A bionic duty cycle gait control method, program, device and storage medium for an elastic robotic fish

Abstract

The present application belongs to the technical field of underwater bionic robots, and particularly relates to a bionic duty cycle gait control method for an elastic robotic fish, a program, an equipment and a storage medium. The present application is improved based on a traditional Hopf oscillator, a CPG oscillator simulating the red muscle activation mode of fish is constructed, a CPG network model based on the improved Hopf central pattern generator is constructed according to the structure of the elastic robotic fish, and actuator control mapping is completed; a proportional-integral-derivative (PID) CPG closed-loop controller is constructed, and through continuous iteration of the motion cycle, the actual maximum angle of each joint gradually approaches the maximum angle expected value. According to the physiological characteristics of the red muscle activation of fish, the present application constructs a general gait control method for bionic robotic fish, which can effectively reduce the power consumption of the elastic robotic fish and improve the efficiency.

Description

A biomimetic duty cycle gait control method, program, device, and storage medium for an elastic robotic fish. Technical Field

[0001] This invention belongs to the field of underwater biomimetic robot technology, specifically relating to a biomimetic duty cycle gait control method, program, device, and storage medium for elastic robotic fish. Background Technology

[0002] Thanks to their high efficiency and maneuverability, robotic fish have shown great potential in fields such as resource exploration and marine ranching monitoring. Robotic fish can achieve rhythmic movements similar to real fish through gait control technology. Employing efficient gait control strategies is a fundamental condition for robotic fish to complete more advanced tasks. With researchers' in-depth study of the kinematics and fluid dynamics of robotic fish, a series of gait control methods have emerged. A typical example is the trajectory approximation method, which continuously discretizes the midline trajectory of a swimming fish and adjusts motion parameters to make the positions of each joint of the robotic fish approximate the fish's trajectory at corresponding moments. However, the effectiveness of this method is limited by the number of joints in the robotic fish. Another typical control method is the periodic signal generation method. By generating periodic signals with phase differences, the robotic fish's joints can mimic the movement of real fish. This method is simple in principle and easy to implement.

[0003] While the aforementioned gait control methods have achieved some success, none of them reference the muscle activation patterns of real fish. Fish asynchronously activate muscles on both sides of their body axis via neural signals, generating tail-beating movements. The red muscle, which generates the primary driving force, is divided into several segments along the fish's body axis by the diaphragm. The driving patterns of red muscle differ at different locations; the closer to the tail, the shorter the activation time of the red muscle within a cycle, and the muscles on the same side cease activation at the same time. To replicate the movement control mode of fish red muscle in a robotic fish, this invention proposes a biomimetic duty cycle gait control method for elastic robotic fish to improve its swimming performance.

[0004] Similar technical solutions to this invention include the biomimetic robotic fish control methods disclosed in CN115390442A and CN110909859A. CN115390442A discloses a deep reinforcement learning-based biomimetic robotic fish control method, device, and storage medium technology, employing deep reinforcement learning and CPG joint control. CN110909859A discloses a biomimetic robotic fish motion control method and system based on adversarial structured control, combining global control signals and local compensation control signals for model adversarial training. Neither of these methods involves control methods inspired by fish muscle control. They are significantly different from this invention. Summary of the Invention

[0005] The purpose of this invention is to provide a biomimetic duty cycle gait control method, program, device, and storage medium for elastic robotic fish.

[0006] A biomimetic duty cycle gait control method for an elastic robotic fish includes the following steps:

[0007] The Hopf central pattern generator was improved to simulate the red muscle activation mode of fish. The CPG oscillator closer to the front of the elastic robotic fish body is activated earlier in a motion cycle, and all CPG oscillators reach the peak amplitude at the same time.

[0008] Based on the structure of the elastic robotic fish, a CPG network model based on an improved Hopf central mode generator is constructed, with each CPG oscillator corresponding to an actuator of the elastic robotic fish. Each CPG oscillator is numbered, and the phase of the first CPG oscillator is used as the initial phase. The phase constraint is propagated to the remaining CPG oscillators through the coupling relationship between adjacent numbered CPG oscillators, thus completing the actuator control mapping of the elastic robotic fish.

[0009] The oscillator frequency, the maximum expected angle of the actuator corresponding to each CPG oscillator, and the duty cycle are used as optimization objectives to construct an objective vector; a weighted fitness function is constructed with the dual objectives of maximizing motion speed and minimizing transportation cost; the objective vector is optimized in combination with constraints to obtain the optimal objective vector with the largest corresponding weighted fitness function value.

[0010] The elastic robotic fish performs biomimetic duty cycle gait control based on the optimal target vector.

[0011] Furthermore, in the CPG network model based on the improved Hopf central pattern generator, the expression for the CPG oscillator is:

[0012]

[0013] in, For the first State variables of excited neurons in each CPG oscillator; For the first The state variables of the inhibited neurons of each CPG oscillator; For the first The angular frequency of the periodic control signal for each CPG oscillator. ; The frequency of the oscillator; For the first Duty cycle of each CPG oscillator; and The convergence factor; In each motion cycle CPG oscillator activation start time. ; In each motion cycle The activation end time of each CPG oscillator. .

[0014] Furthermore, the frequency of the oscillator The maximum expected angle of the actuator corresponding to each CPG oscillator Duty cycle As the optimization objective, construct the objective vector. ; The total number of CPG oscillators;

[0015] A weighted fitness function is constructed with the dual objectives of maximizing movement speed and minimizing transportation cost. ;

[0016]

[0017] in, and These are the weighting coefficients. ; This is the theoretical maximum speed of the elastic robotic fish; The maximum permissible transportation cost for the flexible robotic fish; and The actual maximum speed and transportation cost of the flexible robotic fish.

[0018] Furthermore, the actual maximum speed of the elastic robotic fish With transportation costs The calculation method is as follows:

[0019] Step 1.1: Initialize the motion cycle Set the maximum exercise cycle and error threshold ;

[0020] Step 1.2: In the first... Movement cycle, collecting real-time movement speed of the elastic robotic fish Energy consumption The actual maximum angle of the actuator corresponding to each CPG oscillator was collected. Calculate the angle error of the actuator corresponding to each CPG oscillator. ;

[0021]

[0022] Step 1.3: If or If the iteration stops, calculate the actual maximum speed of the elastic robotic fish. With transportation costs ;

[0023]

[0024]

[0025] Otherwise, a discrete position PID controller is used to update the amplitude of each CPG oscillator. Each CPG oscillator is based on its duty cycle. Compared with the updated amplitude Output periodic control signals and transmit them to the corresponding actuators to perform control, then return to step 1.2.

[0026] Furthermore, a discrete position PID controller is used to update the amplitude of each CPG oscillator. Specifically:

[0027]

[0028] in, , , For the first The proportional coefficient, integral coefficient, and differential coefficient of each CPG oscillator; The duration of a single motion cycle. .

[0029] Furthermore, the constraints include:

[0030] Based on the structural characteristics of the elastic robotic fish, the upper limit of the oscillator frequency is determined. and lower limit ;

[0031] Based on the maximum swing angle limit of the joints of the flexible robotic fish, determine the upper limit of the maximum angle of the actuator. and lower limit ;

[0032] Set the upper limit of the duty cycle. and lower limit .

[0033] Furthermore, a genetic algorithm is used to optimize the target vector, including the following steps:

[0034] Step 2.1: Generate an initial population through uniformly distributed random sampling, with each individual representing a set of target vectors;

[0035] Step 2.2: Using the tournament selection method, randomly select individuals from the population and enter the next generation population with the individual corresponding to the highest fitness value; repeat this step until the number of individuals in the next generation population reaches the preset size;

[0036] Step 2.3: Crossover operation: Perform arithmetic crossover on the selected parent individuals to generate offspring individuals;

[0037] Step 2.4: Mutation operation: Perform Gaussian mutation on the offspring after crossover to avoid the algorithm getting trapped in local optima;

[0038] After mutation, check whether the parameter meets the constraints. If it exceeds the range, truncate to the boundary value and set the parameter to its boundary value.

[0039] The individual with the highest fitness value in the population after the mutation operation is selected as the optimal individual.

[0040] Step 2.5: If the maximum number of iterations is reached, or the change in the fitness value of the best individual in multiple consecutive iterations is less than the threshold, then stop the iteration and output the best individual, i.e., the optimal target vector; otherwise, return to step 2.2.

[0041] A computer device includes a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above-described biomimetic duty cycle gait control method for an elastic robotic fish.

[0042] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described biomimetic duty cycle gait control method for an elastic robotic fish.

[0043] A computer program product includes computer instructions that, when executed by a processor, implement the steps of the aforementioned biomimetic duty cycle gait control method for an elastic robotic fish.

[0044] The beneficial effects of this invention are as follows:

[0045] This invention improves upon the traditional Hopf oscillator by constructing a CPG oscillator that simulates the red muscle activation mode of fish. Based on the structure of the elastic robotic fish, a CPG network model based on an improved Hopf central pattern generator is built to complete the actuator control mapping. A proportional-integral-derivative (PID)-CPG closed-loop controller is constructed, and through continuous iteration of the motion cycle, the actual maximum angle of each joint gradually approaches the expected maximum angle value. Based on the physiological characteristics of red muscle activation in fish, this invention constructs a general gait control method for biomimetic robotic fish, which can effectively reduce the power consumption and improve the efficiency of the elastic robotic fish, and has broad application prospects in water quality detection, underwater reconnaissance, and other fields. Attached Figure Description

[0046] Figure 1 shows an embodiment of the CPG oscillator allocation for three types of mechanical fish.

[0047] Figure 2 is a schematic diagram of the PID-CPG closed-loop controller.

[0048] Figure 3 shows the convergence of angular error with the motion cycle.

[0049] Figure 4 is a diagram illustrating the principle of red muscle activation in fish.

[0050] Figure 5 shows the amplitude output of the oscillator.

[0051] Figure 6 shows the output of the synthesized amplitude of the oscillator.

[0052] Figure 7 is a diagram of the overall architecture of the present invention. Detailed Implementation

[0053] The present invention will now be further described with reference to the accompanying drawings.

[0054] A biomimetic duty cycle gait control method for an elastic robotic fish includes the following steps:

[0055] Step 1: Improve the Hopf central pattern generator to simulate the red muscle activation mode of fish;

[0056] The activation mechanism of red muscle in fish is shown in Figure 4. The fish body is divided into multiple segments by septa from front to back, and each segment is connected by red muscle. Biomechanics has shown that the red muscle closer to the front of the body is activated earlier in a movement cycle, and the red muscle in each part reaches the peak contractile force at the same time.

[0057] Based on the above biological principles, this invention improves the traditional Hopf oscillator to simulate the red muscle activation mode of fish. The mathematical expression for a single oscillator is:

[0058]

[0059] in, For the first State variables of excited neurons in each CPG oscillator; For the first The state variables of the inhibited neurons of each CPG oscillator; For the first The angular frequency of the periodic control signal for each CPG oscillator. ; , , for , , Regarding time The first derivative; and The convergence factor; In each motion cycle CPG oscillator activation start time. ; In each motion cycle The activation end time of each CPG oscillator. ; For the first Duty cycle of each CPG oscillator;

[0060] Step 2: Based on the structure of the elastic robotic fish, construct a CPG network model based on the Hopf central pattern generator, where each CPG oscillator corresponds to one actuator of the elastic robotic fish;

[0061] Each CPG oscillator is assigned a number, and the phase of the first CPG oscillator is used as the initial phase. The phase constraint is propagated to the remaining CPG oscillators through the coupling relationship between adjacent numbered CPG oscillators, thus completing the actuator control mapping of the elastic robotic fish.

[0062] Step 3: Set the oscillator frequency The maximum expected angle of the actuator corresponding to each CPG oscillator Duty cycle As the optimization objective, construct the objective vector. ;

[0063] A weighted fitness function is constructed with the dual objectives of maximizing movement speed and minimizing transportation cost. ;

[0064]

[0065] in, and These are the weighting coefficients. For example, in water quality testing scenarios where energy efficiency is a priority, energy-saving methods can be adopted. , Underwater reconnaissance scenarios emphasize speed, taking , ;

[0066] This is the theoretical maximum speed of the elastic robotic fish; The maximum permissible transportation cost for the flexible robotic fish;

[0067] and The actual maximum speed and transportation cost of the flexible robotic fish are calculated as follows:

[0068] Step 3.1: Initialize the motion cycle Set the maximum exercise cycle and error threshold ;

[0069] Step 3.2: In the first... Movement cycle, collecting real-time movement speed of the elastic robotic fish Energy consumption The actual maximum angle of the actuator corresponding to each CPG oscillator was collected. Calculate the angle error of the actuator corresponding to each CPG oscillator. ;

[0070]

[0071] Step 3.3: If or If the iteration stops, calculate the actual maximum speed of the elastic robotic fish. With transportation costs ;

[0072]

[0073]

[0074] Otherwise, a discrete position PID controller is used to update the amplitude of each CPG oscillator. Each CPG oscillator is based on its duty cycle. Compared with the updated amplitude Output a periodic control signal, transmit it to the corresponding actuator to perform control, and return to step 3.2;

[0075]

[0076] in, , , For the first The proportional coefficient, integral coefficient, and differential coefficient of each CPG oscillator; The duration of a single motion cycle. , The frequency of the oscillator;

[0077] Step 4: Determine the upper limit of the oscillator frequency based on the structural characteristics of the elastic robotic fish. and lower limit ;

[0078] Based on the maximum swing angle limit of the joints of the flexible robotic fish, determine the upper limit of the maximum angle of the actuator. and lower limit ;

[0079] Set the upper limit of the duty cycle. and lower limit ; acceptable This avoids insufficient driving force due to an excessively small duty cycle or a surge in power consumption due to an excessively large duty cycle.

[0080] For the target vector Perform optimization to obtain the optimal target vector with the largest corresponding weighted fitness function value. ;

[0081] Step 5: Based on the optimal target vector Perform biomimetic duty cycle gait control for the elastic robotic fish.

[0082] In step 4, a genetic algorithm can be used to analyze the target vector. The optimization process includes the following steps:

[0083] Step 4.1: Generate a scale of [size missing] through uniformly distributed random sampling. The initial population, where each individual represents a set of target vectors. , ;

[0084]

[0085]

[0086]

[0087]

[0088] Step 4.2: Using the tournament selection method, randomly select m individuals from the population and choose their fitness values. The highest-ranking individual enters the next generation of the population, and the cycle repeats. Secondly, ensure the preservation of superior genes in the population;

[0089] Step 4.3: Crossover operation: Perform arithmetic crossover on the selected parent individuals to generate offspring individuals;

[0090] For the parent generation individual and

[0091] offspring individuals The formula for calculation is:

[0092]

[0093]

[0094]

[0095] in, This is a crossover factor used to ensure a smooth transition of child parameters within the range of parent parameters. ;

[0096] Step 4.4: Mutation operation: Perform Gaussian mutation on the offspring after crossover to avoid the algorithm getting trapped in local optima;

[0097] For the first generation of individuals Parameters Parameters after mutation

[0098] in, This indicates that the mean is 0 and the variance is 0. ;

[0099] After mutation, check whether the parameter meets the constraints. If it exceeds the range, truncate to the boundary value and set the parameter to its boundary value.

[0100] Step 4.5: Select the individual with the highest fitness value in the population after the mutation operation as the optimal individual;

[0101] If the maximum number of iterations is reached, or the change in the fitness value of the optimal individual in multiple consecutive iterations is less than a threshold (i.e., convergence is achieved), then the iteration stops, and the optimal individual, i.e., the optimal target vector, is output. .

[0102] Example 1:

[0103] The following uses three typical dual-joint robotic fish structures to illustrate this point. The number of joints is just for illustration; in fact, it can be extended to robotic fish with more joints.

[0104] As shown in Figure 1, for the fluid-driven flexible robotic fish 1, the oscillator CPG1 corresponds to the pressure of the fluid cavity 1.1 on the left side of the first joint, the oscillator CPG2 corresponds to the pressure of the fluid cavity 1.2 on the right side of the first joint, the oscillator CPG3 corresponds to the pressure of the fluid cavity 1.3 on the left side of the first joint, and the oscillator CPG4 corresponds to the pressure of the fluid cavity 1.4 on the right side of the first joint. The elasticity of the robotic fish body is provided by the flexible actuator itself.

[0105] For the cable-operated robotic fish 2, the oscillator CPG1 corresponds to the tension on the left cable 2.1 of the first joint, the oscillator CPG2 corresponds to the tension on the right cable 2.2 of the first joint, the oscillator CPG3 corresponds to the tension on the left cable 2.3 of the first joint, and the oscillator CPG4 corresponds to the tension on the right cable 2.4 of the first joint. The elasticity of the robotic fish body is provided by the elastic element 2.5 located on the centerline of the fish body.

[0106] For the servo-driven robotic fish 3, there is no symmetrical drive mechanism. Therefore, the torque generated by the servos is decomposed into the sum of clockwise and counterclockwise virtual torques. Oscillator CPG1 corresponds to the counterclockwise virtual torque on the first servo 3.1, oscillator CPG2 corresponds to the clockwise virtual torque on the first servo 3.1, oscillator CPG3 corresponds to the counterclockwise virtual torque on the second servo 3.2, and oscillator CPG4 corresponds to the counterclockwise virtual torque on the second servo 3.2. The elasticity of the robotic fish's body is provided by the torsion spring 3.3 on the first servo 3.1 and the torsion spring 3.4 on the second servo 3.2. The phase difference between oscillators CPG1 and CPG2 is Φ12, and the phase difference between oscillators CPG3 and CPG4 is Φ34.

[0107] This invention designs an oscillator mimicking the activation mechanism of red muscle in fish, as shown in Figure 4. The fish body is divided into multiple segments by septa from front to back, each connected by red muscle. Biomechanical evidence shows that red muscle closer to the front of the body activates earlier within a single movement cycle, and all parts of the red muscle reach their peak contractile force simultaneously. Based on these biological principles, this invention improves the traditional Hopf model to simulate the red muscle activation mechanism in fish, constructing a CPG network model based on an improved Hopf central pattern generator. The mathematical expression for a single CPG oscillator is:

[0108]

[0109] and Representing the The state variables of excitatory and inhibitory neurons of each CPG oscillator; It is the first The time of one CPG oscillator within one cycle , The frequency of the oscillator. For the first The phase of a CPG oscillator;

[0110] , , Oscillators CPG1 and CPG2, and oscillators CPG3 and CPG4, are pairs of oscillators that drive the same joint. .

[0111] For the fluid-driven flexible robotic fish 1 and the wire-pulled robotic fish 2, the oscillator amplitude directly controls the single-sided actuator. For the series-driven servo motor robotic fish 3, the torque generated by the first servo motor 3.1 is CPG1 + CPG2, and the torque generated by the second servo motor 3.2 is CPG3 + CPG4. Therefore, the present invention can adjust the movement of the robotic fish by adjusting the duty cycle of the oscillator. For example, when... , , At that time, the output of the amplitude of each oscillator is shown in Figure 5. After the amplitude of a pair of oscillators in the same joint is combined, the combined amplitude of the oscillators is shown in Figure 6. The stage when the combined amplitude of the oscillators is not zero is the stage when the actuator is in action, and the stage when the combined amplitude of the oscillators is zero is the stage when the actuator is not in action. At this time, the fish body returns to the centerline by relying on passive stiffness and surrounding hydrodynamic movement, thereby achieving the purpose of reducing power consumption and improving efficiency.

[0112] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A biomimetic duty cycle gait control method for an elastic robotic fish, characterized in that: An improved Hopf central pattern generator was used to simulate the red muscle activation mechanism in fish. CPG oscillators closer to the front of the elastic robotic fish's body activated earlier within a single motion cycle, and all CPG oscillators reached their peak amplitude at the same time. Based on the structure of the elastic robotic fish, a CPG network model based on the improved Hopf central pattern generator was constructed, with each CPG oscillator corresponding to one actuator of the elastic robotic fish. Each CPG oscillator was numbered, and the phase of the first CPG oscillator was used as the initial phase. The phase constraint was propagated to the remaining CPG oscillators through the coupling relationship between adjacent numbered CPG oscillators. The CPG oscillator is used to complete the actuator control mapping of the elastic robotic fish. The oscillator frequency, the maximum expected angle of the actuator corresponding to each CPG oscillator, and the duty cycle are used as optimization objectives to construct a target vector. A weighted fitness function is constructed with the dual objectives of maximizing motion speed and minimizing transportation cost. The target vector is optimized under constraints to obtain the optimal target vector with the largest corresponding weighted fitness function value. Based on the optimal target vector, the biomimetic duty cycle gait control of the elastic robotic fish is executed. The expression for the CPG oscillator in the CPG network model based on the improved Hopf central pattern generator is: in, For the first State variables of excited neurons in a CPG oscillator; For the first The state variables of the inhibited neurons of each CPG oscillator; For the first The angular frequency of the periodic control signal for each CPG oscillator. ； The frequency of the oscillator; For the first Duty cycle of each CPG oscillator; and The convergence factor; In each motion cycle CPG oscillator activation start time. ； In each motion cycle The activation end time of each CPG oscillator. ； For the first The amplitude of each CPG oscillator.

2. The biomimetic duty cycle gait control method for an elastic robotic fish according to claim 1, characterized in that: The frequency of the oscillator The maximum expected angle of the actuator corresponding to each CPG oscillator Duty cycle As the optimization objective, construct the objective vector. ； Let the total number of CPG oscillators be denoted as ; a weighted fitness function is constructed with the dual objectives of maximizing motion speed and minimizing transportation cost as the primary objectives. ； in, and These are the weighting coefficients. ； This is the theoretical maximum speed of the elastic robotic fish; The maximum permissible transportation cost for the flexible robotic fish; and The actual maximum speed and transportation cost of the flexible robotic fish.

3. The biomimetic duty cycle gait control method for an elastic robotic fish according to claim 2, characterized in that: The actual maximum speed of the elastic robotic fish With transportation costs The calculation method is as follows: Step 1.1: Initialize the motion cycle Set the maximum exercise cycle and error threshold ； Step 1.2: In the first... Movement cycle, collecting real-time movement speed of the elastic robotic fish Energy consumption The actual maximum angle of the actuator corresponding to each CPG oscillator was collected. Calculate the angle error of the actuator corresponding to each CPG oscillator. ； Step 1.3: If or If the iteration stops, calculate the actual maximum speed of the elastic robotic fish. With transportation costs ； in, The duration of a single motion cycle. ； Otherwise, a discrete position PID controller is used to update the amplitude of each CPG oscillator. Each CPG oscillator is based on its duty cycle. Compared with the updated amplitude Output periodic control signals and transmit them to the corresponding actuators to perform control, then return to step 1.

2.

4. The biomimetic duty cycle gait control method for an elastic robotic fish according to claim 3, characterized in that: The discrete position PID controller is used to update the amplitude of each CPG oscillator. Specifically: in, 、 、 For the first The proportional coefficient, integral coefficient, and differential coefficient of each CPG oscillator.

5. The biomimetic duty cycle gait control method for an elastic robotic fish according to claim 2, characterized in that: The constraints include: determining the upper limit of the oscillator frequency based on the structural characteristics of the elastic robotic fish. and lower limit Based on the maximum swing angle limit of the joints of the elastic robotic fish, determine the upper limit of the maximum angle of the actuator. and lower limit Set the upper limit of the duty cycle. and lower limit 。 6. The biomimetic duty cycle gait control method for an elastic robotic fish according to claim 5, characterized in that: The genetic algorithm is used to optimize the target vector, including the following steps: Step 2.1: Generate an initial population through uniformly distributed random sampling, with each individual representing a target vector; Step 2.2: Use tournament selection to randomly select individuals from the population, and the individual with the highest corresponding fitness value is added to the next generation population; repeat this step until the number of individuals in the next generation population reaches the preset size; Step 2.3: Crossover operation: Perform arithmetic crossover on the selected parent individuals to generate offspring individuals; Step 2.4: Mutation operation: Perform Gaussian mutation on the crossover offspring individuals to avoid the algorithm getting trapped in local optima; after mutation, check whether the parameters meet the constraints. If they exceed the range, truncate to the boundary value and set the parameters equal to the boundary value; select the individual with the highest corresponding fitness value in the population after the mutation operation as the optimal individual; Step 2.5: If the maximum number of iterations is reached, or the change in the fitness value of the optimal individual in multiple consecutive iterations is less than the threshold, stop the iteration and output the optimal individual, i.e., the optimal target vector; otherwise, return to step 2.

2.

7. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When executed by a processor, the computer program implements the steps of the method according to any one of claims 1 to 6.

9. A computer program product comprising computer instructions, characterized in that: When executed by a processor, the computer instructions implement the steps of the method according to any one of claims 1 to 6.

Images