A cross-medium control system and method for water-air multi-rotor unmanned aerial vehicle

Abstract

The present invention discloses a cross-media control system and method for a water-air multi-rotor unmanned aerial vehicle, the system comprising a main controller, a power subsystem, a monitoring subsystem, a sensor subsystem, and a communication subsystem arranged on the water-air multi-rotor unmanned aerial vehicle, wherein the main controller is used to use a humanoid intelligent controller to perform multi-feature fusion pattern recognition on the water-air multi-rotor unmanned aerial vehicle according to information obtained by the monitoring subsystem and the sensor subsystem, determine the optimal heading and expected attitude to be taken by the water-air multi-rotor unmanned aerial vehicle in the next step, and the optimal control strategy to be adopted for adjusting the attitude, and control is performed by improving the motion control level. The present invention adopts multi-modal control, adopts adaptive control strategies for different feature states, and has strong adaptability to complex environments.

Description

Technical Field

[0001] The present invention relates to the technical field of water-air amphibious unmanned aerial vehicles, and in particular to a water-air multi-rotor unmanned aerial vehicle cross-medium control system and method that combines humanoid intelligent control with self-disturbance rejection control. Background Art

[0002] Since the 21st century, the field of unmanned aerial vehicles has developed rapidly, and various new types of unmanned aerial vehicles have been born one after another. Among them, multi-rotor aircraft are widely used by people because of their advantages such as flexibility and stability. In the military, they are often used for reconnaissance and patrolling; in the civilian field, they can be used for daily entertainment photography, plant protection, water environment monitoring, and search and rescue missions in disasters. At present, traditional unmanned aerial vehicles can generally only work in the air and cannot meet the needs of continuous reconnaissance operations in water, land, and air scenes, such as continuous reconnaissance operations in water, land, and air in oceans, rivers, lakes, and dams. Amphibious multi-rotor drones are a new type of robot that can operate in water and air environments. They have strong environmental adaptability and effectively solve the problem that traditional drones cannot meet the needs of continuous reconnaissance operations in water, land, and air scenes. They have now become a hot topic of research in various countries. Amphibious multi-rotor drones need to overcome the problem of water-air medium cross-domain when operating. When crossing the medium, there are many complex factors such as environmental mutations (the physical properties of water and air are very different) and interference from the external environment (such as water flow, wind and waves, temperature and pressure, etc.). Traditional cross-medium control methods have low robustness and are difficult to maintain system stability in complex environments, resulting in cross-medium failure. Currently, there is no good control method or system that can enable amphibious multi-rotor drones to achieve stable cross-medium in complex and changing environments. Summary of the invention

[0003] In view of the above-mentioned deficiencies in the prior art, the present invention provides a cross-media control system and method for a water-to-air multi-rotor UAV that combines human-like intelligent control with self-disturbance rejection control.

[0004] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0005] In a first aspect, the present invention proposes a cross-media control system for a water-air multi-rotor UAV, including a main controller, a power subsystem, a monitoring subsystem, a sensor subsystem, and a communication subsystem arranged on the water-air multi-rotor UAV;

[0006] The power subsystem is used to provide a driving signal to the water-air multi-rotor UAV and to supply power to the main controller;

[0007] The monitoring subsystem is used to monitor the internal and external environmental information of the water-air multi-rotor UAV in real time;

[0008] The sensor subsystem is used to obtain the body information of the water-air multi-rotor UAV and the navigation information in the underwater environment;

[0009] The communication subsystem is used to receive external command information and transmit battery power, humidity inside the control box and temperature of the main controller, and the body posture and position information.

[0010] The main controller is used to use a humanoid intelligent controller to perform multi-feature fusion pattern recognition on the water-air multi-rotor UAV according to the information obtained by the monitoring subsystem and the sensor subsystem, determine the optimal heading and expected attitude to be taken by the water-air multi-rotor UAV in the next step and the optimal control strategy to be adopted for adjusting the attitude, and control it by improving the motion control level.

[0011] Further, the humanoid intelligent controller includes a task adaptation level, a parameter correction level and an improved motion control level;

[0012] The task adaptation level is used to use an expert system to perform a fusion analysis on the information obtained by the monitoring subsystem and the sensor subsystem to determine the optimal speed, posture or other operations that the water-air multi-rotor UAV needs to achieve in the face of the current situation;

[0013] The parameter correction stage is used to correct the threshold value in the feature model and adjust the parameters of the motion control stage under different feature states;

[0014] The improved motion control stage is used to adopt an anti-disturbance controller, estimate the real-time attitude and total system disturbance of the water-air multi-rotor UAV through an extended state observer, and compensate the control rate according to the stable system input signal, the estimated real-time attitude and the total disturbance of the system through nonlinear error feedback, and generate the final control quantity of the water-air multi-rotor UAV system. Finally, the system input signal is tracked and differentiated by a tracking differentiator to obtain a stable system input signal.

[0015] Further, the expert system includes a knowledge acquisition structure, an expert knowledge base, a comprehensive database and an inference engine;

[0016] The knowledge acquisition structure is used to acquire expert knowledge and store it in the expert knowledge base according to the set rules;

[0017] The expert knowledge base is used to represent the stored expert knowledge in a set form and distinguish whether the expert knowledge is quantifiable;

[0018] The comprehensive database is used to store the real-time monitoring data of the monitoring subsystem and the sensor subsystem on the water-air multi-rotor UAV;

[0019] The inference engine is used to match and solve the initial facts provided by the comprehensive database with the knowledge in the expert knowledge base, and output the countermeasure with the highest matching degree to the main controller.

[0020] Furthermore, when controlling the water-to-air multi-rotor drone to enter water,

[0021] The humanoid intelligent controller uses multi-feature fusion pattern recognition based on the body information and environmental information obtained from the monitoring subsystem and the sensor subsystem to make an intelligent judgment on whether to enter the water in the current environment, and issues two preliminary instructions to the main controller: to enter the water or not;

[0022] When the command is to enter the water, the humanoid intelligent controller is used to determine the UAV's water entry direction, water entry posture and water entry speed according to environmental information, and to control the UAV's posture, heading and speed; when the UAV is detected to be in contact with the water surface, the humanoid intelligent controller is used to adjust the UAV's posture to be coplanar with the horizontal plane and make the UAV sink into the water by its own gravity;

[0023] When the instruction is not to enter the water, the humanoid intelligent controller is used to judge the environmental information; if it is a hardware failure, an alarm is sent through the communication subsystem and the drone returns home autonomously; if it is due to environmental factors, the communication subsystem transmits information that the current water environment is poor and autonomously searches for surrounding waters suitable for entering the water. If a better environment suitable for entering the water is found, an entry instruction is issued, and the main controller controls the drone to execute the above entry instruction; otherwise, it is judged that the current environment is bad and it is impossible to enter the water, and the drone returns home autonomously.

[0024] Furthermore, when the water-to-air multi-rotor drone is controlled to exit the water,

[0025] The main controller receives the water exit command, controls the drone to drive to the horizontal plane, and uses the humanoid intelligent controller to adjust the drone's attitude so that the fixed rotor surface is the top surface and coplanar with the horizontal plane;

[0026] The humanoid intelligent controller is used to perform multi-feature fusion pattern recognition based on the body information and environmental information obtained from the monitoring subsystem and the sensor subsystem, to make an intelligent judgment on whether water is being discharged in the current environment, and to issue two preliminary instructions to the main controller: to discharge water or not to discharge water;

[0027] When the instruction is to exit the water, the humanoid intelligent controller is used to determine the exit direction, exit posture and exit speed of the UAV according to the environmental information, and to control the posture, heading and speed of the UAV to make the UAV leave the water;

[0028] When the instruction is not to exit the water temporarily, the humanoid intelligent controller sends an instruction to search for the water exit environment to the main controller; when the humanoid intelligent controller senses that the current environment meets the water exit requirements, it sends a water exit instruction to the main controller to complete the UAV's exit from the water; if the current environment does not meet the water exit requirements, an alarm is sent through the communication subsystem and its own posture is maintained.

[0029] In a second aspect, the present invention proposes a cross-medium control method for a water-air multi-rotor UAV using the above system, comprising the following steps:

[0030] S1, determining whether the control instruction received by the main controller is a water entry instruction or a water exit instruction;

[0031] If it is a water entry instruction, jump to step S2;

[0032] If it is a water outlet instruction, jump to step S7;

[0033] S2, using a humanoid intelligent controller to perform multi-feature fusion pattern recognition based on the body information and environmental information obtained from the monitoring subsystem and the sensor subsystem to determine whether to enter the water in the current environment;

[0034] If yes, jump to step S3;

[0035] Otherwise jump to step S5;

[0036] S3, using a humanoid intelligent controller to determine the UAV's water entry direction, water entry posture and water entry speed according to environmental information, and control the UAV's posture, heading and speed;

[0037] S4: After detecting that the drone has touched the water surface, the humanoid intelligent controller is used to adjust the drone's posture to be coplanar with the horizontal plane and make the drone sink into the water by its own gravity, and the process ends;

[0038] S5, using humanoid intelligent controller to judge environmental information;

[0039] If it is a hardware failure, an alarm will be sent through the communication subsystem and the aircraft will return home autonomously;

[0040] If it is due to environmental factors, the communication subsystem transmits information that the current water environment is poor and autonomously searches for surrounding waters suitable for entering the water, and then jumps to step S6;

[0041] S6, judging whether the current environment meets the water outlet requirements;

[0042] If yes, jump to step S3;

[0043] Otherwise, return home autonomously;

[0044] S7, receiving the water exit command through the main controller, controlling the UAV to drive to the horizontal plane and using the humanoid intelligent controller to adjust the attitude of the UAV so that the fixed rotor surface is the top surface and coplanar with the horizontal plane;

[0045] S8, using a humanoid intelligent controller to perform multi-feature fusion pattern recognition based on the body information and environmental information obtained from the monitoring subsystem and the sensor subsystem to determine whether water is currently being discharged in the environment;

[0046] If yes, jump to step S9;

[0047] Otherwise jump to step S10;

[0048] S9, using a humanoid intelligent controller to determine the water exit direction, water exit posture and water exit speed of the UAV according to environmental information, and controlling the posture, heading and speed of the UAV to make the UAV leave the water;

[0049] S10, using a humanoid intelligent controller to send a water outlet environment search instruction to the main controller to determine whether the current environment meets the water outlet requirements;

[0050] If yes, a water exit command is sent to the main controller to complete the drone exiting the water;

[0051] Otherwise, send an alarm through the communication subsystem and maintain its own posture.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] (1) The cross-medium control method and system of a water-air multi-rotor UAV that combines human-like intelligent control with anti-disturbance control provided by the present invention can be adapted to a variety of rotor-type water-air amphibious UAVs.

[0054] (2) The cross-media control system and method of the water-air multi-rotor UAV that combines humanoid intelligent control with anti-disturbance control provided by the present invention retains the original intelligent decision-making characteristics of humanoid intelligent control, online feature recognition and feature memory, and multi-modal control that combines qualitative decision-making with quantitative control, adopts adaptive control strategies for different feature states, and has strong adaptability to complex environments.

[0055] (3) The present invention adopts an anti-disturbance control controller to improve the traditional motion control level of human-like intelligent control, which effectively prevents overshoot of the control quantity, increases the limit of the influence of noise and environmental disturbances, and utilizes the anti-disturbance property of the anti-disturbance control controller itself to improve the robustness of the control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a schematic diagram of the structure of a water-air multi-rotor UAV in Example 1 of the present invention;

[0057] Figure 2 This is a schematic diagram of the structure of a water-air multi-rotor UAV control system in Example 1 of the present invention;

[0058] Figure 3 This is a schematic diagram of the structure of a humanoid intelligent controller in Embodiment 2 of the present invention;

[0059] Figure 4 Schematic diagram of the structure of the active disturbance rejection controller in Example 2 of the present invention;

[0060] Figure 5 This is a schematic diagram of a humanoid intelligent control characteristic model in Example 2 of the present invention;

[0061] Figure 6 This is a schematic diagram of the structure of the human-like intelligent control expert system in Example 3 of the present invention;

[0062] Figure 7 Schematic diagram of the cross-medium control method of the water-air multi-rotor UAV in Example 5 of the present invention.

[0063] The accompanying drawings are marked as follows: 1. frame, 2. fixed rotor, 3. waterproof rotor motor, 4. waterproof compartment. DETAILED DESCRIPTION

[0064] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0065] Example 1

[0066] This embodiment is specifically directed to a water-air multi-rotor drone, such as Figure 1 As shown, the structure of the water-air six-rotor UAV includes a frame 1, multiple fixed rotors 2, multiple waterproof rotor motors 3 for driving the fixed rotors, and a waterproof compartment 4 for installing components such as a main controller.

[0067] This embodiment provides a cross-medium control system for a water-to-air multi-rotor UAV that combines human-like intelligent control and anti-disturbance control for the above-mentioned water-to-air multi-rotor UAV. Figure 2 As shown, the control system includes a main controller, a power subsystem, a monitoring subsystem, a sensor subsystem, and a communication subsystem arranged on the water-air multi-rotor UAV;

[0068] The power subsystem is used to provide a driving signal to the water-air multi-rotor UAV and to supply power to the main controller; it specifically includes a power module connected to the main controller, a fixed rotor, a waterproof rotor motor, and a rotor motor drive module. Specifically, the waterproof rotor motor is connected to the main controller through the rotor motor drive module; the power module uses a large-capacity lithium battery to supply power to the waterproof rotor motor, and supplies power to the main controller through a voltage regulator and step-down module; the main controller outputs a PWM control signal to the rotor motor drive module to control the speed of the waterproof rotor motor to control the heading and attitude of the UAV.

[0069] The monitoring subsystem is used to monitor the internal and external environmental information of the water-air multi-rotor UAV in real time; it specifically includes a wind speed and direction sensor, a humidity and temperature sensor, and a power supply and power detection module connected to the main controller. Specifically, the wind speed and direction sensor, the humidity and temperature sensor, and the power supply and power detection module in the monitoring subsystem are connected to the main controller. The wind speed and direction sensor is used to detect the wind speed and direction data of the ambient airspace; the humidity and temperature sensor is used to monitor the internal humidity of the control box and the temperature of the main controller; the power supply and power module is used to collect battery status such as battery voltage, current current, power, and remaining power.

[0070] The sensor subsystem is used to obtain the body information of the water-air multi-rotor UAV and the navigation information in the underwater environment; it specifically includes an inertial measurement sensor connected to the main control, a GPS positioning module, and an underwater acoustic positioning module. Specifically, the inertial measurement sensor includes a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer, which are used to collect the attitude and position information of the UAV; the GPS positioning module is used to obtain the longitude and latitude and altitude information of the UAV; the underwater acoustic positioning module 7 is used to provide accurate acoustic heading, absolute position and underwater speed for the underwater body.

[0071] The communication subsystem is used to receive external command information, and transmit information such as battery power, humidity inside the control box and temperature of the main controller, body posture, and position to the outside; it specifically includes a wireless communication module and a wired communication interface connected to the main controller. Specifically, the wireless communication module and wired communication interface are used for the drone to receive command information from the outside, and transmit information such as battery power, humidity inside the control box and temperature of the main controller, body posture, and position to the outside.

[0072] The main controller is used to use a humanoid intelligent controller to perform multi-feature fusion pattern recognition on the water-air multi-rotor UAV according to the information obtained by the monitoring subsystem and the sensor subsystem, determine the optimal heading and expected attitude to be taken by the water-air multi-rotor UAV in the next step and the optimal control strategy to be adopted for adjusting the attitude, and control it by improving the motion control level.

[0073] Example 2

[0074] like Figure 3 As shown, the humanoid intelligent controller in this embodiment starts from the perspective of macroscopic structure simulation and behavioral function simulation of the human brain, simulates the empirical behavior and knowledge structure of experts in the field, establishes an expert knowledge model that combines qualitative and quantitative dynamic features of system feature identification, combines the dynamic feature model of the controlled system and the controller qualitatively and quantitatively, performs online feature identification and feature memory according to the real-time feature state of the controlled system, and uses a heuristic reasoning rule set to determine the optimal control strategy. During the cross-medium process of the water-air multi-rotor UAV, the humanoid intelligent control expert system determines the optimal heading and expected attitude of the water-air multi-rotor UAV by performing multi-feature fusion pattern recognition on the internal and external environmental information obtained by the UAV monitoring subsystem and sensor subsystem and its own posture, position, altitude and other information, and controls it through the motion control level.

[0075] In this embodiment, the humanoid intelligent controller includes a task adaptation level, a parameter correction level, and an improved motion control level;

[0076] The task adaptation level is used to use an expert system to perform a fusion analysis on the information obtained by the monitoring subsystem and the sensor subsystem to determine the optimal speed, posture or other operations that the water-air multi-rotor UAV needs to achieve in the face of the current situation;

[0077] The parameter correction stage is used to correct the threshold value in the feature model and adjust the parameters of the motion control stage under different feature states;

[0078] The improved motion control stage is used to adopt an anti-disturbance controller, estimate the real-time attitude and total system disturbance of the water-air multi-rotor UAV through an extended state observer, and compensate the control rate according to the stable system input signal, the estimated real-time attitude and the total disturbance of the system through nonlinear error feedback, and generate the final control quantity of the water-air multi-rotor UAV system. Finally, the system input signal is tracked and differentiated by a tracking differentiator to obtain a stable system input signal.

[0079] The control prototype algorithm of traditional human-like intelligent control is prone to overshoot of the control system and does not fully simulate human control strategies. At the motion control level, since the switching of control modes depends on the error of the controlled variable and its rate of change, the impact of noise on system stability is increased, and the water-air multi-rotor UAV will inevitably be affected by noise and environmental disturbances during the cross-medium process. If the traditional motion control level is still used, the control accuracy and stability will be difficult to meet the ideal requirements.

[0080] In order to solve the problems of overshoot of control quantity, low control accuracy and low robustness when facing complex environmental disturbances such as noise in the above-mentioned traditional motion control level, this embodiment adopts an anti-disturbance control controller to improve the traditional motion control level to execute the control strategy determined by the upper-level algorithm of human-simulated intelligent control.

[0081] The ADRC has the advantages of strong anti-interference ability and does not rely on accurate system models. The tracking differentiator of the ADRC prevents the occurrence of overshoot and makes the control effect more stable. The state observer of the ADRC can also observe the real-time state of the controlled variable and the total disturbance of the system and eliminate the total disturbance by compensating the control rate through the nonlinear error feedback law. The ADRC itself has self-interference resistance, which makes up for the shortcomings of the traditional motion control level of the original human-like intelligent control.

[0082] In this embodiment, the anti-disturbance controller includes a tracking differentiator (TD), an extended state observer (ESO) and a nonlinear error feedback (NLSEF); the tracking differentiator TD is used to track and differentiate the system input signal to obtain a stable system input signal; the extended state observer ESO is used to estimate the real-time state of the attitude angle of the water-to-air multi-rotor UAV and the total system disturbance; the nonlinear error feedback NLSEF is used to compensate the control rate according to the stable system input signal, the estimated real-time attitude and the total disturbance of the system, and generate the final control quantity of the water-to-air multi-rotor UAV.

[0083] The anti-disturbance controller used in this embodiment has the characteristics of multi-modal control. For the second-order nonlinear system, ADRC uses the extended state observer (ESO) to observe the actual value of the regulated variable and the total disturbance of the system, and uses the control quantity output by nonlinear error feedback to overcome the system disturbance so that the actual controller becomes an ideal series double integrator, and finally adopts proportional differential modal control. The mathematical model of the water-air multi-rotor UAV is a second-order nonlinear mathematical model. The multi-modal control can be realized by correcting the proportional differential coefficient of the ADRC control subsystem through the parameter correction layer of the human-like intelligent control subsystem.

[0084] The input of the tracking differentiator is the desired attitude angle θ o (desired pitch angle, desired roll angle or desired yaw angle), the output is θ o1 and θ o2 θ o1 and θ o2 They are θ o The tracking signal and the differential signal of:

[0085] d=rh0

[0086] d0=h0d

[0087]

[0088] y=θ o1 +h0θ o2

[0089]

[0090]

[0091] Among them, h0 is the filter factor, r is the fast factor, the larger h0 is, the smaller the static error is, the smaller the overshoot is at the beginning, the smaller the initial error is, and the better the filtering effect is, but it will cause the output signal to rise too slowly, the fastness is not good, and the tracking signal phase loss is more; the larger the fast factor r is, the faster the output value and the target value are tracked, but it will amplify the noise, and sign(·) is the sign function. The d, d0, a0, a and y calculated by the two parameters r and h0 to be adjusted are used to define the fhan(·) function. fhan(·) is the fastest tracking function, which is mainly used to avoid the chatter phenomenon in steady state.

[0092] Then the second-order continuous maximum feedback system of the tracking differentiator is obtained as:

[0093]

[0094]

[0095] θ o1 (t+1)=θ o1 (t)+hθ o2 (t)

[0096] θ o2 (t+1)=θ o2 (t)+h·fhan(θ o1 -θ o ,r,h0)

[0097] in, and is θ o1 and θ o2 The derivative of can be integrated to get the next moment θ o1 and θ o2 ; h is the operation step; t is the sampling time; θ o1 (t) and θ o2 (t) are the tracking input signal, θ o1 (t) is the first-order differential signal.

[0098] The input of the extended state observer is the actual attitude angle (pitch angle, roll angle, yaw angle) of the water-air multi-rotor aircraft, the product of the system control quantity U and the coefficient b0. The output is the observed attitude angle value z1, the observed attitude angle change rate z2 and the observed total disturbance z3. In the extended state observer ESO, the observed value is regarded as the actual value, the observed change rate is regarded as the actual change rate, and the observed total disturbance is the total internal and external disturbance suffered by the system. Divided by b0, the output control quantity u0 of the state error feedback control law is subtracted to obtain the final control quantity U of the system.

[0099] U(t)=u0(t)-z3(t) / b0

[0100] Among them, t is the sampling time, U(t) is the final control output of the ADRC system at time t, u0(t) is the control quantity output by the state error feedback control law at time t, z3(t) is the total system disturbance observed at time t, and b0 is the output compensation factor.

[0101] The extended state observer is expressed as:

[0102] e(t)=z1(t)-x1(t)

[0103]

[0104]

[0105]

[0106] Where t is the sampling time, e(t) is the deviation between x1(t) (drone attitude angle) and ESO observation value z1(t) at time t; z2(t) is the observation value of x2(t) (drone attitude angle change rate); z3(t) is the observation value of x3(t) (system total disturbance); and are the derivatives of z1(t), z2(t) and z3(t), respectively, which can be integrated to obtain z1, z2 and z3 at the next moment; β 01 ,β 02 ,β 03 is the observer gain parameter to be adjusted and is positive, β 01 Too large will cause oscillation or even divergence, β 02 Too small will cause the output to diverge, too large will produce dense high-frequency noise; β 03 If it is too large, the system will oscillate; if it is too small, the tracking speed of the output curve will be reduced; α1, α2 are nonlinear factors, usually α1 = 0.5, α2 = 0.25; δ is the filter factor, in order to ensure the linear interval width of the fal(·) function, generally δ = 0.01; fal(·) is a nonlinear function, the form is:

[0107]

[0108] Among them, e is the value of e(t), and α is α1 and α2.

[0109] The nonlinear error feedback input is the attitude angle error e1 (tracking attitude angle θ o1 -Observed attitude angle z1), rate of change error e2 (tracking attitude angle change rate θ o2 -Observe the attitude angle change rate z2), output the control value u0. The nonlinear error feedback is processed as follows:

[0110] e1=θ o1 -z1

[0111] e2=θ o2 -z2

[0112] u0=β1fal(e1,α1,δ)+β2fal(e2,α2,δ)

[0113] Among them, u0 is the error feedback control quantity, β1, β2 are the gain parameters to be adjusted, e1 is the system attitude angle deviation, and e2 is the system attitude angle change rate deviation.

[0114] The characteristic model of the parameter correction level is basically the same as the characteristic model of the traditional operation control level. Figure 5 In order to improve the control accuracy, the parameters involved in the motion control level are as follows: p 、w d and k et al.

[0115] Let the feature element set of parameter correction level be:

[0116] Q={q1,q2,q3,q4,q5,q6,q7,q8}

[0117] in:

[0118] q1={|e|≥e1}, q2={|e|≥e4}, q3={|e|≥e2}, q4={|e|≥e3}

[0119]

[0120] In the formula, Q represents the feature element set, and contains q1 to q88 subsets; e and They represent the error and error change rate between the actual attitude angle and the expected attitude angle of the UAV during the motion, e1, e2, e3, e4 and They are Figure 5The specific values ​​of the characteristic values ​​indicated in the schematic diagram of the human-simulated intelligent control characteristic model can be given by experimental data.

[0121] The characteristic model of the parameter correction stage is:

[0122]

[0123] in, Represents a feature set and contains and Four elements. Specifically:

[0124]

[0125]

[0126]

[0127]

[0128] In the formula, q1~q8 are elements of the feature element set Q. is the inverse set of q1 to q8, Same reason.

[0129] The decision mode set of parameter correction level is:

[0130] ψ1={ψ 11 ,ψ 12 ,ψ 13}

[0131] in:

[0132] ψ 11 :{w p =w p1 ,w d =w d1 ,k1}

[0133] ψ 12 :{w p =w p2 ,w d =w d2 ,k2}

[0134] ψ 13 :{w p =w p3 ,w d =w d3 ,k3}

[0135] Among them, ψ1 and ψ 11 , 12 , 13 Represent the decision mode set and the elements of the set, wp and w d They are β1 and β2 of the nonlinear error feedback output in the motion control stage, and k1~k3 are other ADRC parameters.

[0136] The set of inference rules at the parameter correction level is:

[0137] Ω1={w 11 ,w 12 ,w 13 ,w 14}

[0138] w 11 :

[0139] w 12 :

[0140] w 13 :

[0141] w 14 :

[0142] In the formula, Ω1 and w 11 、w 12 、w 13 、w 14 Represent the inference rule set and the elements of the set, respectively. and and ψ 11 , 12 , 13 The feature set and elements of the decision modal set ψ1.

[0143] Example 3

[0144] The task adaptation level in the humanoid intelligent controller of this embodiment is located above the motion control level and the parameter correction level. As the highest level of the entire control system, the task adaptation level uses an expert system to achieve the best countermeasures for different situations, that is, through the fusion analysis of environmental data and fuselage-related data, it determines the optimal speed, posture or other operations (return, etc.) that the drone should achieve in the current situation (such as entering the water).

[0145] In this embodiment, the expert system of the human-like intelligent controller includes a knowledge acquisition structure, an expert knowledge base, a comprehensive database and an inference engine. Figure 6 As shown;

[0146] The knowledge acquisition structure is used to acquire expert knowledge and store it in the expert knowledge base according to the set rules;

[0147] The expert knowledge base is used to represent the stored expert knowledge in a set form and distinguish whether the expert knowledge is quantifiable;

[0148] The comprehensive database is used to store the real-time monitoring data of the monitoring subsystem and the sensor subsystem on the water-air multi-rotor UAV;

[0149] The inference engine is used to match and solve the initial facts provided by the comprehensive database with the knowledge in the expert knowledge base, and output the countermeasure with the highest matching degree to the main controller.

[0150] In this embodiment, the knowledge acquisition structure of the expert system adopts a semi-automatic knowledge acquisition method, including a traditional manual method and a data mining method.

[0151] The traditional method is that engineers extract knowledge from domain experts and express the knowledge using generative knowledge representation methods. Then the knowledge acquisition structure converts the knowledge into a computer-storable form and stores it in the knowledge base.

[0152] Data mining methods include principal component analysis, which collects previous UAV experimental data, including data such as speed, posture, wind speed in the working area when crossing the medium, and the effects of control methods used in special cases such as facing high wind speed environments. The sample data is reduced in dimensionality and the contribution rate of each data indicator is analyzed to screen out knowledge with high credibility and give it priority in reasoning. Principal component analysis is a commonly used statistical method and will not be described in detail here.

[0153] The knowledge acquisition structure is used to acquire highly credible knowledge to improve the control performance of the controller. After the knowledge is acquired, it is stored in the knowledge base according to certain rules. The rules are shown in the following formula:

[0154] General form:

[0155] IF A i THEN B

[0156] Complex form:

[0157] IF(A1 or A2 or A3...or A i )THEN B

[0158] IF(A1 and A2 and A3...and A i )THEN B

[0159] Among them is A i Known facts, B is the corresponding method, For confidence.

[0160] In this embodiment, the knowledge base of the expert system uses a generative knowledge expression method to represent knowledge, and its establishment method is as follows:

[0161] (1) Based on the traditional method of knowledge acquisition structure, the problems that may be faced when crossing media include poor airspace environment data, the optimal cross-media body posture and speed of the drone under a certain wind speed, the ADRC parameters with good operating effects to maintain the drone posture and speed in complex environments, and some special operations for special situations (such as returning home, maintaining the current state and waiting for staff), etc., to obtain a large amount of drone flight and operation information, screen out valuable experience, and analyze some cross-media failure cases, summarize experience and lessons to obtain the optimal water entry strategy. Finally, these experience knowledge are summarized and stored in the knowledge base through knowledge acquisition structure transformation.

[0162] (2) Based on the data mining method of the knowledge acquisition structure, samples of data from simulations or experiments are collected, including the attitude and speed of the UAV across the medium during simulation, ADRC control parameters, the time required to achieve stable control using a certain set of parameters, and environmental data during the experiment. A principal level analysis is performed to obtain the impact of each factor on water entry. Based on the results, a theoretically better control strategy is constructed, and experiments are conducted to screen out the control strategies that perform better in the experiment and store them in the knowledge base.

[0163] The knowledge in the knowledge base is expressed as follows:

[0164] IF P THEN Q

[0165] in:

[0166] P is the premise, Q is the conclusion, is the confidence level, P and Q both contain multiple data items. A value between 0 and 1.

[0167] Each data item of knowledge in the knowledge base should also be distinguished as to whether it is quantifiable. For non-quantifiable knowledge, a fuzzy judgment method is needed to ensure that the inference engine obtains a distinguishable matching degree when making a matching degree judgment. The specific rules for fuzzy matching degree are as follows:

[0168] The matching degree is sorted as follows: very good match, relatively good match, average match, and slightly good match.

[0169] In addition, the expert knowledge base requires regular maintenance by engineers or staff to remove knowledge that is no longer applicable and add knowledge with better effects, so as to ensure that the knowledge in the expert knowledge base can meet the needs of domain problems.

[0170] In this embodiment, the comprehensive database data source of the expert system is the sensor subsystem and monitoring subsystem of the UAV, which stores various information obtained from the sensor subsystem and the monitoring subsystem during the operation of the UAV, including the body posture, body speed, altitude, battery power, wind speed and direction in the operating airspace, control box humidity and temperature, etc. When the UAV receives a certain instruction through the communication module or detects that a certain data is abnormal under the current working conditions, it provides data to the inference engine.

[0171] In this embodiment, the inference engine of the expert system adopts the forward reasoning method, starting from the initial facts provided by the comprehensive database, matching and solving with the knowledge in the expert knowledge base, and outputting the countermeasure with the highest matching degree; when a matching conflict occurs, that is, some data cannot be matched, the inference engine takes the approach of sorting according to the targetedness of the rules, that is, sorting the currently determined data items (posture, speed, etc.) according to their impact on the current movement, giving priority to matching the data items with higher rankings, and finally outputting a relatively better countermeasure.

[0172] Specifically, the inference engine is a set of executable programs that matches the facts in the comprehensive database with the premises of the knowledge in the expert knowledge base. i , the corresponding data item b of knowledge premise B in the knowledge base i , use the difference method or fuzzy judgment method to obtain the corresponding matching degree, the rules are as follows:

[0173] For quantified data items:

[0174] λ=|a i -b i |

[0175] Where λ is the corresponding matching degree, and the smaller λ is, the higher the matching degree is.

[0176] For non-quantized data items:

[0177] The matching degree is differentiated according to the fuzzy judgment rules, that is, the levels of very matching, relatively matching, generally matching or slightly matching are obtained.

[0178] When a matching conflict occurs, the method of sorting by targeted rules is used to continue to resolve the conflict. For example, when entering the water, the currently provided data items are not comprehensive. According to the knowledge in the knowledge base, in a certain environment, when entering the water, the drone should maintain a certain posture or a certain attitude angle, which has the greatest impact on whether the drone can enter the water stably, followed by the speed of entering the water, etc. Sort the influence of each data item on the movement to be performed by the drone in this environment, and then match them according to the ranking to obtain the optimal countermeasure at this time.

[0179] Example 4

[0180] In this embodiment, in order to better understand the intelligent control of the water-to-air multi-rotor UAV's entry and exit processes by the water-to-air multi-rotor UAV cross-medium control system based on the combination of human-like intelligent control and self-disturbance rejection control, the following is further explained in combination with specific implementation methods.

[0181] (1) When controlling the entry of a multi-rotor drone into water,

[0182] The humanoid intelligent controller uses multi-feature fusion pattern recognition based on the body information and environmental information obtained from the monitoring subsystem and the sensor subsystem to make an intelligent judgment on whether to enter the water in the current environment, and issues two preliminary instructions to the main controller: to enter the water or not;

[0183] When the command is to enter the water, the humanoid intelligent controller is used to determine the UAV's water entry direction, water entry posture and water entry speed according to environmental information, and to control the UAV's posture, heading and speed; when the UAV is detected to be in contact with the water surface, the humanoid intelligent controller is used to adjust the UAV's posture to be coplanar with the horizontal plane and make the UAV sink into the water by its own gravity;

[0184] When the instruction is not to enter the water, the humanoid intelligent controller is used to judge the environmental information; if it is a hardware failure, an alarm is sent through the communication subsystem and the drone returns home autonomously; if it is due to environmental factors, the communication subsystem transmits information that the current water environment is poor and autonomously searches for surrounding waters suitable for entering the water. If a better environment suitable for entering the water is found, an entry instruction is issued, and the main controller controls the drone to execute the above entry instruction; otherwise, it is judged that the current environment is bad and it is impossible to enter the water, and the drone returns home autonomously.

[0185] (2) When controlling the water-to-air multi-rotor drone to exit the water,

[0186] The main controller receives the water exit command, controls the drone to drive to the horizontal plane, and uses the humanoid intelligent controller to adjust the drone's attitude so that the fixed rotor surface is the top surface and coplanar with the horizontal plane;

[0187] The humanoid intelligent controller is used to perform multi-feature fusion pattern recognition based on the body information and environmental information obtained from the monitoring subsystem and the sensor subsystem, to make an intelligent judgment on whether water is being discharged in the current environment, and to issue two preliminary instructions to the main controller: to discharge water or not to discharge water;

[0188] When the instruction is to exit the water, the humanoid intelligent controller is used to determine the exit direction, exit posture and exit speed of the UAV according to the environmental information, and to control the posture, heading and speed of the UAV to make the UAV leave the water;

[0189] When the instruction is not to exit the water temporarily, the humanoid intelligent controller sends an instruction to search for the water exit environment to the main controller; when the humanoid intelligent controller senses that the current environment meets the water exit requirements, it sends a water exit instruction to the main controller to complete the UAV's exit from the water; if the current environment does not meet the water exit requirements, an alarm is sent through the communication subsystem and its own posture is maintained.

[0190] Example 5

[0191] Based on the water-air multi-rotor UAV cross-medium control system based on the combination of human-like intelligent control and anti-disturbance control described in embodiments 1, 2, 3, and 4, this embodiment also provides a water-air multi-rotor UAV cross-medium control method using the above system, such as Figure 7 As shown, the following steps S1 to S10 are included:

[0192] S1, determining whether the control instruction received by the main controller is a water entry instruction or a water exit instruction;

[0193] If it is a water entry instruction, jump to step S2;

[0194] If it is a water outlet instruction, jump to step S7;

[0195] S2, using a humanoid intelligent controller to perform multi-feature fusion pattern recognition based on the body information and environmental information obtained from the monitoring subsystem and the sensor subsystem to determine whether to enter the water in the current environment;

[0196] If yes, jump to step S3;

[0197] Otherwise jump to step S5;

[0198] S3, using a humanoid intelligent controller to determine the UAV's water entry direction, water entry posture and water entry speed according to environmental information, and control the UAV's posture, heading and speed;

[0199] S4: After detecting that the drone has touched the water surface, the humanoid intelligent controller is used to adjust the drone's posture to be coplanar with the horizontal plane and make the drone sink into the water by its own gravity, and the process ends;

[0200] S5, using humanoid intelligent controller to judge environmental information;

[0201] If it is a hardware failure, an alarm will be sent through the communication subsystem and the aircraft will return home autonomously;

[0202] If it is due to environmental factors, the communication subsystem transmits information that the current water environment is poor and autonomously searches for surrounding waters suitable for entering the water, and then jumps to step S6;

[0203] S6, judging whether the current environment meets the water outlet requirements;

[0204] If yes, jump to step S3;

[0205] Otherwise, return home autonomously;

[0206] S7, receiving the water exit command through the main controller, controlling the UAV to drive to the horizontal plane and using the humanoid intelligent controller to adjust the attitude of the UAV so that the fixed rotor surface is the top surface and coplanar with the horizontal plane;

[0207] S8, using a humanoid intelligent controller to perform multi-feature fusion pattern recognition based on the body information and environmental information obtained from the monitoring subsystem and the sensor subsystem to determine whether water is currently being discharged in the environment;

[0208] If yes, jump to step S9;

[0209] Otherwise jump to step S10;

[0210] S9, using a humanoid intelligent controller to determine the water exit direction, water exit posture and water exit speed of the UAV according to environmental information, and controlling the posture, heading and speed of the UAV to make the UAV leave the water;

[0211] S10, using a humanoid intelligent controller to send a water outlet environment search instruction to the main controller to determine whether the current environment meets the water outlet requirements;

[0212] If yes, a water exit command is sent to the main controller to complete the drone exiting the water;

[0213] Otherwise, send an alarm through the communication subsystem and maintain its own posture.

[0214] Example 6

[0215] In order to better understand the cross-medium control method of water-air multi-rotor UAV based on the combination of human-like intelligent control and anti-disturbance control, the following is combined with the attached Figure 7 And specific implementation methods are further described.

[0216] (1) Water entry control

[0217] In the first step, the humanoid intelligent controller performs multi-feature fusion pattern recognition based on information such as its own posture, position, drone battery power, airspace wind speed and direction, control box humidity and temperature obtained from sensors, makes an intelligent judgment on whether to enter the water in the current environment, and sends two preliminary instructions to the main controller: to enter the water or not.

[0218] Step 2: The main controller determines the type of instruction received:

[0219] When the command is to enter the water, the humanoid intelligent controller determines the direction of the drone entering the water, that is, the angle of entry, as well as the posture and speed of entry according to the intelligent decision of the task adaptive level. The main controller controls the speed of the waterproof rotor motor through the rotor motor drive module, and controls the posture, heading and speed of the drone through ADRC. When the drone touches the water, the speed of the waterproof rotor motor is reduced so that the drone lands smoothly on the water surface and then sinks into the water by its own gravity to complete the entry into the water.

[0220] When the instruction is not to enter the water, the humanoid intelligent controller judges the sensor data. If the entry into the water cannot be carried out due to hardware failure or other problems, the UAV will alert the staff through the wireless communication module and return home autonomously; if the current water area cannot be entered due to poor environmental data, the UAV will transmit the information that the current water environment is poor to the staff through the wireless communication module and autonomously search for surrounding water areas suitable for entering the water. If a better environment suitable for entering the water is found, the UAV will issue a water entry instruction to the main controller, and the main controller will control the UAV to execute the above water entry instruction; otherwise, it will be determined that the current environment is too bad to enter the water, and the UAV will return home autonomously.

[0221] (2) Water outlet control

[0222] In the first step, after the main controller receives the water exit command, it controls the UAV to drive towards the horizontal plane and adjusts the UAV attitude through ADRC to fix the rotor surface as the top surface and coplanar with the horizontal plane.

[0223] In the second step, the humanoid intelligent controller performs multi-feature fusion pattern recognition on its own posture, drone battery power, and airspace wind speed and direction information obtained from the sensor, and sends a signal to the main controller as to whether water discharge has been completed at the current position.

[0224] Step 3: The main controller determines the type of instruction received:

[0225] When the instruction is to exit the water, the humanoid intelligent controller determines the exit direction, i.e., the exit angle, as well as the exit posture and speed of the UAV based on the intelligent decision of the task adaptive level. The main controller controls the speed of the waterproof rotor motor through the rotor motor drive module, and controls the posture, heading and speed of the UAV through ADRC to make the UAV leave the water.

[0226] When the instruction is not to discharge water temporarily, the humanoid intelligent controller sends a command to the main controller to search for an environment suitable for discharge water. When the humanoid intelligent controller senses that the current environment is relatively suitable, it sends a discharge water command to the main controller to complete the discharge of the drone. If the current environment is too bad to discharge water successfully, the drone sends an alarm to the staff through the communication module and maintains its posture and waits for the staff

[0227] Example 7

[0228] In order to better understand the response of the ADRC motion control level to human-like intelligent control, the following is a further explanation in conjunction with a specific implementation method:

[0229] (1) Water entry control

[0230] When the humanoid intelligent control expert system makes the initial command not to enter the water:

[0231] When the expert system determines that the problem is a hardware failure, it conveys the return command to the main controller. At this time, based on the knowledge of the humanoid intelligent control expert system, it first obtains the current position through the GPS module to obtain the return path, obtains the current nose direction through the sensor system, and then gives the angle that should be turned to make the drone nose face the starting point. Then the main controller adjusts the drone's attitude through ADRC to make its nose face the starting point. When it is detected that the nose rotation is completed, the ADRC control parameters for rapid return are obtained based on the knowledge of the humanoid intelligent control expert system. The main controller uses the ADRC with the control parameters to control the drone to fly to the starting point and complete the return.

[0232] When the expert system determines that the environmental data is poor, it transmits the command of finding a more suitable environment to the main controller, and determines whether to search based on the current power measured by the power module according to the expert system knowledge base. The search area is determined according to the power, specifically, the search is carried out around a certain radius circle in the initially selected water entry airspace, and the radius of the circle gradually increases. The radius of the circle is determined according to the current power status, and the radius adopted should ensure that the environmental data will change during the search. The maximum power consumption of the search should not include the power consumed by the return of the drone and the two cross-medium crossings of the drone. When the environmental data cannot be transformed into more suitable data during and after the search and it is judged that no search is to be carried out, the return command is transmitted to the main controller, and the drone returns using ADRC control according to the return operation; when the environment is detected to be suitable during the search process, the water entry command is transmitted to the main controller.

[0233] When the humanoid intelligent control gives the initial instruction to enter the water:

[0234] The humanoid intelligent control expert system matches the entry and exit strategies for the current situation, including the expected entry posture, speed, and anti-disturbance parameters. The main controller controls the drone through ADRC to adjust the posture and flight speed to achieve the expected posture for entering the water. After contacting the horizontal plane, the humanoid intelligent control expert system matches the control strategy of cross-medium transition to make the top surface of the drone flush with the water surface, allowing it to sink by its own gravity. When sinking, it is detected that the drone's condition has changed suddenly, and the expert system quickly matches countermeasures. For example, if the drone's posture suddenly changes due to wind and waves, it is not conducive to the execution of the next task. The expert system matches the posture that the drone should reach at the next moment to resist the disturbance caused by the sudden change until it runs stably. The main controller controls the drone's posture through ADRC.

[0235] When entering the water, the parameter correction layer of the human-like intelligent control adjusts the ADRC parameters online according to the characteristic model, making the control effect more accurate and stable.

[0236] (2) Water outlet control

[0237] When the drone is out of water:

[0238] The humanoid intelligent control expert system determines the speed and attitude that the UAV should take to reach the horizontal plane as quickly as possible based on the water flow conditions. The main controller then controls the UAV through ADRC to perform corresponding operations to reach the horizontal plane and make the top surface of the UAV flush with the horizontal plane.

[0239] When the UAV leaves the water, due to the sudden change of the environment, the data in the comprehensive database of the expert system will mutate. At this time, the inference engine quickly infers and obtains the corresponding strategy, giving the expected speed and posture that the UAV should achieve to counteract this change. The ADRC control of the parameter correction layer online corrects the parameters to achieve the expected speed and posture of the UAV, so that the top surface of the UAV is flush with the horizontal plane.

[0240] The expert system determines the desired attitude and speed for rapid water exit based on airspace data. The main controller makes the UAV fly out of the water after ADRC adjustment. At this time, the attitude and speed of the UAV are disturbed due to the sudden change of environment and the disturbance of water droplets adhering to the fuselage, which is inconsistent with the expected attitude and even causes a mutation. In response to these situations, the expert system will obtain the corresponding strategy based on the expert knowledge base, and the parameter correction layer will perform online correction on the ADRC controller parameters to enable the UAV to smoothly and quickly leave the water and complete the water exit.

[0241] Finally, wait for instructions from the staff, including returning, waiting on the spot, etc.

[0242] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0243] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0244] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0245] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

[0246] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims

1. A cross-medium control system for a water-air multi-rotor UAV, characterized in that: It includes a main controller, a power subsystem, a monitoring subsystem, a sensor subsystem, and a communication subsystem installed on the water-air multi-rotor UAV; The power subsystem is used to provide a driving signal to the water-air multi-rotor UAV and to supply power to the main controller; The monitoring subsystem is used to monitor the internal and external environmental information of the water-air multi-rotor UAV in real time; The sensor subsystem is used to obtain the body information of the water-air multi-rotor UAV and the navigation information in the underwater environment; The communication subsystem is used to receive external command information and transmit battery power, humidity inside the control box and temperature of the main controller, body posture, and position information; The main controller is used to use a humanoid intelligent controller to perform multi-feature fusion pattern recognition on the water-air multi-rotor UAV according to the information obtained by the monitoring subsystem and the sensor subsystem, determine the optimal heading and expected attitude to be taken by the water-air multi-rotor UAV in the next step and the optimal control strategy to be adopted for adjusting the attitude, and control it by improving the motion control level; The humanoid intelligent controller includes a task adaptation level, a parameter correction level and an improved motion control level; The task adaptation level is used to use an expert system to perform a fusion analysis on the information obtained by the monitoring subsystem and the sensor subsystem to determine the optimal speed, posture or other operations that the water-air multi-rotor UAV needs to achieve in the face of the current situation; The parameter correction stage is used to correct the threshold value in the feature model and adjust the parameters of the motion control stage under different feature states; The improved motion control stage is used to adopt an anti-disturbance controller, estimate the real-time attitude and total system disturbance of the water-air multi-rotor UAV through an extended state observer, and compensate the control rate according to the stable system input signal, the estimated real-time attitude and the total disturbance of the system through nonlinear error feedback, and generate the final control quantity of the water-air multi-rotor UAV system. Finally, the system input signal is tracked and differentiated by a tracking differentiator to obtain a stable system input signal.

2. The cross-medium control system of a water-air multi-rotor UAV according to claim 1 is characterized in that: The expert system includes a knowledge acquisition structure, an expert knowledge base, a comprehensive database and an inference engine; The knowledge acquisition structure is used to acquire expert knowledge and store it in the expert knowledge base according to the set rules; The expert knowledge base is used to represent the stored expert knowledge in a set form and distinguish whether the expert knowledge is quantifiable; The comprehensive database is used to store the real-time monitoring data of the monitoring subsystem and the sensor subsystem on the water-air multi-rotor UAV; The inference engine is used to match and solve the initial facts provided by the comprehensive database with the knowledge in the expert knowledge base, and output the countermeasure with the highest matching degree to the main controller.

3. The cross-medium control system of a water-air multi-rotor UAV according to claim 1 is characterized in that: When controlling the entry of a multi-rotor drone into water, The humanoid intelligent controller uses multi-feature fusion pattern recognition based on the body information and environmental information obtained from the monitoring subsystem and the sensor subsystem to make an intelligent judgment on whether to enter the water in the current environment, and issues two preliminary instructions to the main controller: to enter the water or not; When the command is to enter the water, the humanoid intelligent controller is used to determine the UAV's water entry direction, water entry posture and water entry speed according to environmental information, and to control the UAV's posture, heading and speed; when the UAV is detected to be in contact with the water surface, the humanoid intelligent controller is used to adjust the UAV's posture to be coplanar with the horizontal plane and make the UAV sink into the water by its own gravity; When the instruction is not to enter the water, the humanoid intelligent controller is used to judge the environmental information; if it is a hardware failure, an alarm is sent through the communication subsystem and the craft returns autonomously; If it is due to environmental factors, the communication subsystem will transmit information that the current water environment is poor and autonomously search for surrounding waters suitable for entry. If a better environment suitable for entry is found, an entry command will be issued and the main controller will control the drone to execute the above entry command; otherwise, it will be judged that the current environment is bad and entry is impossible, and it will return autonomously.

4. The cross-medium control system of a water-air multi-rotor UAV according to claim 1 is characterized in that: When controlling the water-to-air multi-rotor drone out of water, The main controller receives the water exit command, controls the drone to drive to the horizontal plane, and uses the humanoid intelligent controller to adjust the drone's attitude so that the fixed rotor surface is the top surface and coplanar with the horizontal plane; The humanoid intelligent controller is used to perform multi-feature fusion pattern recognition based on the body information and environmental information obtained from the monitoring subsystem and the sensor subsystem, to make an intelligent judgment on whether water is being discharged in the current environment, and to issue two preliminary instructions to the main controller: to discharge water or not to discharge water; When the instruction is to exit the water, the humanoid intelligent controller is used to determine the exit direction, exit posture and exit speed of the UAV according to the environmental information, and to control the posture, heading and speed of the UAV to make the UAV leave the water; When the instruction is not to exit the water temporarily, the humanoid intelligent controller sends an instruction to search for the water exit environment to the main controller; when the humanoid intelligent controller senses that the current environment meets the water exit requirements, it sends a water exit instruction to the main controller to complete the UAV's exit from the water; if the current environment does not meet the water exit requirements, an alarm is sent through the communication subsystem and its own posture is maintained.

5. A cross-medium control method for a water-air multi-rotor UAV using the system as claimed in any one of claims 1 to 4, characterized in that: The following steps are involved: S1, determining whether the control instruction received by the main controller is a water entry instruction or a water exit instruction; If it is a water entry instruction, jump to step S2; If it is a water outlet instruction, jump to step S7; S2, using a humanoid intelligent controller to perform multi-feature fusion pattern recognition based on the body information and environmental information obtained from the monitoring subsystem and the sensor subsystem to determine whether to enter the water in the current environment; If yes, jump to step S3; Otherwise jump to step S5; S3, using a humanoid intelligent controller to determine the UAV's water entry direction, water entry posture and water entry speed according to environmental information, and control the UAV's posture, heading and speed; S4: After detecting that the drone has touched the water surface, the humanoid intelligent controller is used to adjust the drone's posture to be coplanar with the horizontal plane and make the drone sink into the water by its own gravity, and the process ends; S5, using humanoid intelligent controller to judge environmental information; If it is a hardware failure, an alarm will be sent through the communication subsystem and the aircraft will return home autonomously; If it is due to environmental factors, the communication subsystem transmits information that the current water environment is poor and autonomously searches for surrounding waters suitable for entering the water, and then jumps to step S6; S6, judging whether the current environment meets the water outlet requirements; If yes, jump to step S3; Otherwise, return home autonomously; S7, receiving the water exit command through the main controller, controlling the UAV to drive to the horizontal plane and using the humanoid intelligent controller to adjust the attitude of the UAV so that the fixed rotor surface is the top surface and coplanar with the horizontal plane; S8, using a humanoid intelligent controller to perform multi-feature fusion pattern recognition based on the body information and environmental information obtained from the monitoring subsystem and the sensor subsystem to determine whether water is currently being discharged in the environment; If yes, jump to step S9; Otherwise jump to step S10; S9, using a humanoid intelligent controller to determine the water exit direction, water exit posture and water exit speed of the UAV according to environmental information, and controlling the posture, heading and speed of the UAV to make the UAV leave the water; S10, using a humanoid intelligent controller to send a water outlet environment search instruction to the main controller to determine whether the current environment meets the water outlet requirements; If yes, a water exit command is sent to the main controller to complete the drone exiting the water; Otherwise, send an alarm through the communication subsystem and maintain its own posture.

Images