An online lift compensation method for scene understanding
By constructing a dynamic model and designing voltage reduction observer and anti-interference control law, the problem of insufficient lift caused by battery voltage reduction by multi-rotor drones is solved, and the high accuracy of the drone in large maneuvers and long-term flight is achieved and stability control.
Patent Information
- Application Number
- CN202411227089.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-03
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-09-03
AI Technical Summary
In large maneuvering or long-term flight scenarios, multi-rotor drones are insufficient in lift due to reduced battery voltage, which affects flight stability and accuracy. The existing technology has not effectively solved this problem.
By theoretically analyzing the impact of battery voltage reduction on drones, building a dynamic model, and designing a voltage reduction observer and anti-interference control law, using the interference observer and PID control principle, real-time compensation of lift loss and decoupling lift loss With the drone state, an anti-interference control law is designed for voltage reduction.
Without increasing the load of the drone, high-precision flight and stable control are achieved, ensuring the safety and stability of the drone when the battery voltage is reduced, and improving the ability of large maneuvers and long-term flights.
Smart Images

Figure CN118732707B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of flying robot control, and specifically relates to an online lift compensation method for scene understanding, which is suitable for unmanned aerial vehicle control systems that need to be in large maneuvering flight or long-term flight scenarios and need to achieve stable and safe control. Background Art
[0002] In recent years, small drones with multi-rotors as driving mechanisms have been widely used in disaster warning, geological survey, emergency rescue, defense, power, aerial photography and other fields. When the drone is in a large maneuver flight or long-term flight and other similar scenarios, due to the reduction of the onboard battery voltage, the actual lift generated by each motor is difficult to reach the expected value, resulting in a significant reduction in the safety and stability of the rotor drone.
[0003] The control theory related to anti-interference of UAVs is relatively mature. At present, there are few control methods that have been experimentally verified for the interference caused by the reduction of battery voltage. Chinese patent application CN202010249028.4 proposes a lithium battery state estimation method for small rotary-wing pure electric UAVs, but does not involve the impact of changes in the lift of the UAV caused by changes in the battery state. Chinese patent application CN202211373979.8 proposes a rotor UAV control method based on a lift feedback power device, but it requires real-time measurement of lift with the help of force sensors at the hardware level, which increases the weight of the UAV. In addition, some scientific researchers have studied the impact of wind, rain and other types of interference on rotor UAVs from the field of control algorithms or have made contributions to the design of long-life batteries for rotor UAVs, but these works have failed to design corresponding anti-interference control laws for rotor UAVs in response to the impact of battery voltage reduction. Summary of the invention
[0004] In order to overcome the shortcomings of existing research content and methods, for a flying robot system based on a multi-rotor UAV, the present invention provides an online lift compensation method for scene understanding to solve the problem of insufficient lift and reduced trajectory flight accuracy caused by reduced battery voltage when the UAV performs large maneuvering flights or long-term flight missions. The present invention first conducts a theoretical analysis of some of the effects caused by the reduction in battery voltage of the UAV; then, on this basis, a UAV dynamics model that considers the impact of battery voltage reduction is constructed in combination with the basic dynamics model of the UAV. Secondly, in terms of UAV control, the voltage reduction observer that is decoupled from the UAV attitude is designed by using the concepts and experience related to interference observers and PID control, and the parameters of the observer are reasonably selected, and on this basis, an anti-interference control law for voltage reduction is designed.
[0005] The solution provided by the present invention can realize real-time compensation of the lift loss of the UAV at the software level with the help of the control algorithm, and realize the decoupling between the lift loss estimation value of the UAV and the UAV state by using the lift loss model. Without increasing the load of the UAV, the high-precision flight and stability of the rotor UAV under the condition of reduced battery voltage are guaranteed to complete various operation tasks.
[0006] In order to achieve the above object, the present invention adopts the following technical scheme:
[0007] An online lift compensation method for scene understanding comprises the following steps:
[0008] The first step is to theoretically analyze the impact of battery voltage reduction on the UAV posture control;
[0009] The second step is to build a UAV dynamics model that takes into account the impact of battery voltage reduction;
[0010] The third step is to generate a smooth attitude control signal that can be achieved by the UAV according to the Hopf fiberization method;
[0011] The fourth step is to design the position loop voltage reduction disturbance observer based on the established UAV dynamics model;
[0012] The fifth step is to design the position loop control law based on the observed disturbance value to suppress the disturbance.
[0013] Furthermore, in the first step, the following theoretical analysis is performed on the relationship between the lift generated by the motor, the motor speed and the battery voltage:
[0014] During the flight of the rotor drone, the controller calculates and processes the expected signal and the real-time status information of the drone to obtain the expected ESC input signal. The ESC adjusts the output voltage or current according to the size of its input signal, thereby achieving motor speed control. However, the quantitative relationship between the ESC input signal and the output signal needs to be calibrated to achieve precise control of the motor speed. Generally, this quantitative relationship does not change with the drone control system signal.
[0015] As the onboard lithium battery drives the motor and generates circuit heat loss, its battery power will gradually decrease. As the onboard lithium battery power decreases, its output voltage decreases accordingly, and then the output motor operating voltage decreases when the electronic control command signal remains unchanged. For brushless motors, the number of rotations per minute is related to the motor input voltage, and the relevant parameter is the Kv value. The motor parameter Kv value describes the relationship between the peak voltage and speed of the motor under no-load conditions, and the unit is RPM / V, where RPM is the abbreviation of Revolutions Per Minute, which means the number of rotations per minute of the motor, and V represents the input voltage of the motor.
[0016] It can be seen that as the motor input voltage decreases, the motor speed decreases accordingly, and the drone control effect expected by the controller cannot be achieved. One of the most significant effects of voltage reduction is the reduction of the total lift of the drone, which specifically corresponds to the reduction of the lift of each motor, resulting in poor drone position control and attitude control, increasing the safety risk of drone flight.
[0017] In the present invention, it is considered that the voltage reduction has the most direct impact on the rotation speed of each motor of the rotary-wing UAV, which in turn affects the lift generated by each motor. This interference is mainly located in the position control loop of the UAV, that is, the position loop lift loss interference. The attitude loop influence caused by the different lift of each motor is not considered.
[0018] Furthermore, the second step includes:
[0019] The known dynamic equation of the rotor UAV is:
[0020] ,
[0021] in, is the body coordinate system, is the geographic coordinate system, Indicates the rotation matrix corresponding to the conversion of the body coordinate system coordinates into the geographic coordinate system coordinates. represents the total lift of the drone, Represents the thrust generated by all motors of the drone, Indicates the gravity acting on the drone. Indicates the quality of the drone, represents the acceleration due to gravity, represents the UAV’s moment of inertia, Indicates the rotational angular velocity of the drone, with the superscript represents the first-order derivative, such as express The first derivative of Indicates the control torque of the drone. Represents the transpose of a vector or matrix. Specifically, it can be expressed as:
[0022] .
[0023] Thus, the UAV dynamics model considering the impact of battery voltage reduction is obtained:
[0024] ,
[0025] in, Represents the lift loss vector of the rotorcraft UAV.
[0026] Furthermore, the third step includes:
[0027] The Hopf fiberization method is used to generate smooth attitude control signals that can be achieved by the UAV. The required auxiliary variables are known. and , , ,in The definition is as follows:
[0028] ,
[0029] Further, we can get , , And its first-order derivative is:
[0030] ,
[0031] .
[0032] From this, we can get the expected angular velocity signal of the rotorcraft for:
[0033] ,
[0034] in, , express , .
[0035] To obtain the desired attitude angle information, auxiliary variables can be used , Calculating quaternions , and then obtain the rotation matrix , auxiliary variable expressions and their relationship with quaternions The relationship is as follows:
[0036] ,
[0037] ,
[0038] .
[0039] Quaternions With the rotation matrix The mathematical details of the conversion are as follows:
[0040] .
[0041] Furthermore, from the rotation matrix element structure, we can get:
[0042] .
[0043] Due to the limited range of the actan function, the atan2 function is used instead of atan as the function for solving the inverse tangent value:
[0044] .
[0045] At this point, the desired signal for the attitude loop control of the rotorcraft UAV is obtained.
[0046] Furthermore, the fourth step includes:
[0047] The lift loss model decoupled from the attitude information of the rotorcraft is constructed as follows:
[0048] ,
[0049] in, Indicates the lift loss of the drone, which can be considered to be caused by the reduction of battery voltage and other factors.
[0050] First, the voltage drop observer is designed by using auxiliary variables , the observer gain matrix can be designed , that is, generate is a 3×3 square matrix of diagonal elements, and the observer structure is finally obtained:
[0051] ,
[0052] in, Indicates the speed of the drone in the geographic coordinate system, , The output of the observer Observed value of .
[0053] Furthermore, the fifth step includes: designing the position loop anti-interference control law:
[0054] make represents the position control error, where the subscript Indicates the expected signal, Indicates location information. represents the speed control error, Represents position information, and the position loop anti-interference control law can be designed as follows:
[0055] ,
[0056] in, and represents the controller gain, An estimate of the lift loss vector for the rotorcraft, provided by the voltage drop observer. represents the second-order derivative of the desired trajectory of the rotorcraft, that is, the expected acceleration corresponding to the desired trajectory, Represents the expected lift of the rotary wing UAV in the geographic coordinate system.
[0057] The beneficial effects of the present invention are:
[0058] The present invention can realize real-time compensation of lift when a rotor UAV encounters a battery voltage reduction problem, thereby ensuring its control capability of safe and stable flight in the air. Compared with the traditional anti-interference control method, the present invention first conducts a theoretical analysis of the influence of battery voltage reduction on the attitude control of a rotor UAV, and constructs a UAV dynamics model that considers the influence of battery voltage reduction. On this basis, the lift loss of the rotor UAV is decoupled from its own state, and a battery voltage reduction observer is designed. Finally, a rotor UAV control law that is resistant to battery voltage reduction interference is provided. The present invention can be applied to most existing rotor UAVs to truly give them the ability to fly with high maneuverability and high precision or for a long time, and ensure that the UAV has a high flight quality when performing large maneuverability flight or long-term flight missions. BRIEF DESCRIPTION OF THE DRAWINGS
[0059] Figure 1 The present invention is a flow chart of an online lift compensation method for scene understanding. DETAILED DESCRIPTION
[0060] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.
[0061] For the design of the drone controller, since the operating frequency of the attitude loop of the drone is much greater than that of the position loop, the controllers of the position loop and attitude loop can be designed separately according to the time scale separation principle. The present invention targets the interference effect of the battery voltage reduction, and believes that it is mainly concentrated in the position loop, so the present invention only targets the design of the position loop control law.
[0062] like Figure 1 As shown, an online lift compensation method for scene understanding of the present invention specifically includes the following steps:
[0063] The first step is to theoretically analyze the impact of battery voltage reduction on drone posture control:
[0064] During the flight of the rotor drone, the controller calculates and processes the expected signal and the real-time status information of the drone to obtain the expected ESC input signal. The ESC adjusts the output voltage or current according to the size of its input signal, thereby achieving motor speed control. However, the quantitative relationship between the ESC input signal and the output signal needs to be calibrated to achieve precise control of the motor speed. Generally, this quantitative relationship does not change with the drone control system signal.
[0065] As the onboard lithium battery drives the motor and generates circuit heat loss, its battery power will gradually decrease. As the onboard lithium battery power decreases, its output voltage decreases accordingly, and then the output motor operating voltage decreases when the electronic control command signal remains unchanged. For brushless motors, the number of rotations per minute is related to the motor input voltage, and the relevant parameter is the Kv value. The motor parameter Kv value describes the relationship between the peak voltage and speed of the motor under no-load conditions, and the unit is RPM / V, where RPM is the abbreviation of Revolutions Per Minute, which means the number of rotations per minute of the motor, and V represents the input voltage of the motor.
[0066] It can be seen that as the motor input voltage decreases, the motor speed decreases accordingly, and the drone control effect expected by the controller cannot be achieved. One of the most significant effects of voltage reduction is the reduction of the total lift of the drone, which specifically corresponds to the reduction of the lift of each motor, resulting in poor drone position control and attitude control, increasing the safety risk of drone flight.
[0067] In the present invention, it is considered that the voltage reduction has the most direct impact on the rotation speed of each motor of the rotary-wing UAV, which in turn affects the lift generated by each motor. This interference is mainly located in the position control loop of the UAV, that is, the position loop lift loss interference. The attitude loop influence caused by the different lift of each motor is not considered.
[0068] The second step is to build a rotorcraft UAV dynamics model based on theoretical analysis, taking into account the impact of battery voltage reduction:
[0069] The known dynamic equation of the rotor UAV is:
[0070] ,
[0071] in, is the body coordinate system, is the geographic coordinate system, Indicates the rotation matrix corresponding to the conversion of the body coordinate system coordinates into the geographic coordinate system coordinates. represents the total lift of the drone, Represents the thrust generated by all motors of the drone, Indicates the gravity acting on the drone. Indicates the quality of the drone, represents the acceleration due to gravity, represents the moment of inertia of the drone, Indicates the angular velocity of the drone, with the superscript represents the first-order derivative, such as express The first derivative of Indicates the control torque of the drone. Represents the transpose of a vector or matrix. Specifically, it can be expressed as:
[0072] .
[0073] Thus, the UAV dynamics model considering the impact of battery voltage reduction is obtained:
[0074] ,
[0075] in, Represents the lift loss vector of the rotor UAV.
[0076] The third step is to use the Hopf fiberization method to generate a smooth attitude control signal that the drone can achieve:
[0077] The required auxiliary variables are known and , , ,in The definition is as follows:
[0078] ,
[0079] Further, we can get , , And its first-order derivative is:
[0080] ,
[0081] .
[0082] From this, we can get the desired angular velocity signal of the rotorcraft for:
[0083] ,
[0084] in, , express , .
[0085] To obtain the desired attitude angle information, auxiliary variables can be used , Calculating quaternions , and then obtain the rotation matrix , auxiliary variable expressions and their relationship with quaternions The relationship is as follows:
[0086] ,
[0087] ,
[0088] .
[0089] Quaternions With the rotation matrix The mathematical details of the conversion are as follows:
[0090] ,
[0091] Furthermore, from the rotation matrix element structure, we can get:
[0092] .
[0093] Due to the limited range of the actan function, the atan2 function is used instead of atan as the function for solving the inverse tangent value:
[0094] .
[0095] At this point, the desired signal for the attitude loop control of the rotorcraft UAV is obtained.
[0096] The fourth step is to construct a lift loss model and design a battery voltage drop disturbance observer:
[0097] The lift loss model decoupled from the attitude information of the rotorcraft is constructed as follows:
[0098] ,
[0099] in, Indicates the lift loss of the drone, which can be considered to be caused by the reduction of battery voltage and other factors.
[0100] First, the voltage drop observer is designed by using auxiliary variables , the observer gain matrix can be designed , that is, generate is a 3×3 square matrix of diagonal elements, and the observer structure is finally obtained:
[0101] ,
[0102] in, Indicates the speed of the drone in the geographic coordinate system, , The output of the observer Observed value of .
[0103] The fifth step is to design the anti-interference control law of the position loop:
[0104] make represents the position control error, where the subscript Indicates the expected signal, Indicates location information. represents the speed control error, Represents position information, and the position loop anti-interference control law can be designed as follows:
[0105] ,
[0106] in, and represents the controller gain, represents the lift loss vector of the rotorcraft estimated by the voltage drop observer, represents the second-order derivative of the desired trajectory of the rotorcraft, that is, the expected acceleration corresponding to the desired trajectory, Represents the expected lift of the rotary wing UAV in the geographic coordinate system.
[0107] The contents not described in detail in the specification of the present invention belong to the prior art known to the professional and technical personnel in this field.
Claims
1. An online lift compensation method for scene understanding, characterized in that: The steps include: The first step is to theoretically analyze the impact of battery voltage reduction on the UAV posture control; The second step is to build a UAV dynamics model that takes into account the impact of battery voltage reduction; The third step is to generate a smooth attitude control signal that can be achieved by the UAV according to the Hopf fiberization method; The fourth step is to design the position loop voltage reduction disturbance observer based on the established UAV dynamics model; Step 5: Design the position loop control law based on the observed disturbance value to suppress the disturbance; The second step includes: The known dynamic equation of the rotor UAV is: , in, is the body coordinate system, Represents the body coordinate system axis, is the geographic coordinate system, Represents the geographic coordinate system axis; Indicates the acceleration of the drone, the superscript Indicates that the variable is a physical quantity in the geographic coordinate system. Indicates the rotation matrix corresponding to the conversion of the body coordinate system coordinates into the geographic coordinate system coordinates. represents the total lift of the drone, Indicates the thrust generated by all motors of the drone, with the superscript Indicates that the variable is a physical quantity in the body coordinate system Indicates the gravity acting on the drone. Indicates the quality of the drone, represents the acceleration due to gravity, represents the UAV’s moment of inertia, Indicates the rotational angular velocity of the drone, with the superscript represents the first-order derivative, express The first derivative of Indicates the control torque of the drone; superscript symbol Represents the transpose of a vector or matrix; Specifically expressed as: , in, They represent the rolling angular velocity, pitch angular velocity and yaw angular velocity of the UAV respectively; Thus, the UAV dynamics model considering the impact of battery voltage reduction is obtained: , in, Represents the acceleration of the drone in the geographic coordinate system, represents the lift loss vector of the rotor UAV; The third step comprises: The Hopf fibration method is used to generate smooth attitude control signals that can be achieved by UAVs; The required auxiliary variables are known and ,in The definition is as follows: , get And its first-order derivative is: , ; The desired angular velocity signal of the rotorcraft is obtained. for: , in, , express , , Indicates the yaw angle of the drone; In order to obtain the desired attitude angle information, the auxiliary variable , Calculating quaternions , and then obtain the rotation matrix , auxiliary variable expressions and their relationship with quaternions The relationship is as follows: , , ; Quaternions With the rotation matrix The mathematical details of the conversion are as follows: , From the rotation matrix element structure we get: ; Due to the limited range of the actan function, the atan2 function is used instead of atan as the function for solving the inverse tangent value: ; So far, the desired signal of the attitude loop control of the rotor UAV has been obtained; The fourth step comprises: The lift loss model decoupled from the attitude information of the rotorcraft is constructed as follows: , in, Indicates the lift loss of the drone, which is believed to be caused by the reduction of battery voltage and other factors. Specifically indicate the roll angle, pitch angle and yaw angle of the drone; First, the voltage drop observer is designed by using auxiliary variables , the observer gain matrix can be designed , that is, generate is a 3×3 square matrix of diagonal elements, and the observer structure is finally obtained: , in, Indicates the speed of the drone in the geographic coordinate system, , The output of the observer is The observed value of Represents the total lift of the drone in the geographic coordinate system; The fifth step includes: designing the anti-interference control law of the position loop: make represents the position control error, where the subscript Indicates the expected signal, Indicates location information. represents the speed control error, Represents position information, and the position loop anti-interference control law can be designed as follows: , in, represents the expected acceleration in the geographic coordinate system, and represents the controller gain, represents the lift loss vector of the rotorcraft estimated by the voltage drop observer, represents the second-order derivative of the desired trajectory of the rotorcraft, that is, the expected acceleration corresponding to the desired trajectory, Represents the expected lift of the rotary wing UAV in the geographic coordinate system.
2. The online lift compensation method for scene understanding according to claim 1, characterized in that: In the first step, the impact of battery voltage reduction on the drone's attitude control is analyzed: During the flight of the rotorcraft, the controller calculates and processes the expected signal and the real-time status information of the drone to obtain the expected ESC input signal. The ESC adjusts the output voltage or current according to the size of the input signal, thereby realizing the speed control of the motor. However, the quantitative relationship between the ESC input signal and the output signal needs to be calibrated to realize the precise control of the motor speed. This quantitative relationship does not change with the UAV control system signal. Since the onboard lithium battery drives the motor to operate and generates circuit heat loss, its battery power will gradually decrease; as the onboard lithium battery power decreases, its output voltage decreases accordingly, and then the output motor working voltage decreases when the electronic control command signal remains unchanged; for brushless motors, its number of rotations per minute is related to the motor input voltage, and the relevant parameter is the Kv value; the motor parameter Kv value describes the relationship between the peak voltage and speed of the motor under no-load conditions, and the unit is RPM / V, where RPM is the abbreviation of Revolutions Per Minute, which means the number of rotations per minute of the motor, and V means the input voltage of the motor; It can be seen that as the motor input voltage decreases, the motor speed decreases accordingly, and the drone control effect expected by the controller cannot be achieved; one of the most significant effects of voltage reduction is the reduction of the total lift of the drone, which specifically corresponds to the reduction of the lift of each motor, resulting in poor drone position control and attitude control effects, increasing the safety risk of drone flight; The most direct impact of voltage reduction is the rotation speed of each motor of the rotorcraft, which in turn affects the lift generated by each motor. This interference is located in the position control loop of the drone, that is, the lift loss interference of the position loop; the influence of the attitude loop caused by the different lift of each motor is not considered.
