A monitoring type lifesaving and rescue unmanned aerial vehicle capable of floating on water and a rescue method thereof

Abstract

The application relates to the field of water rescue, and particularly discloses a monitoring type lifesaving and rescuing unmanned plane capable of floating on water and waiting and a lifesaving and rescuing method thereof, which comprises a waterproof central machine body, four foldable machine arms symmetrically hinged to the periphery of the machine body in a cross shape, a brushless motor and a propeller installed at the tail end of the machine arm, a flexible solar photovoltaic panel covering the top of the machine body and the upper surfaces of the machine arms, and an electromagnetic hook type life buoy throwing mechanism arranged at the bottom of the machine body. The application realizes long-term unattended monitoring through floating on water and solar power supply, realizes active anti-drifting on the water surface and rapid aerial response through the machine arm winding and unwinding bimodal structure and propeller thrust control, improves search accuracy by predicting the position of a fallen person through an LSTM neural network combined with real-time hydrological data, ensures the reliability of the rescue by landing on the water surface after throwing the life buoy and performing audio and video safety confirmation, integrates the functions of hydrological monitoring and emergency rescue, and effectively shortens the rescue response time.

Description

Technical Field

[0001] This invention belongs to the field of water rescue, specifically a monitoring-type life-saving and rescue drone that can float on the water surface and standby, and its rescue method. Background Technology

[0002] Water rescue is a crucial issue in public safety, encompassing scenarios such as drowning searches, flood emergency response, and water accident rescue. Traditional water rescue primarily relies on manually operated rescue boats and jet skis. These methods are limited by response speed, personnel availability, and weather conditions, often making it difficult to reach the accident site within the critical rescue window. With the development of drone technology, land-based multi-rotor drones are increasingly being used in water search and rescue. These drones expand the search area from an aerial perspective and can carry and deliver rescue equipment. However, these drones must take off from the shore, their flight paths are limited by terrain, they require additional flight time to reach the accident area, and they cannot remain on the water for extended periods, making 24 / 7 unattended monitoring difficult.

[0003] Existing technologies include floating platforms, such as water rescue robots or simple floating drone platforms. These devices can float on the water surface for extended periods for monitoring or standby. Water rescue robots employ a boat-shaped structure and are propelled by propellers, possessing a degree of autonomous navigation capability. However, their movement speed is relatively slow, and they cannot cross obstacles on the water surface. Simple floating drone platforms are mostly passively floating structures, lacking active position correction capabilities. They are easily affected by water currents and waves, deviating from the preset monitoring area. Furthermore, they rely on battery power, resulting in limited endurance and making it difficult to support long-term continuous monitoring tasks.

[0004] In summary, the main shortcomings of existing water rescue technologies are that existing equipment cannot simultaneously accommodate long-term standby on the water and rapid aerial response, and lacks a dual-mode structural design that combines self-sustaining floating on the water with vertical take-off and landing capabilities. This results in a time window delay in rescue response, making it difficult to meet the demand for efficient and accurate rescue within the golden rescue time. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a monitoring-type life-saving and rescue drone capable of floating and remaining on the water surface, along with its rescue method. This solves the problems in existing technologies where long-term water surface standby and rapid aerial response are not simultaneously possible, and the lack of a dual-mode water-air structure leads to time window delays in rescue response.

[0006] This invention provides a monitoring-type life-saving and rescue drone capable of floating and remaining in standby mode on water, comprising:

[0007] The central body is waterproof, and a sealed cavity is formed inside it;

[0008] Four foldable arms are symmetrically hinged in a cross shape around the waterproof central body. Each foldable arm is connected to the waterproof central body via a rotating servo motor and can switch between a folded state and an unfolded state.

[0009] A brushless motor and a propeller are mounted at the end of each of the aforementioned foldable arms;

[0010] Flexible solar photovoltaic panels cover the top of the waterproof central body and the upper surface of each of the foldable arms;

[0011] An electromagnetic hook-type lifebuoy delivery mechanism is located at the bottom of the waterproof central body;

[0012] The flight control computer is built into the cavity of the waterproof central fuselage;

[0013] The hydrological monitoring sensor array is built into the cavity of the waterproof central body and is used to collect real-time hydrological data.

[0014] A dual-light gimbal is installed at the front or bottom of the waterproof central body;

[0015] The wireless communication module is built into the cavity of the waterproof central body.

[0016] Preferably, the foldable arm is attached to the side wall of the waterproof central fuselage in the retracted state, so that the overall length of the drone is less than its length in the unfolded state; the foldable arm is horizontally radial in the unfolded state.

[0017] The present invention also provides a monitoring and rescue control method, applied to the aforementioned drone, comprising the following steps:

[0018] Water surface standby steps: The UAV floats on the water surface, and the flexible solar photovoltaic panel supplies power to the hydrological monitoring sensor group and the wireless communication module; the flight control computer controls the propeller to rotate, generating thrust to correct the UAV's water surface drift.

[0019] Takeoff switching steps: When the wireless communication module receives an emergency call, the flight control computer controls the rotary servo to drive the foldable arm to switch from the retracted state to the unfolded state, starts the brushless motor and propeller, and controls the UAV to take off vertically.

[0020] Steps for predicting the location of a person who has fallen into the water: Obtain real-time hydrological data collected by the hydrological monitoring sensor group and initial location information from the emergency rescue instructions; input the real-time hydrological data and the initial location information into a pre-trained LSTM neural network model to predict the drift trajectory and estimated location of the person who has fallen into the water.

[0021] Search and rescue steps: Control the drone to fly to the estimated location, use the dual-light gimbal to search for and confirm the target of the person who fell into the water, and then control the electromagnetic hook-type lifebuoy deployment mechanism to release the lifebuoy;

[0022] Safety confirmation and return procedure: Control the drone to land on the water surface and control the foldable arm to switch to the retracted state. Use the dual-light gimbal and the wireless communication module to conduct video and voice interaction with the person who fell into the water. After confirming safety, control the drone to return with the arm retracted.

[0023] Preferably, the training process of the LSTM neural network model includes:

[0024] Collect historical water rescue data, which includes historical hydrological data, historical initial location of falling into the water, historical drift trajectory, and historical final rescue location.

[0025] The historical hydrological data and the historical initial location of falling into the water are used as input features, and the historical drift trajectory or the historical final rescue location is used as output label to train the LSTM neural network, thereby obtaining the pre-trained LSTM neural network model.

[0026] Preferably, the LSTM neural network model includes a forget gate, an input gate, an output gate, and a cell state update mechanism, and its calculation process is expressed as follows:

[0027] Forgotten Gate: ;

[0028] Input Gate: ;

[0029] Cell status update: ;

[0030] Output gate: ;

[0031] in, The input consists of the current time-series hydrological data and the initial location information of the water droplet. This is the predicted drift position information output at the current moment. Output for the forget gate. For input gate output, For output gate output, The current cell state, Candidate cell state, and These are the weight matrices and bias terms for each gate. It is the sigmoid activation function. The hyperbolic tangent activation function is used. This represents element-wise multiplication;

[0032] Before inputting the feature factor data into the LSTM neural network model, each feature factor data is first standardized; after obtaining the prediction result, destandardization is performed.

[0033] Preferably, in the water surface standby step, the specific method by which the flight control computer controls the propeller rotation to correct the UAV's water surface drift is as follows:

[0034] Acquire the real-time location information and preset anchoring location information of the drone;

[0035] Calculate the drift deviation between the real-time location information and the anchoring location information;

[0036] Based on the drift deviation and the real-time hydrological data, the rotational speed and direction of each propeller are determined to generate a resultant force opposite to the drift direction, thereby keeping the UAV at the anchored position.

[0037] Preferably, in the safety confirmation and return steps, the video and voice interaction includes: establishing a real-time communication link with the remote rescue center through the wireless communication module, transmitting the on-site video collected by the dual-light PTZ and the voice information of the person in the water to the remote rescue center in real time, and receiving voice instructions from the remote rescue center to broadcast to the person in the water.

[0038] Preferably, the electromagnetic hook-type lifebuoy deployment mechanism includes an electromagnet, a hook, and a lifebuoy. The electromagnet is electrically connected to the flight control computer, and the flight control computer controls the on / off state of the electromagnet to control the attraction and release of the hook.

[0039] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described monitoring and rescue control method.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] By setting up a waterproof central fuselage and foldable arms, combined with a dual-mode switching structure for floating standby on the water surface and flying in the air, the drone can float on the water for a long time and be on standby. After receiving an emergency rescue command, it can quickly switch to flight mode to rush to the rescue site. This effectively solves the problem of response delay caused by the need for existing land-based drones to take off from the shore, and significantly shortens the rescue response time.

[0042] By using flexible solar photovoltaic panels to cover the top of the fuselage and the upper surface of the arm, solar energy is used to continuously power the hydrological monitoring sensor group and communication module, realizing long-term unattended operation and energy self-sufficiency of the equipment, overcoming the shortcomings of insufficient endurance of existing simple floating platforms, and reducing operation and maintenance costs and energy consumption.

[0043] By controlling the propeller to rotate at low speed through the flight control computer to generate thrust, and combining real-time hydrological data with the preset anchor position to correct drift, the UAV can actively resist the influence of water flow and waves and maintain a stable anchor position while waiting on the water surface. This solves the problem that existing floating platforms lack active position correction capabilities and improves standby stability and the continuity of hydrological monitoring data.

[0044] By introducing an LSTM neural network model, real-time hydrological data and the initial location information of the person who fell into the water are input into the model to dynamically predict the drift trajectory and estimated location of the person who fell into the water. This enables the drone to accurately locate the search area according to changes in the hydrological environment, overcoming the shortcomings of existing technologies such as blind searching and lack of hydrological data support, and greatly improving the search and positioning accuracy and rescue efficiency.

[0045] After deploying the lifebuoy, this invention controls the drone to land on the water and switch to a retracted state. Through a dual-light gimbal and wireless communication module, it establishes video and voice interaction with the person in the water to achieve safe confirmation of the rescue result. This solves the problem of the lack of a post-rescue confirmation mechanism in existing technologies, ensuring the success rate of rescue and the safety of personnel. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of the overall structure of the UAV of the present invention;

[0047] Figure 2 This is a top-view structural diagram of the UAV of the present invention;

[0048] Figure 3 This is a diagram showing the relationship between the core control and sensing modules of the UAV of this invention;

[0049] Figure 4 This is a diagram showing the connection and control relationship of the foldable arm of the UAV of the present invention;

[0050] Figure 5 This is a schematic diagram illustrating the state switching of the foldable arm of the UAV of the present invention;

[0051] Figure 6 This is a detailed schematic diagram of the foldable arm of the UAV in Embodiment 6 of the present invention. Detailed Implementation

[0052] 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.

[0053] Example 1: This example provides a monitoring-type rescue drone that can float on water and remain in standby mode. For example... Figures 1-3 As shown, the drone mainly includes a waterproof central fuselage 100, four foldable arms 200, a brushless motor and propeller 300, a flexible solar photovoltaic panel 400, an electromagnetic hook-type lifebuoy delivery mechanism 500, a flight control computer 600, a hydrological monitoring sensor group 700, a dual-light gimbal 800, and a wireless communication module 900.

[0054] The waterproof central fuselage 100 is made of lightweight, high-strength waterproof material, and has a sealed cavity inside to house the flight control computer 600, the hydrological monitoring sensor group 700, the wireless communication module 900, and other electronic components.

[0055] The bottom of the waterproof central body 100 is equipped with an electromagnetic hook-type life ring deployment mechanism 500, which is used to release the life ring during rescue.

[0056] A dual-light gimbal 800 is installed at the front or bottom of the waterproof central body 100 for acquiring visible light images and thermal imaging images.

[0057] like Figure 4 As shown, four foldable arms 200 are symmetrically hinged in a cross shape around the periphery of the waterproof central fuselage 100. Each foldable arm 200 is connected to the waterproof central fuselage 100 via a rotary servo 210. The rotary servo 210 is controlled by the flight control computer 600, driving the foldable arm 200 to switch between a retracted state and an extended state, as detailed below. Figure 5 As shown.

[0058] In the retracted state, the foldable arm 200 fits against the side wall of the waterproof central fuselage 100, making the overall length of the drone shorter than that in the unfolded state. At this time, the drone has a streamlined structure, which is suitable for floating on the water surface.

[0059] In its unfolded state, the foldable arms 200 are arranged in a horizontal radial pattern, with brushless motors and propellers 300 installed at the ends of each arm providing lift, enabling the drone to take off and land vertically and fly in the air.

[0060] Flexible solar photovoltaic panels 400 cover the top of the waterproof central fuselage 100 and the upper surface of each foldable arm 200. When the UAV floats on the water, the flexible solar photovoltaic panels 400 receive solar energy and convert it into electrical energy to power the flight control computer 600, hydrological monitoring sensor group 700, dual-light gimbal 800, wireless communication module 900 and other electrical equipment, and store excess electrical energy in the built-in battery to achieve long-term unattended standby.

[0061] The flight control computer 600 is built into the cavity of the waterproof central fuselage 100. As the core control unit of the UAV, it is responsible for flight control, data processing, mode switching control and rescue process management.

[0062] The hydrological monitoring sensor group 700 is also built into the cavity of the waterproof central body 100 for collecting real-time hydrological data;

[0063] In a preferred embodiment, the hydrological monitoring sensor group 700 includes a water flow velocity sensor, a water flow direction sensor, a wind speed sensor, a wind direction sensor, and a wave height sensor. Accordingly, the real-time hydrological data includes real-time water flow velocity, real-time water flow direction, real-time wind speed, real-time wind direction, and real-time wave height.

[0064] The dual-light gimbal 800 includes a visible light camera and a thermal imaging camera. The visible light camera is used to capture visible light images of the person who has fallen into the water, while the thermal imaging camera is used to identify the location of the person by body temperature characteristics at night or in low visibility conditions. The dual-light gimbal 800 is connected to the waterproof central body 100 via a gimbal motor, which allows for angle adjustment in the horizontal and vertical directions to expand the search range.

[0065] The wireless communication module 900 is built into the cavity of the waterproof central body 100. It supports multiple communication methods such as 4G, 5G or satellite communication, and is used to interact with the remote rescue center, receive emergency rescue instructions and transmit on-site video, audio and drone status information.

[0066] Example 2: Based on Example 1, this example further defines the electromagnetic hook-type lifebuoy delivery mechanism 500;

[0067] The electromagnetic hook-type lifebuoy deployment mechanism 500 includes an electromagnet 510, a hook 520, and a lifebuoy 530. The electromagnet 510 is electrically connected to the flight control computer 600, and the hook 520 is mechanically connected to the armature of the electromagnet 510.

[0068] In standby mode, the electromagnet 510 is energized to generate magnetic force, attracting the armature and keeping the hook 520 in a closed state, thereby attaching the life ring 530.

[0069] When the lifebuoy needs to be deployed, the flight control computer 600 controls the electromagnet 510 to lose power, the magnetic force disappears, the hook 520 opens under the action of gravity, and the lifebuoy 530 is released.

[0070] Example 3: This example provides a monitoring and rescue control method based on the above-mentioned UAV, which is executed by the flight control computer 600; the method includes the following steps:

[0071] Step S1: Standby at the water surface

[0072] The drone floats on the water surface, and the flexible solar photovoltaic panel 400 powers the hydrological monitoring sensor group 700 and the wireless communication module 900, enabling unattended standby. During this period, the hydrological monitoring sensor group 700 continuously collects real-time hydrological data and uploads it to the remote rescue center through the wireless communication module 900 for aquatic environment monitoring.

[0073] Meanwhile, the flight control computer 600 controls the propellers to rotate at low speed, generating thrust to correct the drone's water surface drift. Specifically, the flight control computer 600 acquires the drone's real-time position information and preset anchoring position information, calculates the drift deviation between the real-time position information and the anchoring position information, and then determines the rotation speed and direction of each propeller based on the drift deviation and real-time hydrological data to generate a resultant force opposite to the drift direction, keeping the drone near the anchoring position and achieving autonomous water surface anchoring.

[0074] Step S2: Takeoff Switch

[0075] When the wireless communication module 900 receives an emergency rescue command from the remote rescue center, the flight control computer 600 controls the rotary servo 210 to drive the foldable arm 200 from the retracted state to the unfolded state, and then starts the brushless motor and propeller 300 to control the UAV to take off vertically and fly towards the waters where the incident occurred.

[0076] Step S3: Predict the location of the person who fell into the water

[0077] The flight control computer 600 acquires real-time hydrological data collected by the hydrological monitoring sensor group 700 and the initial location information of the person who fell into the water contained in the emergency rescue instructions (such as the location provided in the emergency call, mobile phone positioning, etc.); it inputs the real-time hydrological data and initial location information into a pre-trained LSTM neural network model to predict the drift trajectory and estimated location of the person who fell into the water.

[0078] The training process of the LSTM neural network model is as follows: First, historical water rescue data is collected, including historical hydrological data, historical initial position of falling into the water, historical drift trajectory, and historical final rescue position; then, historical hydrological data and historical initial position of falling into the water are used as input features, and historical drift trajectory or historical final rescue position is used as output label to train the LSTM neural network and obtain a pre-trained LSTM neural network model.

[0079] In this embodiment, the LSTM neural network model includes a forget gate, an input gate, an output gate, and a cell state update mechanism, and its calculation process is represented as follows:

[0080] Forgotten Gate: ;

[0081] Input Gate: ;

[0082] Cell status update: ;

[0083] Output gate: ;

[0084] in, The input consists of the current time-series hydrological data and the initial location information of the water droplet. This is the predicted drift position information output at the current moment. Output for the forget gate. For input gate output, For output gate output, The current cell state, Candidate cell state, and These are the weight matrices and bias terms for each gate. It is the sigmoid activation function. The hyperbolic tangent activation function is used. This indicates element-wise multiplication.

[0085] By using the aforementioned LSTM neural network model, combined with real-time hydrological data and the initial location of the person who fell into the water, it is possible to dynamically predict the drifting location of the person at different points in time, thereby significantly improving the search and positioning accuracy.

[0086] In this embodiment, the feature factors input to the LSTM neural network model include time (month, day, hour, minute), water flow direction (angle), water flow velocity (meters / second), and temperature (degrees Celsius). Before inputting the above feature factor data into the LSTM neural network model, each feature factor data is first standardized to eliminate the influence between different units, which facilitates model training and inference; after obtaining the prediction results, destandardization is performed to restore them to the actual physical units.

[0087] The input data is constructed using a sliding window approach, with a sliding window size of 1. Let the time series data be... The sequence of water flow directions is The water flow velocity sequence is The temperature sequence is ;

[0088] The input data format is shown in the table below: Table 1 Training Sample Construction Format (Sliding Window = 1)

[0089] Sample number P (time) Q (water flow direction) R (water flow velocity) S (temperature) 1 2 3 ... ... ... ... ... n

[0090] Table 2 Input data format for the forecast period

[0091] Forecast period P (time) Q (water flow direction) R (water flow velocity) S (temperature) Next period

[0092] Step S4: Search and Rescue

[0093] The flight control computer 600 controls the drone to fly to the estimated location and activates the dual-light gimbal 800 to search. The dual-light gimbal 800 simultaneously acquires visible light images and thermal imaging images. The flight control computer 600 performs fusion analysis on the acquired images to identify and confirm the target of the person in the water. After the person in the water is identified, the flight control computer 600 controls the electromagnetic hook-type lifebuoy delivery mechanism 500 to release the lifebuoy 530, so that the lifebuoy 530 lands on the water surface near the person in the water for the person to grab.

[0094] Step S5: Safety Confirmation and Return

[0095] After the lifebuoy is deployed, the flight control computer 600 controls the drone to land on the water and controls the foldable arms 200 to switch to the retracted state, allowing the drone to float on the water in a compact form. Through the dual-light gimbal 800 and the wireless communication module 900, a real-time communication link is established with the remote rescue center. The on-site video collected by the dual-light gimbal 800 and the voice information of the person in the water are transmitted to the remote rescue center in real time. The drone also receives voice commands from the remote rescue center and broadcasts them to the person in the water, realizing video and voice interaction with the person in the water and confirming that the person is safe. After confirming that everything is in order, the flight control computer 600 controls the drone to return with the arms retracted.

[0096] Example 4: This example further optimizes the water surface drift correction method based on Example 3. When the flight control computer 600 determines the rotational speed and direction of each propeller based on the drift deviation and real-time hydrological data, it uses a PID control algorithm to calculate the thrust vector required by each propeller. Specifically, the flight control computer 600 uses the drift deviation as the input of the PID controller to calculate the magnitude and direction of the resultant force required to counteract the drift. Then, combined with information such as water flow velocity, water flow direction, wind speed, and wind direction in the real-time hydrological data, it calculates and distributes the thrust of each propeller to ensure that the UAV can still stably maintain its anchor position under complex hydrological conditions.

[0097] Example 5: This example further optimizes the safety confirmation steps based on Example 3. After the drone lands on the water, the flight control computer 600 automatically activates the tracking mode of the dual-light gimbal 800 to continuously track the location of the person in the water, ensuring that the person remains within the video monitoring range. At the same time, the flight control computer 600 transmits the vital signs information of the person in the water (such as body temperature distribution and movement status in thermal imaging images) in real time through the wireless communication module 900, allowing the remote rescue center to assess the person's physical condition. Once the remote rescue center confirms that the person in the water has been rescued by other rescue forces or has reached the shore on their own, the flight control computer 600 controls the drone to return to base.

[0098] Example 6: This example, based on Example 1, adds a specific structure to the foldable arm, such as... Figure 6 As shown, the foldable arm includes a fixed arm 211, a folding arm 212, and a rotary servo motor 210. The fixed arm 211 is fixed to the periphery of the waterproof central body 100, and the rotary servo motor 210 is fixedly mounted on the fixed arm 211, with its output shaft connected to the folding arm 212. Driven by the rotary servo motor 210, the folding arm 212 rotates relative to the fixed arm 211 between a retracted state and an extended state. In the retracted state, the folding arm 212 is against the side wall of the waterproof central body 100; in the extended state, the folding arm 212 extends horizontally outward.

[0099] It should be noted that the technical features in the above embodiments can be combined with each other to form more implementation methods without conflict. Those skilled in the art should understand that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A monitoring type lifesaving and rescue drone capable of water surface floating standby, characterized in that, include: The central body is waterproof, and a sealed cavity is formed inside it; Four foldable arms are symmetrically hinged in a cross shape around the waterproof central body. Each foldable arm is connected to the waterproof central body via a rotating servo motor and can switch between a folded state and an unfolded state. A brushless motor and a propeller are mounted at the end of each of the aforementioned foldable arms; Flexible solar photovoltaic panels cover the top of the waterproof central body and the upper surface of each of the foldable arms; An electromagnetic hook-type lifebuoy delivery mechanism is located at the bottom of the waterproof central body; The flight control computer is built into the cavity of the waterproof central fuselage; The hydrological monitoring sensor array is built into the cavity of the waterproof central body and is used to collect real-time hydrological data. A dual-light gimbal is installed at the front or bottom of the waterproof central body; The wireless communication module is built into the cavity of the waterproof central body.

2. The monitoring-type life-saving and rescue drone capable of floating and remaining in standby mode on the water surface according to claim 1, characterized in that, The foldable arm is attached to the side wall of the waterproof central fuselage in the retracted state, so that the overall length of the drone is less than the length in the unfolded state; the foldable arm is horizontally radial in the unfolded state.

3. A monitoring and rescue control method, applied to the UAV as described in claim 1 or 2, characterized in that, Includes the following steps: Water surface standby steps: The UAV floats on the water surface, and the flexible solar photovoltaic panel supplies power to the hydrological monitoring sensor group and the wireless communication module; the flight control computer controls the propeller to rotate, generating thrust to correct the UAV's water surface drift. Takeoff switching steps: When the wireless communication module receives an emergency call, the flight control computer controls the rotary servo to drive the foldable arm to switch from the retracted state to the unfolded state, starts the brushless motor and propeller, and controls the UAV to take off vertically. Steps for predicting the location of a person who has fallen into the water: Obtain real-time hydrological data collected by the hydrological monitoring sensor group and initial location information from the emergency rescue instructions; input the real-time hydrological data and the initial location information into a pre-trained LSTM neural network model to predict the drift trajectory and estimated location of the person who has fallen into the water. Search and rescue steps: Control the drone to fly to the estimated location, use the dual-light gimbal to search for and confirm the target of the person who fell into the water, and then control the electromagnetic hook-type lifebuoy deployment mechanism to release the lifebuoy; Safety confirmation and return procedure: Control the drone to land on the water surface and control the foldable arm to switch to the retracted state. Use the dual-light gimbal and the wireless communication module to conduct video and voice interaction with the person who fell into the water. After confirming safety, control the drone to return with the arm retracted.

4. The monitoring and rescue control method according to claim 3, characterized in that, The training process of the LSTM neural network model includes: Collect historical water rescue data, which includes historical hydrological data, historical initial location of falling into the water, historical drift trajectory, and historical final rescue location. The historical hydrological data and the historical initial location of falling into the water are used as input features, and the historical drift trajectory or the historical final rescue location is used as output label to train the LSTM neural network, thereby obtaining the pre-trained LSTM neural network model.

5. The monitoring and rescue control method according to claim 4, characterized in that, The LSTM neural network model includes a forget gate, an input gate, an output gate, and a cell state update mechanism. Its calculation process is expressed as follows: Forgotten Gate: ; Input Gate: ; Cell status update: ; Output gate: ; in, The input consists of the current time-series hydrological data and the initial location information of the water droplet. This is the predicted drift position information output at the current moment. Output for the forget gate. For input gate output, For output gate output, The current cell state, Candidate cell state, and These are the weight matrices and bias terms for each gate. It is the sigmoid activation function. The hyperbolic tangent activation function is used. This represents element-wise multiplication; Before inputting the feature factor data into the LSTM neural network model, each feature factor data is first standardized; after obtaining the prediction result, destandardization is performed.

6. The monitoring and rescue control method according to claim 3, characterized in that, In the water surface standby step, the specific method by which the flight control computer controls the propeller rotation to correct the UAV's water surface drift is as follows: Acquire the real-time location information and preset anchoring location information of the drone; Calculate the drift deviation between the real-time location information and the anchoring location information; Based on the drift deviation and the real-time hydrological data, the rotational speed and direction of each propeller are determined to generate a resultant force opposite to the drift direction, thereby keeping the UAV at the anchored position.

7. The monitoring and rescue control method according to claim 3, characterized in that, In the safety confirmation and return process, the video and voice interaction includes: establishing a real-time communication link with the remote rescue center through the wireless communication module, transmitting the on-site video collected by the dual-light PTZ and the voice information of the person in the water to the remote rescue center in real time, and receiving voice instructions from the remote rescue center to broadcast to the person in the water.

8. The monitoring-type life-saving and rescue drone capable of floating and remaining in standby mode on the water surface according to claim 1, characterized in that, The electromagnetic hook-type lifebuoy deployment mechanism includes an electromagnet, a hook, and a lifebuoy. The electromagnet is electrically connected to the flight control computer, and the flight control computer controls the on / off state of the electromagnet to control the attraction and release of the hook.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the monitoring and rescue control method according to any one of claims 3 to 7.

Images