which communicates only with the preset projection coordinates on the ground. They can take any shape, such as a robot dog, humanoid, cylindrical, or cubic, and can move manually or automatically, or follow humans to the location of an indoor fire to carry out firefighting activities.

3. The indoor intelligent firefighter according to claim 1, characterized in that, The multi-head monitoring system and at least one independent sensing system each include at least one composite probe consisting of a visible light camera and an infrared thermometer and / or an infrared camera, and a power supply system for the composite probe. The composite probe contains a chip, and the chip stores an indoor floor plan diagram set. The multi-head monitoring system comprises a bearing-type base detachably connected to the wall and ceiling, a hollow sleeve rod, a rack rod, a composite probe bracket fixedly connected to the hollow sleeve rod, a rotary motor, and a lifting motor. The bearing-type base includes an inner ring and an outer ring. The inner ring is detachably and fixedly connected to the top of the wall. The hollow sleeve rod extends detachably from the bottom of the outer ring. The rack rod is vertically and flexibly fitted inside the hollow sleeve rod, and a hanging spring is fixedly connected between the top of the rack rod and the inner wall of the hollow sleeve rod. The composite probe bracket includes a ring and a composite probe seat that can be tilted and rotated on the ring. The ring is fixedly connected to the hollow sleeve rod by a spring or a connecting rod. The outer surface of the rack rod is provided with one more straight rack than the number of composite probes. Each composite probe holder has an arc-shaped rack near the rotation axis of the hollow sleeve rod. The arc-shaped rack meshes with the corresponding straight rack, and the remaining straight rack meshes with the second drive gear at the output end of the lifting motor. The rack rod is driven to move up and down by the lifting motor, thereby adjusting the pitch angle of all the arc-shaped racks as a whole during the movement. The lifting motor bracket is detachably mounted on the hollow sleeve rod, and the lifting motor is detachably and fixedly placed on the lifting motor bracket. The lifting motor can switch between forward and reverse rotation, thereby enabling all the composite probes to move up and down. The composite probe mounted on the probe holder can swing back and forth around a set pitch angle. The rotary motor is detachably fixed by a rotary motor bracket that is detachably connected to the wall top. The side of the outer ring has a circumferential gear that meshes with the first drive gear on the output end of the rotary motor. Thus, the rotary motor drives the outer ring to rotate, causing the hollow sleeve rod, rack rod, composite probe bracket, lifting motor bracket, and lifting motor to rotate as a whole. Rotary encoders are installed at the output ends of the rotary motor and the lifting motor. The horizontal coordinate position of each composite probe holder is obtained by acquiring the rotary encoder signal through the control and analysis circuit. The independent sensing system is designed to be placed anywhere indoors, and / or installed in indoor furnishings, and / or directly integrated into the firefighter's interior.

4. The indoor intelligent firefighter according to claim 3, characterized in that, Both the lifting motor and the rotating motor are servo motors that are wirelessly remotely controlled by the control and analysis circuit to switch between rotation and reversal. Both the lifting motor and the rotating motor are preset to work alternately periodically, with a cycle of 5 minutes to 1 hour.

5. The indoor intelligent firefighter according to claim 3 or 4, characterized in that, The chip processes and analyzes visible light images and temperature and / or thermal images in real time to identify real-time fire situations.

6. The indoor intelligent firefighter according to claim 5, characterized in that, The identification method includes inputting real-time visible light images and thermal images into a pre-trained convolutional neural network (CNN), and / or inputting real-time visible light images into a pre-trained convolutional neural network and referencing real-time temperature measurements from an infrared thermometer to identify the fire.

7. The indoor intelligent firefighter according to claim 6, characterized in that, When a real-time visible light image and a thermal image are input into a pre-trained convolutional neural network, if one identifies a fire, a suspicious situation is reported to the firefighter's main unit; if both are identified, the firefighter's main unit is immediately activated; and / or, If a pre-trained convolutional neural network detects a fire when a real-time visible light image is input, but the real-time temperature measurement does not exceed a threshold, it reports a suspicious situation to the main firefighter unit. If the pre-trained convolutional neural network does not detect a fire when a visible light image is input, but the real-time temperature measurement exceeds a threshold, or if the pre-trained convolutional neural network detects a fire when a visible light image is input, and the real-time temperature measurement exceeds a threshold, the main firefighter unit is immediately activated. When activating the firefighter body, the internal processor of the firefighter performs path planning based on the indoor floor plan transmitted by the chip of the network node in the room that generates the fire signal.

8. The indoor intelligent firefighter according to claim 7, characterized in that, The path planning includes the Floyd algorithm, Dijkstra's algorithm, and A... Algorithm, D Algorithm, any one of the ant colony algorithms.

9. The indoor intelligent firefighter according to claim 1, characterized in that, When multiple network nodes enter the main step of activating the firefighter, they transmit the corresponding multiple ground projection coordinates to the firefighter. The firefighter's internal processor performs path planning, which specifically includes linking the multiple coordinates end to end to form line segments, triangles, or polygons. In the case of line segments, the path planning is performed with the midpoint coordinates as the target coordinates. For the triangle case, the coordinates of its perpendicular point are taken as the target coordinates for path planning; For polygons, the polygon is divided into a minimum number of triangles, and the orthocenter of each triangle is found. If all orthocenters coincide, the coordinates of the orthocenter are the target coordinates. If there are non-coincident orthocenters, the orthocenters are connected sequentially. If a line segment is formed, the midpoint is taken. If a triangle is formed, the orthocenter is taken. If a polygon is formed, the polygon is divided into a minimum number of triangles until the orthocenter of the triangle formed by the midpoints of the line segments or the orthocenters is obtained. The coordinates of the midpoint or orthocenter are then used as the target coordinates for path planning.

10. The indoor intelligent firefighter according to claim 1, characterized in that, The firefighter is equipped with a high-definition camera to observe the real-time situation in its direction of movement. Upon activating the firefighter and / or in response to any suspicious circumstances, the corresponding network nodes communicate with a remote smart mobile device. Users can observe the real-time situation on this remote smart mobile device and can reversibly switch between displaying an indoor floor plan marked with the locations of the corresponding network nodes. Users can also manually draw paths on the indoor floor plan or have the firefighter plan its own paths for movement. The movement method is as follows: S1 is based on the firefighter's current orientation as the reference direction, forming a reference ray in the diagram. S2 calculates the acute angle between the reference ray and the manually drawn path or the path planned by the firefighter at fixed intervals in the current moving position. If the acute angle is within a preset angle range, the firefighter continues to move in the current direction; otherwise, the firefighter changes direction so that the acute angle is within the preset angle range and continues to move in the changed direction. The manually drawn path is a broken line formed by connecting at least one straight line segment, and the preset angle range is ±0.5°. When the distance between the current position coordinates and the target coordinates is the preset distance, the target position is considered to have been reached, and the firefighter stops moving. The fixed interval is 0.2-0.5s, and the preset distance is 0.5-1m. The firefighter also has a microphone to identify the distress calls of people indoors and to communicate with a remote fire station.

Images