A method and device for controlling unmanned aerial vehicle (UAV) navigation on water surface, unmanned aerial vehicle (UAV), and storage medium

Abstract

The present invention relates to the field of drone control technology, and in particular to a drone surface navigation control method, device, drone and storage medium. The drone surface navigation control method takes into account the drone's energy consumption and the reward mechanism for controlling the switching frequency, and processes the drone's current state quantity and the drone's surface control signal at the previous moment as inputs, thereby reducing the drone's energy consumption and avoiding frequent switching of the drone's power structure. The prediction result is obtained through the GRU network, and the surface control signal at the current moment is obtained to control the opening or closing of the power structure corresponding to the drone, thereby controlling the drone to reach the target point position and making the navigation path smooth, and realizing the control of surface navigation of the existing drone's power structure. At the same time, the drone's current state quantity is directly converted to the drone's surface control signal, realizing end-to-end drone control, which is more suitable for various drone surface navigation controls with water speed and heading control capabilities.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle control, and in particular to a method and device for controlling unmanned aerial vehicle surface navigation, an unmanned aerial vehicle and a storage medium. Background Art

[0002] With the rapid development of UAV technology, multi-purpose UAVs with cross-domain capabilities in water and air are increasingly used in various special environments. However, the increasing expansion and complexity of application scenarios have put forward higher requirements on the environmental adaptability of UAVs. There are essential differences in the dynamic characteristics of the water surface and the air. When the UAV wades, due to the large surface resistance, the propeller in the UAV power structure is difficult to rotate in the water as the throttle increases or decreases as in the air. After wading, the UAV cannot continuously adjust the speed as the throttle changes, but is equivalent to a switch-like state of only opening or closing. The power output corresponding to the switch-like state brings certain difficulties to the precise control of the UAV's navigation on the water surface. Under the current research background, it is a difficult problem in the use of UAVs to reuse the power structure of existing UAVs as simply as possible to realize the surface navigation control of UAVs. Summary of the invention

[0003] In order to solve the above technical problems, the present invention provides a method and device for controlling the surface navigation of an unmanned aerial vehicle, an unmanned aerial vehicle and a storage medium.

[0004] To achieve the above object, the present invention provides a method for controlling a UAV to navigate on the water, comprising:

[0005] Acquire the current state quantity of the drone, wherein the current state quantity of the drone includes the distance between the current position of the drone and the position of the target point, and the heading angle between the current position of the drone and the position of the target point;

[0006] Acquire a water surface control signal of the drone at the last moment, wherein the water surface control signal includes forward, close, right turn, and left turn, and each water surface control signal corresponds to a group of drone propeller control signals, wherein the drone includes a first power structure and a second power structure located on both sides, and the first power structure and the second power structure each include at least one propeller, and the propeller is at least partially immersed in water when the drone is sailing on the water surface, and the water surface control of the drone is achieved by controlling the opening or closing of the propellers of the first power structure and the second power structure;

[0007] The current state quantity of the UAV and the water surface control signal of the UAV at the previous moment are processed based on the reward mechanism to obtain and output a first water surface control signal sequence;

[0008] The GRU network obtains the first water surface control signal sequence and obtains a prediction result;

[0009] The UAV obtains the water surface control signal at the current moment corresponding to the prediction result, and controls the opening or closing of the corresponding propeller until the UAV reaches the target point position.

[0010] Optionally, the steps for obtaining the current state of the drone include:

[0011] Establish a water gliding coordinate system, the origin of which is located at the UAV fuselage, the x-axis of which is the projection of the UAV in the horizontal plane, the z-axis of which is vertically downward, and the y-axis of which is determined by the right-hand rule;

[0012] Based on the water gliding coordinate system, the coordinates of the current position of the UAV in the water gliding coordinate system are obtained as follows: , the coordinates of the target point position are ;

[0013] Based on the coordinates of the current position of the drone and the coordinates of the target point position, the distance between the current position of the drone and the target point position in the hydroplaning coordinate system is obtained as follows: ; The heading angle between the current position of the UAV and the position of the target point in the water gliding coordinate system of the UAV is: ;

[0014] The heading angle between the current position of the UAV and the target point position is obtained based on the inertial coordinate system and the water gliding coordinate system:

[0015] , where g is the water gliding coordinate system, N is the inertial coordinate system, q0, q1, q2, q3 are attitude quaternions, is the x-axis and y-axis projection corresponding to the current position of the drone in the inertial coordinate system, are the x-axis and y-axis projections corresponding to the current position of the UAV in the water gliding coordinate system, and t is the current time.

[0016] Optionally, the propeller control signal includes: the first power structure is turned on, the first power structure is turned off, the second power structure is turned on, and the second power structure is turned off;

[0017] When the water surface control signal is stop, the first power structure is shut down and the second power structure is shut down;

[0018] When the water surface control signal is to turn left, the first power structure is closed and the second power structure is opened;

[0019] When the water surface control signal is forward, the first power structure is turned on and the second power structure is turned on;

[0020] When the water surface control signal is to turn right, the first power structure is turned on and the second power structure is turned off.

[0021] Optionally, the step of processing the current state quantity of the drone and the water surface control signal of the drone at the previous moment based on the reward mechanism to obtain and output a first water surface control signal sequence includes:

[0022] Obtain the current state of the drone;

[0023] Processing the current state of the drone through the first fully connected layer;

[0024] Obtaining the surface control signal of the UAV at the last moment;

[0025] Processing the surface control signal of the UAV at the last moment through the second fully connected layer;

[0026] The current state quantity of the UAV processed by the first fully connected layer and the water surface control signal of the UAV at the previous moment processed by the second fully connected layer are processed based on the reward mechanism to obtain and output the first water surface control signal sequence.

[0027] Optionally, the reward mechanism is:

[0028] ,

[0029] Among them, R is the reward value of the reward mechanism, To terminate the reward, is the distance between the current position of the drone and the position of the target point, Penalty for surface control signal, To control the switching penalty, is the distance adjustment parameter, is a positive number. Optionally, the termination reward for:

[0030] ,in, is the distance between the current position of the drone and the position of the target point, is the distance adjustment parameter, A is a constant and not equal to 0;

[0031] The drone surface control signal penalty for:

[0032] ,in, , , is a constant, ;

[0033] The control switching penalty The Markov table is used to quantitatively define the switching penalty of the UAV's surface control signal.

[0034] Optionally, the UAV is a binary vector powered UAV.

[0035] The present invention also provides a UAV surface navigation control device, comprising:

[0036] A first control unit is used to obtain a current state quantity of the drone, wherein the current state quantity of the drone includes a distance between a current position of the drone and a target point position, and a heading angle between the current position of the drone and the target point position;

[0037] A second control unit is used to obtain a water surface control signal of the drone at the last moment, wherein the water surface control signal includes forward, close, right turn, and left turn, and each water surface control signal corresponds to a group of drone propeller control signals, wherein the drone includes a first power structure and a second power structure located on both sides, and the first power structure and the second power structure each include at least one propeller and the propeller is at least partially immersed in water when the drone is sailing on the water surface, and the water surface control of the drone is achieved by controlling the opening or closing of the propellers of the first power structure and the second power structure;

[0038] A third control unit is used to process the current state quantity of the drone and the water surface control signal of the drone at the previous moment based on the reward mechanism, and obtain and output a first water surface control signal sequence;

[0039] A fourth control unit, used for the GRU network to obtain the first water surface control signal sequence and obtain a prediction result;

[0040] The fifth control unit is used for the UAV to obtain the water surface control signal corresponding to the prediction result at the current moment, and control the opening or closing of the corresponding propeller until the UAV reaches the target point position.

[0041] The present invention also provides a UAV, comprising: a memory, a processor, and a UAV surface navigation control program stored in the memory and executable on the processor, wherein the UAV surface navigation control is configured to implement the steps of the above-mentioned UAV surface navigation control method.

[0042] The present invention also provides a storage medium, comprising: a UAV surface navigation control program is stored on the storage medium, and a UAV surface navigation control method is implemented when the UAV surface navigation control program is executed by a processor.

[0043] In summary, the advantages and beneficial effects of the present invention are:

[0044] The present invention provides a method, device, drone and storage medium for controlling the navigation of a drone on the water surface. The method comprises obtaining the current state quantity of the drone, the current state quantity of the drone includes the distance between the current position of the drone and the position of a target point, and the heading angle between the current position of the drone and the position of the target point; obtaining the surface control signal of the drone at the last moment, the surface control signal includes forward, close, right turn, and left turn, and each surface control signal corresponds to a group of drone propeller control signals; processing the current state quantity of the drone and the surface control signal of the drone at the last moment based on a reward mechanism, obtaining and outputting a first surface control signal sequence; a GRU network obtains the first surface control signal sequence and obtains a prediction result; the drone obtains the surface control signal at the current moment corresponding to the prediction result, and controls the opening or closing of the corresponding propeller until the drone reaches the target point position.

[0045] The reward mechanism takes into account the distance between the current position of the drone and the target point, the energy consumption of the drone, and the control switching frequency. The current state of the drone and the surface control signal of the drone at the last moment are processed as inputs to reduce the energy consumption of the drone and avoid frequent switching of the power structure of the drone. The prediction result is further obtained through the GRU network to obtain the surface control signal at the current moment, and the corresponding power structure is controlled to be turned on or off, thereby controlling the drone to reach the target point. The energy consumption and frequent switching of the control of the unmanned surface navigation are controlled to make the navigation path of the drone smooth, and realize the control of the drone surface navigation of the existing drone power structure. At the same time, the drone current state is directly converted to the drone surface control signal, which realizes end-to-end drone control, simplifies the control of the drone, and is more suitable for various drone surface navigation control methods with water speed and heading control capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A flow chart of a method for controlling a drone's surface navigation provided by an embodiment of the present invention;

[0047] Figure 2 A schematic diagram of a UAV according to a method for controlling UAV surface navigation provided by an embodiment of the present invention;

[0048] Figure 3 A schematic diagram of a surface control signal switching penalty multiplier coefficient of a UAV surface navigation control method provided by an embodiment of the present invention;

[0049] Figure 4A schematic diagram of a GRU network flow of a method for controlling the surface navigation of an unmanned aerial vehicle provided by an embodiment of the present invention;

[0050] Figure 5 A schematic diagram of water navigation simulation of a method for controlling surface navigation of an unmanned aerial vehicle provided in an embodiment of the present invention.

[0051] 1. UAV, 21. first propeller, 22. second propeller, 3. GNSS device. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in combination with the specific implementation methods of the present invention and the corresponding drawings. Obviously, the described implementation methods are only part of the implementation methods of the present invention, not all implementation methods. Based on the implementation methods of the present invention, all other implementation methods obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0053] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0054] The present invention provides a method for controlling the navigation of an unmanned aerial vehicle on the water surface. Figure 1 As shown, including:

[0055] Step S10, obtaining the current state quantity of the drone, wherein the current state quantity of the drone includes the distance between the current position of the drone and the position of the target point, and the heading angle between the current position of the drone and the position of the target point;

[0056] Step S20, obtaining a water surface control signal of the drone at the last moment, wherein the water surface control signal includes forward, close, right turn, and left turn, and each water surface control signal corresponds to a group of drone propeller control signals, wherein the drone includes a first power structure and a second power structure located on both sides, and the first power structure and the second power structure each include at least one propeller and the propeller is at least partially immersed in water when the drone is sailing on the water surface, and the water surface control of the drone is achieved by controlling the opening or closing of the propellers of the first power structure and the second power structure;

[0057] Step S30, processing the current state quantity of the drone and the water surface control signal of the drone at the previous moment based on the reward mechanism, and obtaining and outputting a first water surface control signal sequence;

[0058] Step S40, the GRU network acquires the first water surface control signal sequence and obtains a prediction result;

[0059] Step S50, the UAV obtains the water surface control signal corresponding to the prediction result at the current moment, and controls the opening or closing of the corresponding power structure until the UAV reaches the target point position.

[0060] Specifically, step S10 is executed to obtain the current state quantity of the drone, where the current state quantity of the drone includes the distance between the current position of the drone and the position of the target point, and the heading angle between the current position of the drone and the position of the target point.

[0061] In an embodiment of the present invention, the step of obtaining the current state of the drone includes:

[0062] Establish a water gliding coordinate system, the origin of which coincides with the center of gravity of the UAV, the x-axis of which is the projection of the UAV in the horizontal plane, the z-axis of which is vertically downward, and the y-axis of which is determined by the right-hand rule;

[0063] Based on the water gliding coordinate system, the coordinates of the current position of the UAV in the water gliding coordinate system are obtained as follows: , the coordinates of the target point position are ;

[0064] Based on the coordinates of the current position of the drone and the coordinates of the target point position, the distance between the current position of the drone and the target point position in the hydroplaning coordinate system is obtained as follows: ; The heading angle between the current position of the UAV and the position of the target point in the water gliding coordinate system of the UAV is: ;

[0065] Acquire a heading angle between the current position of the UAV and the position of the target point, and a distance between the current position of the UAV and the position of the target point based on an inertial coordinate system and the hydroplaning coordinate system;

[0066] Specifically, the heading angle between the current position of the drone and the target point position is:

[0067] , where g is the water gliding coordinate system, N is the inertial coordinate system, q0, q1, q2, q3 are attitude quaternions, is the x-axis and y-axis projection corresponding to the current position of the drone in the inertial coordinate system, is the x-axis and y-axis projection corresponding to the current position of the drone in the inertial coordinate system, and t is the current time.

[0068] In an embodiment of the present invention, the distance between the current position of the drone and the target point position is obtained by a positioning system carried by the drone, for example: Figure 2 As shown, the GNSS device 3.

[0069] In the embodiment of the present invention, the attitude quaternions q0, q1, q2, q3 are used as parameters to describe the rotation relationship between the gliding coordinate system and the inertial coordinate system, so as to facilitate the UAV to execute the surface control signal.

[0070] A water gliding coordinate system is established with the center of mass of the drone as the origin. Based on the water gliding coordinate system, the distance between the current position of the drone and the target point and the heading angle between the current position of the drone and the target point are determined from the perspective of the drone, so as to facilitate the navigation control of the drone.

[0071] Execute step S20 to obtain the surface control signal of the UAV at the last moment, wherein the surface control signal includes forward, closed, right turn, and left turn, and each surface control signal corresponds to a group of UAV propeller control signals, wherein the UAV includes a first power structure and a second power structure located on both sides, and the first power structure and the second power structure each include at least one propeller and the propeller is at least partially immersed in water when the UAV is sailing on the water surface, and the surface control of the UAV is achieved by controlling the opening or closing of the propellers of the first power structure and the second power structure.

[0072] In an embodiment of the present invention, the first power structure and the second power structure each include two or more propellers.

[0073] In an embodiment of the present invention, the first power structure and the second power structure of the UAV can provide power both in the air and in wading water. When the UAV wades water, due to the large water surface resistance, the propellers in the first power structure and the second power structure cannot continuously adjust the speed as the throttle increases or decreases as in the air. After wading water, the UAV cannot continuously adjust the speed as the throttle changes, but is in a state of switching power output. Under the current research conditions, the power structure of the existing UAV is adopted, and the UAV control signal is directly output based on the current position of the UAV to achieve end-to-end human-machine surface navigation control.

[0074] In an embodiment of the present invention, the first power structure and the second power structure are located on the left and right sides of the tail of the UAV, wherein the first power structure is located on the left side of the UAV and the second power structure is located on the right side of the UAV.

[0075] In an embodiment of the present invention, the propeller control signal includes: first power structure on, first power structure off, second power structure on, second power structure off;

[0076] When the water surface control signal is stop, the first power structure is shut down and the second power structure is shut down;

[0077] When the water surface control signal is to turn left, the first power structure is closed and the second power structure is opened;

[0078] When the water surface control signal is forward, the first power structure is turned on and the second power structure is turned on;

[0079] When the water surface control signal is to turn right, the first power structure is turned on and the second power structure is turned off.

[0080] In an embodiment of the present invention, the UAV controls the opening or closing of the corresponding propeller based on the surface control signal of the UAV, so as to achieve the purpose of controlling the switching of the motion state of the UAV (i.e., the UAV action). Correspondingly, the motion state of the UAV is forward, closed, turn right, and turn left.

[0081] Based on the power provided by the unmanned first power structure and the second power structure on the water surface, the surface navigation of the unmanned aerial vehicle is controlled.

[0082] In the embodiment of the present invention, Figure 2 As shown, the UAV is a binary vector-powered UAV 1, the first power structure includes a first propeller 21, the second power structure includes a second propeller 22, the first power structure and the second power structure are located on both sides of the tail of the binary vector-powered UAV 1, wherein the first power structure is located on the left side of the UAV, and the second power structure is located on the right side of the UAV.

[0083] Execute step S30 to process the current state quantity of the drone and the water surface control signal of the drone at the previous moment based on the reward mechanism to obtain and output a first water surface control signal sequence.

[0084] In the embodiment of the present invention, Figure 4 As shown, the DQN controller is used to convert the current state of the drone into , the surface control signal of the drone at the last moment Processing is performed to obtain and output a first sequence.

[0085] In an embodiment of the present invention, the steps of using a DQN controller to process based on a reward mechanism to obtain and output a first water surface control signal sequence include:

[0086] Obtain the current state of the drone;

[0087] Processing the current state of the drone through the first fully connected layer;

[0088] Obtaining the surface control signal of the UAV at the last moment;

[0089] Processing the surface control signal of the UAV at the last moment through the second fully connected layer;

[0090] The current state quantity of the UAV processed by the first fully connected layer and the water surface control signal of the UAV at the previous moment processed by the second fully connected layer are processed (added) based on the reward mechanism to obtain and output the first water surface control signal sequence.

[0091] In this embodiment of the present invention, the current state of the drone is , the current state of the UAV is the current state of the UAV obtained based on the inertial coordinate system and the water gliding coordinate system, where k is the current moment, is the distance between the current position of the drone and the target point at the current moment, is the heading angle between the current position of the drone and the position of the target point at the current moment, It is the current state of the drone at the current moment.

[0092] In the embodiment of the present invention, the time interval between the current moment and the previous moment is controlled by the frequency of the DQN controller. Specifically, the frequency of the DQN controller is 10 Hz, so that the time interval is 0.1 s.

[0093] In other embodiments, the frequency range of the DQN controller is 10HZ~20HZ, which can effectively control the UAV's water surface gliding control.

[0094] In an embodiment of the present invention, the first fully connected layer and the second fully connected layer are both a neural network layer that performs multiplication and offset calculations. The current state quantity of the drone and the surface control signal of the drone at the previous moment can be processed by the first fully connected layer and the second fully connected layer before subsequent calculations can be performed.

[0095] In the embodiment of the present invention, the reward mechanism is:

[0096] ,

[0097] Among them, R is the reward value of the reward mechanism, To terminate the reward, is the distance between the current position of the drone and the position of the target point, Penalty for drone surface control signal, To control the switching penalty, is the distance adjustment parameter, Is a positive number.

[0098] In this embodiment of the present invention, the termination reward for:

[0099] ,

[0100] in, is the distance between the current position of the drone and the position of the target point, is the distance adjustment parameter, is a positive number, A is a constant and is not equal to 0, specifically A is 100, that is:

[0101] .

[0102] In this embodiment of the present invention, the distance between the current position of the drone and the target point position is greater than the distance adjustment parameter When the target point is within a certain range, the target point is judged to be far away, and the termination reward is set. is 0, when the distance between the current position of the drone and the target point is less than the distance adjustment parameter When the target point is within a certain range, the target point is judged to be close, and the termination reward is set at this time. is 100, driving the UAV to reach the target point as quickly as possible.

[0103] In other embodiments, when the distance between the current position of the drone and the target point position is greater than or less than a certain range of the distance adjustment parameter, the termination reward is set according to the specific situation, wherein the termination reward when the distance between the current position of the drone and the target point position is greater than the distance adjustment parameter is less than the termination reward when the distance between the current position of the drone and the target point position is less than the distance adjustment parameter.

[0104] In this embodiment of the present invention, the UAV surface control signal penalty for:

[0105] ,in, , , is a constant, ;

[0106] Specific, for example ,

[0107] Among them, forward, close, turn right, and turn left are the surface control signals of the UAV.

[0108] In an embodiment of the present invention, the corresponding propeller is turned on or off based on the surface control signal of the UAV, thereby generating corresponding energy consumption. For example, when moving forward, the first power structure is turned on and the second power structure is turned on; when turning right, the first power structure is turned off and the second power structure is turned on. When moving forward, both the first power structure and the second power structure need to consume energy, while when turning right, only the second power structure needs to be turned on. Therefore, the energy consumed when moving forward is greater than the energy consumption when turning right. In order to reduce energy consumption, more energy consumption is set and more penalties are given. Specifically, when moving forward, both power structures are turned on and both need to consume energy. At this time, the energy consumed is the largest, and the penalty given is 0.15; when turning left and right, only one power structure is turned on to consume energy, and the other power structure is not turned on and does not consume energy. The penalty given is 0.1. Both the first power structure and the second power structure are turned off and do not consume energy. The penalty given is 0, thereby achieving the purpose of reducing energy consumption.

[0109] In the embodiment of the present invention, the control switching penalty The Markov table is used to quantitatively define the switching penalty of the UAV's surface control signal.

[0110] In order to avoid frequent switching of the surface control signal of the UAV, which causes the first power structure and the second power structure to be frequently switched on and off, causing the UAV to frequently switch left, right, forward, and shut down, for example: the UAV keeps turning left or right, causing the UAV to be unable to move forward stably and consume more energy. Therefore, the penalty obtained by setting the surface control signal of the UAV to control the opening of the first power structure and the second power structure is greater than the penalty obtained by closing the first power structure and the second power structure, prompting the controller to control in a smooth manner.

[0111] In the embodiment of the present invention, since the surface control signal of the UAV is discrete, the action of the UAV is also discrete. Therefore, the present invention uses a Markov table to quantify the switching penalty of the surface control signal of the UAV, such as Figure 3 As shown, specifically, when the surface control signal of the UAV switches from stop to right turn, the penalty multiplier is 2; when the surface control signal of the UAV switches from right turn to stop, the penalty multiplier is 1, where, Figure 3 The numbers in the box represent the penalty multiplier.

[0112] In order to control the movement of the UAV, it is necessary to control the opening or closing of the first power structure and the second power structure of the UAV to achieve a suitable state of average power (the average value of the output power per unit time) to achieve the purpose of controlling the direction of the UAV. However, the rotation of the propeller requires some additional energy, which is not conducive to energy-saving control. In addition, to achieve the average power, the power structure can be frequently switched on and off, or the power structure can be slowly switched on and off, or the average power can be achieved by only opening it once. Since energy is wasted in the process of frequently switching the power structure on and off, and the faster the switching speed, the more energy is consumed, and even damage is caused. Therefore, by setting different penalties, the control purpose can be achieved at one time instead of frequently switching the power structure on and off.

[0113] Execute step S40, such as Figure 4 As shown, the GRU network obtains the first water surface control signal sequence and obtains a prediction result.

[0114] In the embodiment of the present invention, the GRU network obtains the first sequence through the GRU layer, the third fully connected layer, and the softmax (normalized exponential function) to output the prediction result.

[0115] The present invention adopts a DQN controller, takes the continuous current state quantity of the UAV as input, and sets a reward mechanism to better solve the problem of frequent opening and closing of the first power structure and the second power structure caused by the frequent switching of the UAV surface control signal. At the same time, a GRU network is introduced to realize the memory of the past motion state of the UAV, and the surface control signal of the UAV at the current moment is jointly regulated based on the past motion state of the UAV and the current state quantity of the UAV, thereby avoiding the control of the UAV by only a single variable. For example, if only judgment is made based on the current state quantity of the UAV, control overshoot will occur. By introducing a GRU network, the current state quantity of the UAV over a period of time in the past and the surface control signal of the UAV at the previous moment are memorized, thereby solving the overshoot problem caused by a single control variable, and outputting the discrete surface control signal of the UAV at the current moment, thereby realizing that the UAV finally outputs the surface control signal of the UAV at the current moment from the current state quantity of the UAV, thereby realizing end-to-end control.

[0116] Step S50 is executed, the UAV obtains the water surface control signal at the current moment corresponding to the prediction result , control the opening or closing of the corresponding power structure until the UAV reaches the target point.

[0117] In an embodiment of the present invention, based on the water surface control signal at the current moment , obtain a corresponding set of drone propeller control signals, control the corresponding power structure to open or close, and enable the drone to execute the surface control signal at the current moment, so that the drone can reach the target point.

[0118] In the embodiment of the present invention, Figure 5 As shown, in the simulation environment, with the horizontal plane as the reference plane, the UAV's surface navigation control is verified in the two-dimensional vertical-horizontal direction to obtain the position of the UAV reaching the target point and the steering ability. The dotted line in the simulation result is the line connecting the UAV and the target position, and the solid line is the trajectory of the UAV's historical navigation. As can be seen from the figure, the UAV has the ability to track complex tracks on the water.

[0119] The present invention also provides a UAV surface navigation control device, comprising:

[0120] A first control unit is used to obtain a current state quantity of the drone, wherein the current state quantity of the drone includes a distance between a current position of the drone and a target point position, and a heading angle between the current position of the drone and the target point position;

[0121] A second control unit is used to obtain a water surface control signal of the drone at the last moment, wherein the water surface control signal includes forward, close, right turn, and left turn, and each water surface control signal corresponds to a group of drone propeller control signals, wherein the drone includes a first power structure and a second power structure located on both sides, and the first power structure and the second power structure each include at least one propeller and the propeller is at least partially immersed in water when the drone is sailing on the water surface, and the water surface control of the drone is achieved by controlling the opening or closing of the propellers of the first power structure and the second power structure;

[0122] A third control unit processes the current state quantity of the drone and the water surface control signal of the drone at the previous moment based on the reward mechanism, and obtains and outputs a first water surface control signal sequence;

[0123] A fourth control unit, used for the GRU network to obtain the first water surface control signal sequence and obtain a prediction result;

[0124] The fifth control unit is used for the UAV to obtain the water surface control signal corresponding to the prediction result at the current moment, and control the opening or closing of the corresponding propeller until the UAV reaches the target point position.

[0125] The present invention also provides a UAV, comprising: a memory, a processor, and a UAV surface navigation control program stored in the memory and executable on the processor, wherein the UAV surface navigation control is configured to implement the steps of the above-mentioned UAV surface navigation control method.

[0126] The present invention also provides a storage medium, comprising: a UAV surface navigation control program is stored on the storage medium, and the UAV surface navigation control program implements the above-mentioned UAV surface navigation control method when executed by a processor.

[0127] Finally, it is to be noted that any modification or equivalent replacement of part or all of the technical features based on the device structure of the present invention and the technical solution of the described embodiment, which does not deviate from the essence of the corresponding technical solution of the present invention, belongs to the patent scope of the device structure of the present invention and the described implementation scheme.

Claims

1. A method for controlling the navigation of an unmanned aerial vehicle on the water surface, characterized in that: include: Acquire the current state quantity of the drone, wherein the current state quantity of the drone includes the distance between the current position of the drone and the position of the target point, and the heading angle between the current position of the drone and the position of the target point; Acquire a water surface control signal of the drone at the last moment, wherein the water surface control signal includes forward, close, right turn, and left turn, and each water surface control signal corresponds to a group of drone propeller control signals, wherein the drone includes a first power structure and a second power structure located on both sides, and the first power structure and the second power structure each include at least one propeller, and the propeller is at least partially immersed in water when the drone is sailing on the water surface, and the water surface control of the drone is achieved by controlling the opening or closing of the propellers of the first power structure and the second power structure; The current state quantity of the drone and the surface control signal of the drone at the last moment are processed based on the reward mechanism to obtain and output a first surface control signal sequence. The step of processing the current state quantity of the drone and the surface control signal of the drone at the last moment based on the reward mechanism to obtain and output the first surface control signal sequence includes: obtaining the current state quantity of the drone; processing the current state quantity of the drone through the first fully connected layer; obtaining the surface control signal of the drone at the last moment; processing the surface control signal of the drone at the last moment through the second fully connected layer; processing the current state quantity of the drone processed by the first fully connected layer and the surface control signal of the drone at the last moment processed by the second fully connected layer based on the reward mechanism to obtain and output the first surface control signal sequence; wherein the reward mechanism is: , where R is the reward value of the reward mechanism, To terminate the reward, is the distance between the current position of the drone and the position of the target point, Penalty for surface control signal, To control the switching penalty, is the distance adjustment parameter, is a positive number; the termination reward for: ,in, is the distance between the current position of the drone and the position of the target point, is the distance adjustment parameter, A is a constant and is not equal to 0; the UAV surface control signal penalty for: ,in, , , is a constant, ; The control switching penalty The Markov table is used to quantitatively define the switching penalty of the UAV's surface control signal; The GRU network obtains the first water surface control signal sequence and obtains a prediction result; The UAV obtains the water surface control signal at the current moment corresponding to the prediction result, and controls the opening or closing of the corresponding propeller until the UAV reaches the target point position.

2. A method for controlling the navigation of an unmanned aerial vehicle on the water surface as claimed in claim 1, characterized in that: The steps to obtain the current state of the drone include: Establish a water gliding coordinate system, the origin of which is located at the UAV fuselage, the x-axis of which is the projection of the line from the center of mass of the UAV to the nose of the UAV in the horizontal plane, the z-axis of which is vertically downward, and the y-axis of which is determined by the right-hand rule; Based on the water gliding coordinate system, the coordinates of the current position of the UAV in the water gliding coordinate system are obtained as follows: , the coordinates of the target point position are ; Based on the coordinates of the current position of the drone and the coordinates of the target point position, the distance between the current position of the drone and the target point position in the hydroplaning coordinate system is obtained as follows: ; The heading angle between the current position of the UAV and the position of the target point in the water gliding coordinate system of the UAV is: ; The heading angle between the current position of the UAV and the target point position is obtained based on the inertial coordinate system and the water gliding coordinate system: , where g is the water gliding coordinate system, N is the inertial coordinate system, q0, q1, q2, q3 are attitude quaternions, , is the x-axis and y-axis projection corresponding to the current position of the drone in the inertial coordinate system, , are the x-axis and y-axis projections corresponding to the current position of the UAV in the water gliding coordinate system, and t is the current time.

3. The method for controlling the surface navigation of an unmanned aerial vehicle according to claim 1, characterized in that: The propeller control signal includes: first power structure on, first power structure off, second power structure on, second power structure off; When the water surface control signal is stop, the first power structure is shut down and the second power structure is shut down; When the water surface control signal is to turn left, the first power structure is closed and the second power structure is opened; When the water surface control signal is forward, the first power structure is turned on and the second power structure is turned on; When the water surface control signal is to turn right, the first power structure is turned on and the second power structure is turned off.

4. The method for controlling the surface navigation of an unmanned aerial vehicle according to claim 1, characterized in that: The UAV is a binary vector powered UAV.

5. A UAV surface navigation control device, characterized in that: include: A first control unit is used to obtain a current state quantity of the drone, wherein the current state quantity of the drone includes a distance between a current position of the drone and a target point position, and a heading angle between the current position of the drone and the target point position; A second control unit is used to obtain a water surface control signal of the drone at the last moment, wherein the water surface control signal includes forward, close, right turn, and left turn, and each water surface control signal corresponds to a group of drone propeller control signals, wherein the drone includes a first power structure and a second power structure located on both sides, and the first power structure and the second power structure each include at least one propeller and the propeller is at least partially immersed in water when the drone is sailing on the water surface, and the water surface control of the drone is achieved by controlling the opening or closing of the propellers of the first power structure and the second power structure; The third control unit is used to process the current state quantity of the drone and the surface control signal of the drone at the last moment based on the reward mechanism to obtain and output a first surface control signal sequence. The step of processing the current state quantity of the drone and the surface control signal of the drone at the last moment based on the reward mechanism to obtain and output the first surface control signal sequence includes: obtaining the current state quantity of the drone; processing the current state quantity of the drone through the first fully connected layer; obtaining the surface control signal of the drone at the last moment; processing the surface control signal of the drone at the last moment through the second fully connected layer; processing the current state quantity of the drone processed by the first fully connected layer and the surface control signal of the drone at the last moment processed by the second fully connected layer based on the reward mechanism to obtain and output the first surface control signal sequence; wherein the reward mechanism is: , where R is the reward value of the reward mechanism, To terminate the reward, is the distance between the current position of the drone and the position of the target point, Penalty for surface control signal, To control the switching penalty, is the distance adjustment parameter, is a positive number; the termination reward for: ,in, is the distance between the current position of the drone and the position of the target point, is the distance adjustment parameter, A is a constant and is not equal to 0; the UAV surface control signal penalty for: ,in, , , is a constant, ; The control switching penalty The Markov table is used to quantitatively define the switching penalty of the UAV's surface control signal; A fourth control unit, used for the GRU network to obtain the first water surface control signal sequence and obtain a prediction result; The fifth control unit is used for the UAV to obtain the water surface control signal corresponding to the prediction result at the current moment, and control the opening or closing of the corresponding propeller until the UAV reaches the target point position.

6. A drone, characterized in that: include: A memory, a processor, and a UAV surface navigation control program stored in the memory and executable on the processor, wherein the UAV surface navigation control is configured to implement the steps of a UAV surface navigation control method as described in any one of claims 1 to 4.

7. A storage medium, characterized in that: include: The storage medium stores a UAV surface navigation control program, and when the UAV surface navigation control program is executed by the processor, a UAV surface navigation control method as described in any one of claims 1 to 4 is implemented.

Images