Bionic snakelike robot based on pneumatic soft body structure and control method

Through bionic snake robots based on pneumatic soft structure and genetic algorithm-optimized PID control, the problem that traditional hardware robots cannot pass through complex pipelines is solved, efficient power plant pipeline detection is achieved, and maintenance costs are reduced.

CN118952181AActive Publication Date: 2024-11-15SHAANXI JUSHOUXIANG TECHNOLOGY CO LTD

Patent Information

Application Number
CN202411145376.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-20
Publication Date
2024-11-15
Estimated Expiration
2044-08-20

AI Technical Summary

Technical Problem

Traditional hard robots cannot pass through various curved or irregular pipes in the power plant, resulting in low detection efficiency and high maintenance costs.

Method used

A bionic snake-shaped robot based on pneumatic soft structure is designed to realize snake-shaped motion through multi-section body mechanisms and flexible air bags, and optimize PID control with air pressure control system and genetic algorithms, which can pass through complex pipelines.

Benefits of technology

It significantly improves detection efficiency, reduces the daily maintenance costs of the power plant, and allows the robot to flexibly pass through various curved or irregular pipes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118952181A_ABST
    Figure CN118952181A_ABST
Patent Text Reader

Abstract

The invention provides a bionic snakelike robot based on a pneumatic soft body structure and a control method, and the bionic snakelike robot comprises a head structure, a body mechanism, a trunk mechanism, a camera, a hollow cavity, a bending sensor, a pressure sensor, a flexible air bag, an air source supply system and an air pressure control system, the trunk mechanism communicates with the air source supply system through a pipeline, the air source supply system is electrically connected with the air pressure control system, and therefore motion control over the bionic snake-shaped robot is achieved. The trunk mechanism of the device is formed by connecting the multiple sections of body mechanisms end to end, the crawling state of a snake is simulated by the trunk mechanism through expansion or contraction of the flexible air bags arranged on the two sides of the body mechanisms, and therefore the robot can pass through various bent or irregular pipelines, the detection efficiency is greatly improved, and the detection cost is reduced. And the daily maintenance cost of a power plant is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of bionic robots, and in particular relates to a bionic snake-like robot based on a pneumatic soft structure and a control method thereof. Background Art

[0002] In order to ensure the safety of normal operation of power plants, it is necessary to regularly inspect the pipelines in power plants to identify wear, corrosion and other potential problems in the pipelines. However, in pipeline inspection in power plants, traditional hard robots are often limited by their rigid structure when operating inside narrow or complex pipelines, and are unable to pass through various curved or irregular pipelines, which seriously restricts the inspection efficiency and increases the maintenance cost of power plants.

[0003] Therefore, there is an urgent need for a bionic robot and a control method that can pass through various curved or irregular pipes. Summary of the invention

[0004] In order to solve the shortcomings and deficiencies of the prior art, a bionic snake-like robot based on a pneumatic soft structure and a control method are provided, thereby solving the problem that the existing hard robot cannot pass through various curved or irregular pipes due to its rigid structure.

[0005] In order to achieve the purpose of the present invention, a bionic snake-like robot based on a pneumatic soft structure is provided, which includes a head structure and a trunk mechanism. The rear end of the head mechanism is fixed to the front end of the trunk mechanism. The trunk mechanism is connected to an air supply system through a pipeline. The air supply system is electrically connected to an air pressure control system, thereby realizing the motion control of the bionic snake-like robot.

[0006] As a further improvement of the above scheme, the torso mechanism is formed by multiple body mechanisms connected end to end, and the body mechanism includes a hexagonal body structure. The front end of the hexagonal body structure is provided with a clamping end of a sliding connection structure, and the rear end is provided with a fixed end of the sliding connection structure. The fixed end of the sliding connection structure in the front body mechanism cooperates with the clamping end of the sliding connection structure in the rear body mechanism to realize the connection between the two body mechanisms. Flexible air bags are installed on both sides of the hexagonal body structure, and the flexible air bags are connected to the air supply system through pipelines, so as to realize the movement of the bionic snake-like robot.

[0007] As a further improvement of the above solution, a plurality of snap-in grooves are provided at the bottom of the hexagonal structure of the body, and scale gripping structures are inserted into the snap-in grooves to increase the friction between the bionic snake-like robot and the ground and thus improve the smoothness of movement.

[0008] As a further improvement of the above scheme, the flexible air bag includes a plurality of folded air cavity structures, and the adjacent surfaces of two air cavity structures are sealed and connected by plastic sealing. An air inlet pipe is provided on one side of the air cavity structure, and the air inlet pipes of two adjacent air cavity structures are connected to the cavity through an air guide pipe. A vent is provided on the outer surface of the air cavity structure at one end of the flexible air bag, and the vent is connected to the air source supply system through a pipeline, thereby realizing the expansion or contraction of the flexible air bag.

[0009] As a further improvement of the above scheme, the air supply system includes an air compressor and a pressure reducing valve. The air compressor is electrically connected to a power supply, and the air compressor is connected to the pressure reducing valve. The pressure reducing valve is connected to the vents in multiple flexible air bags through multiple air supply pipelines, and electrical proportional valves are provided on the multiple air supply pipelines.

[0010] As a further improvement of the above scheme, the air pressure control system includes a 24V constant voltage power supply, a host computer, a PLC control cabinet, a data acquisition card, a pressure sensor, a bending sensor and a camera. The 24V constant voltage power supply and the camera are electrically connected to the power supply. The 24V constant voltage power supply is electrically connected to the electrical proportional valve and the PLC control cabinet for stable power supply. The PLC control cabinet is electrically connected to the electrical proportional valve for controlling the electrical proportional valve. The host computer is electrically connected to the PLC control cabinet for transmitting action instructions. The bending sensor is installed on the front side of the head structure to collect the bending angle signal of the bionic snake-like robot in real time. The camera is symmetrically arranged above the head structure. The pressure sensor is installed at the tail of the trunk mechanism to detect the gas pressure value output by the electrical proportional valve in real time. The camera, bending sensor and pressure sensor are all electrically connected to the data acquisition card, and the data acquisition card is electrically connected to the host computer, so as to transmit the collected information to the host computer for analysis.

[0011] As a further improvement of the above scheme, the bending sensor includes a circuit board and a strain gauge, the circuit board and the strain gauge are connected by a connecting line, the circuit board is electrically connected to a data acquisition card, and the strain gauge is pasted on the side of the front end of the trunk mechanism.

[0012] As a further improvement of the above solution, a hollow cavity is opened in the center of the hexagonal structure of the body to accommodate pipelines and wires so as to improve the mobility of the bionic snake-like robot.

[0013] A control method of a bionic snake-like robot, characterized in that: the body mechanism at the front end of the trunk mechanism is divided into a C area and a D area, the remaining body mechanisms in the trunk mechanism are divided into an A area and a B area, the A area is divided into an a1 half segment and an a2 half segment, and the B area is divided into a b1 half segment and a b2 half segment, and the specific control process is as follows:

[0014] S1, forward

[0015] S1.1, when the A area of ​​the body structure in the trunk structure is inflated, the a1 half of the A area of ​​the body structure moves forward, and the a2 half has a tendency to move backward. Due to the friction between the scale gripping structure and the tube wall, the A area of ​​the body structure moves forward;

[0016] S1.2, when the B area of ​​the body structure in the trunk structure is inflated, the b1 half of the B area of ​​the body structure moves forward, and the b2 half has a tendency to move backward. At the same time, when the A area of ​​the body structure in the trunk structure is deflated, the a1 half of the A area of ​​the body structure has a tendency to move backward, and the a2 half moves forward. Due to the friction between the scale gripping structure and the tube wall, the B area of ​​the body structure moves forward.

[0017] S1.3, when air is inflated into area A of the body structure in the trunk structure, the a1 half of area A of the body structure moves forward, and the a2 half has a tendency to move backward. At the same time, when air is deflated into area B of the body structure in the trunk structure, the b1 half of area B of the body structure has a tendency to move backward, and the b2 half moves forward. Due to the friction between the scale gripping structure and the tube wall, area A of the body structure moves forward.

[0018] S1.4, continuously executing the above S1.2 and S1.3, thus achieving continuous forward movement;

[0019] S2, Turn

[0020] S2.1, the pipeline centerline model is established through the host computer and the set value is input. The camera transmits the image inside the pipeline to the host computer, and the host computer monitors in real time;

[0021] S2.2, left turn: When the host computer detects that the actual pipeline centerline is biased to the left and the deviation value from the pipeline centerline model is far from the set value, the host computer sends a signal to the electric proportional valve through the PLC control cabinet, thereby inflating the D area of ​​the front end of the trunk mechanism, so that the head structure bends to the left. The camera transmits the image in the pipeline in real time and the bending sensor transmits the bending angle signal as a process value to the host computer in real time. The host computer uses the Python program of genetic algorithm to optimize PID for calculation, and passes the three parameters KP, Ki, and Kd generated by the calculation to the PID controller in the host computer. The PID controller calculates the control signal and sends it to the PLC control cabinet through the TCP connection. The PLC control cabinet then sends an adjustment signal to the electric proportional valve, thereby adjusting the inflation valve in the electric proportional valve. The opening size eventually makes the head structure gradually complete the left turn action; when the host computer detects that the deviation value between the actual pipeline centerline and the pipeline centerline model is close to the set value, the camera transmits the image in the pipeline in real time and the bending sensor transmits the bending angle signal as a process value to the host computer in real time. The host computer uses the Python program of genetic algorithm to optimize PID for calculation, and passes the three parameters KP, Ki, and Kd generated by the calculation to the PID controller in the host computer. After the PID controller calculates the control signal, it sends it to the PLC control cabinet through TCP connection. The PLC control cabinet controls the opening size of the inflation valve in the electric proportional valve, so as to deflate the D area of ​​the front end of the trunk mechanism; when the bending angle signal transmitted by the bending sensor to the host computer is 0, the bionic snake-like robot resumes the forward state;

[0022] S2.3, right turn: When the host computer detects that the actual pipeline centerline is biased to the right and the deviation value from the pipeline centerline model is far away from the set value, the host computer sends a signal to the electric proportional valve through the PLC control cabinet, thereby inflating the C area of ​​the front end of the trunk mechanism, so that the head structure bends to the right. The camera transmits the image in the pipeline in real time and the bending sensor transmits the bending angle signal as a process value to the host computer in real time. The host computer uses the Python program of genetic algorithm to optimize PID for calculation, and passes the three parameters KP, Ki, and Kd generated by the calculation to the PID controller in the host computer. The PID controller calculates the control signal and sends it to the PLC control cabinet through TCP connection. The PLC control cabinet controls the opening size of the inflation valve in the electric proportional valve, and finally makes The head structure gradually completes the right turn action; when the host computer detects that the deviation value between the actual pipeline centerline and the pipeline centerline model is close to the set value, the camera transmits the image in the pipeline in real time and the bending sensor transmits the bending angle signal as a process value to the host computer in real time. The host computer uses the genetic algorithm to optimize the PID Python program for calculation, and passes the three parameters KP, Ki, and Kd generated by the calculation to the PID controller in the host computer. After the PID controller calculates the control signal, it sends it to the PLC control cabinet through the TCP connection. The PLC control cabinet controls the opening size of the inflation valve in the electric proportional valve, thereby deflating the C area of ​​the front end of the trunk mechanism. When the bending angle signal transmitted by the bending sensor to the host computer is 0, the bionic snake-like robot resumes the forward state.

[0023] As a further improvement of the above scheme, the genetic algorithm is optimized through genetic operations such as chromosome encoding, calculation of individual fitness function values, selection, crossover and mutation; each individual represents a solution, and by calculating the fitness function value and performing selection, crossover and mutation operations, the population evolves from generation to generation and gradually approaches the optimal solution. After many generations of survival of the fittest, the population continues to multiply and evolve, and finally finds the best solution to the problem.

[0024] The beneficial effects of the present invention are:

[0025] Compared with the prior art, the present invention provides a bionic snake-like robot and control method based on a pneumatic soft structure. The trunk mechanism is formed by connecting multiple body mechanisms end to end, and the expansion or contraction of flexible air bags arranged on both sides of the body mechanism is used to realize the trunk mechanism simulating the crawling state of a snake, so that the robot can pass through various curved or irregular pipelines, greatly improving the detection efficiency and reducing the daily maintenance cost of the power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a structural schematic diagram of the present invention;

[0027] Figure 2It is a schematic diagram of the structure of the body mechanism in the present invention;

[0028] Figure 3 It is a structural schematic diagram of the body structure of the present invention without the scale gripping structure;

[0029] Figure 4 It is a schematic diagram of the connection between the body structure and the flexible air bag in the present invention;

[0030] Figure 5 is a schematic diagram of a flexible air bag in the present invention;

[0031] Figure 6 is a connection diagram of the bending sensor in the present invention;

[0032] Figure 7 is a schematic diagram of the structure of the bending sensor in the present invention;

[0033] Figure 8 It is a connection diagram of the pressure sensor in the present invention;

[0034] Fig. 9 It is a schematic diagram of the structure of the pressure sensor in the present invention;

[0035] Fig.10 It is a connection diagram of the air source supply system, the air pressure control system and the trunk mechanism in the present invention;

[0036] Fig.11 It is a schematic diagram of the division of the trunk mechanism area in the present invention;

[0037] Fig.12 It is a schematic diagram of the trunk mechanism of the present invention when it moves forward;

[0038] Fig.13 It is a schematic diagram of the trunk mechanism of the present invention when turning left / right.

[0039] Among them, 1-head structure; 2-trunk structure; 3-body structure; 4-camera; 5-hollow cavity; 6-bending sensor; 7-pressure sensor; 8-flexible air bag; 9-air supply system; 10-air pressure control system; 301-sliding connection structure; 302-clip slot; 303-body hexagonal structure; 304-scale grip structure; 601-circuit board; 602-connecting wire; 603-strain gauge; 801-air cavity structure; 802-intake pipe; 803-air guide pipe; 804-vent; 901-air compressor; 902-pressure reducing valve; 903-electric proportional valve; 1000-24V constant voltage power supply; 1001-host computer; 1002-PLC control cabinet; 1003-data acquisition card. DETAILED DESCRIPTION

[0040] The specific embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings:

[0041] like Figure 1 As shown, the present invention provides a bionic snake-like robot based on a pneumatic soft structure, comprising a head structure 1 and a trunk mechanism 2, the rear end of the head mechanism 1 is fixed to the front end of the trunk mechanism 2, the trunk mechanism 2 is connected to the air supply system 9 through a pipeline, and the air supply system 9 is electrically connected to the air pressure control system 10, so as to realize the motion control of the bionic snake-like robot. The trunk mechanism 2 is composed of multiple body mechanisms 3 connected end to end, such as Figure 2-4 As shown, the body mechanism 3 includes a hexagonal body structure 303, a front end of the hexagonal body structure 303 is provided with a snap-in end of a sliding connection structure 301, and a rear end is provided with a fixed end of the sliding connection structure 301, the fixed end of the sliding connection structure 301 in the front body mechanism 3 cooperates with the snap-in end of the sliding connection structure 301 in the rear body mechanism 3 to realize the connection between the two body mechanisms 3, flexible air bags 8 are installed on both sides of the hexagonal body structure 303, and the flexible air bags 8 are connected to the air supply system 9 through pipelines to realize the movement of the bionic snake-like robot, a plurality of snap-in grooves 302 are provided at the bottom of the hexagonal body structure 303, and a scale gripping structure 304 is inserted in the snap-in groove 302 to increase the friction between the bionic snake-like robot and the ground and thereby improve the smoothness of the movement, and a hollow cavity 5 is provided in the center of the hexagonal body structure 303 to accommodate pipelines and wires and thereby improve the mobility of the bionic snake-like robot.

[0042] like Figure 5 As shown, the flexible air bag 8 includes a plurality of folded air cavity structures 801, and the adjacent surfaces of two air cavity structures 801 are sealed and connected by plastic sealing. An air inlet pipe 802 is provided on one side of the air cavity structure 801, and the air inlet pipes 802 of two adjacent air cavity structures 801 are connected to the cavity through an air guide pipe 803. The outer surface of the air cavity structure 801 at one end of the flexible air bag 8 is provided with an air vent 804, and the air vent 804 is connected to the air source supply system 9 through a pipeline, thereby realizing the expansion or contraction of the flexible air bag 8.

[0043] like Figure 6-Figure 10As shown, the air source supply system 9 includes an air compressor 901 and a pressure reducing valve 902, the air compressor 901 is electrically connected to the power supply, the air compressor 901 is connected to the pressure reducing valve 802, the pressure reducing valve 902 is connected to the vents in the multiple flexible air bags 8 through a multi-way air supply pipeline, and an electrical proportional valve 903 is arranged on the multi-way air supply pipeline. The air pressure control system 10 includes a 24V constant voltage power supply 1000, a host computer 1001, a PLC control cabinet 1002, a data acquisition card 1003, a pressure sensor 7, a bending sensor 6 and a camera 4, the 24V constant voltage power supply 1000 and the camera 4 are both electrically connected to the power supply, and the 24V constant voltage power supply 1000 is electrically connected to the electrical proportional valve 903 and the PLC control cabinet 1002 for stable power supply. The PLC control cabinet 1002 is electrically connected to the electrical proportional valve 903 for controlling the electrical proportional valve 903. The host computer 1001 is electrically connected to the PLC control cabinet 1002 for transmitting action instructions. The bending sensor 6 is installed on the front side of the head structure 1 to collect the bending angle signal of the bionic snake-like robot in real time. The camera 4 is symmetrically arranged above the head structure 1. The pressure sensor 7 is installed at the tail of the trunk mechanism 2 to detect the gas pressure value output by the electrical proportional valve 903 in real time. The camera 4, the bending sensor 6, and the pressure sensor 7 are all electrically connected to the data acquisition card 1003. The data acquisition card 1003 is electrically connected to the host computer 1001, so as to transmit the collected information to the host computer 1001 for analysis. The bending sensor 6 includes a circuit board 601 and a strain gauge 603 . The circuit board 601 and the strain gauge 603 are connected via a connecting line 602 . The circuit board 601 is electrically connected to the data acquisition card 1003 . The strain gauge 603 is pasted on the side of the front end of the trunk mechanism 3 .

[0044] The present invention also provides a control method for a bionic snake-like robot, such as Fig.11 As shown, the body structure 3 at the front end of the trunk structure 2 is divided into a C area and a D area, and the remaining body structure 3 in the trunk structure 2 is divided into an A area and a B area. The A area is divided into an a1 half section and an a2 half section, and the B area is divided into a b1 half section and a b2 half section. Figure 12-13 As shown, the specific control process is as follows:

[0045] S1, forward

[0046] S1.1, inflate area A of body structure 3 in trunk structure 2, so that half a1 of area A of body structure 3 moves forward, and half a2 tends to move backward. Due to the friction between scale gripping structure 304 and tube wall, area A of body structure 3 moves forward.

[0047] S1.2, inflate the B region of the body structure 3 in the trunk structure 2, then the b1 half of the B region of the body structure 3 moves forward, and the b2 half has a tendency to move backward. At the same time, deflate the A region of the body structure 3 in the trunk structure 2, then the a1 half of the A region of the body structure 3 has a tendency to move backward, and the a2 half moves forward. Due to the friction between the scale gripping structure 304 and the tube wall, the B region of the body structure 3 moves forward.

[0048] S1.3, when air is inflated into area A of the body structure 3 in the trunk structure 2, the a1 half of area A of the body structure 3 moves forward, and the a2 half has a tendency to move backward. At the same time, when air is deflated into area B of the body structure 3 in the trunk structure 2, the b1 half of area B of the body structure 3 has a tendency to move backward, and the b2 half moves forward. Due to the friction between the scale gripping structure 304 and the tube wall, the A area of ​​the body structure 3 moves forward.

[0049] S1.4, continuously executing the above S1.2 and S1.3, thus achieving continuous forward movement;

[0050] S2, Turn

[0051] S2.1, the pipeline centerline model is established through the host computer 1001 and the set value is input, the camera 4 transmits the image in the pipeline to the host computer 1001, and the host computer 1001 monitors in real time;

[0052] S2.2, left turn: when the host computer 1001 detects that the actual pipeline centerline is biased to the left and the deviation value from the pipeline centerline model is far from the set value, the host computer 1001 sends a signal to the electrical proportional valve 903 through the PLC control cabinet 1002, thereby inflating the D area of ​​the front end body structure 3 of the trunk structure 2, so that the head structure 1 bends to the left, the camera 4 transmits the image in the pipeline in real time, and the bending sensor 6 transmits the bending angle signal as a process value to the host computer 1001 in real time, the host computer 1001 uses the genetic algorithm to optimize the PID Python program for calculation, and passes the three parameters KP, Ki, and Kd generated by the calculation to the PID controller in the host computer 1001. After the PID controller calculates the control signal, it sends it to the PLC control cabinet 1002 through the TCP connection. The PLC control cabinet 1002 then sends an adjustment signal to the electrical proportional valve 903, thereby adjusting the electric proportional valve 903 to fill the gap. The opening size of the air valve eventually makes the head structure 1 gradually complete the left turn action; when the host computer 1001 detects that the deviation value between the actual pipeline centerline and the pipeline centerline model is close to the set value, the camera 4 transmits the image in the pipeline in real time and the bending sensor 6 transmits the bending angle signal as a process value to the host computer 1001 in real time, and the host computer 1001 uses the Python program of the genetic algorithm to optimize the PID to perform calculations, and passes the three parameters KP, Ki, and Kd generated by the calculation to the PID controller in the host computer 1001. After the PID controller calculates the control signal, it sends it to the PLC control cabinet 1002 through the TCP connection. The PLC control cabinet 1002 controls the opening size of the inflation valve in the electric proportional valve 903, so that the D area of ​​the body mechanism 3 at the front end of the trunk mechanism 2 is deflated; when the bending angle signal transmitted by the bending sensor 6 to the host computer 1001 is 0, the bionic snake-like robot resumes the forward state;

[0053] S2.3, right turn: when the host computer 1001 detects that the actual pipeline centerline is biased to the right and the deviation value from the pipeline centerline model is far away from the set value, the host computer 1001 sends a signal to the electrical proportional valve 903 through the PLC control cabinet 1002, thereby inflating the C area of ​​the front end body structure 3 of the trunk structure 2, so that the head structure 1 bends to the right, the camera 4 transmits the image in the pipeline in real time and the bending sensor 6 transmits the bending angle signal as a process value to the host computer 1001 in real time, the host computer 1001 uses the genetic algorithm to optimize the PID Python program for calculation, and passes the three parameters KP, Ki, and Kd generated by the calculation to the PID controller in the host computer 1001. After the PID controller calculates the control signal, it sends it to the PLC control cabinet 1002 through the TCP connection. The PLC control cabinet 1002 controls the opening size of the inflation valve in the electrical proportional valve 903. Finally, the head structure 1 gradually completes the right turn action; when the host computer 1001 detects that the deviation value between the actual pipeline centerline and the pipeline centerline model is close to the set value, the camera 4 transmits the image in the pipeline in real time and the bending sensor 6 transmits the bending angle signal as a process value to the host computer 1001 in real time, and the host computer 1001 uses the Python program of genetic algorithm to optimize PID for calculation, and passes the three parameters KP, Ki, and Kd generated by the calculation to the PID controller in the host computer 1001. After the PID controller calculates the control signal, it sends it to the PLC control cabinet 1002 through the TCP connection. The PLC control cabinet 1002 controls the opening size of the inflation valve in the electric proportional valve 903, so that the C area of ​​the front end body mechanism 3 of the trunk mechanism 2 is deflated; when the bending angle signal transmitted by the bending sensor 6 to the host computer 1001 is 0, the bionic snake-like robot resumes the forward state.

[0054] The genetic algorithm is optimized through chromosome encoding, calculation of individual fitness function values, selection, crossover and mutation. Each individual represents a solution. By calculating the fitness function value and performing selection, crossover and mutation operations, the population evolves from generation to generation and gradually approaches the optimal solution. After many generations of survival of the fittest, the population continues to multiply and evolve, and finally finds the best solution to the problem. Among them:

[0055] Initial population: The size of the initial population has a certain impact on the performance and running speed of the algorithm. A small number will lead to a lack of diversity in the population, and too many numbers will slow down the operation, converge prematurely, and overfit. Therefore, the appropriate population size has a positive effect on the algorithm. In this paper, the initial population size is set to 80, that is, the three parameters KP, Ki, and Kd of PID in each generation are 80 each. In order to effectively search for the optimal solution, the ranges of the three parameters are set to: 0≤KP≤0.1, 0≤Ki≤0.03, 0≤Kd≤0.06;

[0056] In order to further improve the diversity and coverage of the population, we introduce the chaotic mapping method to generate the initial population. Specifically, the Logistic mapping is used to generate chaotic sequences and these sequences are mapped to the range of PID parameters. The Logistic mapping is a common chaotic mapping, which is defined as follows:

[0057] x n+1 = r·x n ·(1-x n )

[0058] Among them, r is the chaos parameter, which usually takes a value between [3.57,4];

[0059] After the chaotic sequence is generated, its value is linearly mapped to the range of the PID parameter. Assuming the value of the chaotic sequence is X, and the range of the PID parameter is [a, b], the value after mapping is:

[0060] Y=a+(ba)·X

[0061] This method can effectively improve the optimization effect and efficiency of the genetic algorithm. By reasonably setting the initial population size and introducing chaotic mapping, the algorithm's search ability can be improved while ensuring population diversity, thereby finding a better PID parameter combination.

[0062] Chromosome encoding: In genetic algorithms, the choice of encoding has a direct impact on the performance of the algorithm. Common encoding methods include binary encoding, permutation encoding, and real number encoding. Real number encoding has good convergence and can set parameter precision arbitrarily. Therefore, this paper uses real number encoding for chromosome encoding, and the encoding precision is set to 5 decimal places.

[0063] Fitness function: The design of the fitness function is crucial to the effect of the genetic algorithm. The fitness value can reflect the quality of each individual's solution to the problem and determine the probability of the individual being selected as the offspring, thus affecting the final result. When considering the control effect of the system, hardware disturbances are inevitable, and such disturbances will affect the performance of the control system. Therefore, the integral of the square of the error (ISE) and the dynamic performance of the system are selected as components of the fitness function. The fitness function is as follows:

[0064]

[0065] Where: f is the fitness value; e(t) is the difference between the set value and the process value; OS is the overshoot, when OS is positive, it takes the OS value, when OS is negative, it takes 0; SSE is the steady-state error, RT is the rise time, and PT is the peak time. o 、M e 、M r 、M pis the weight of each indicator, and the influence of each indicator on the fitness value is adjusted by setting the weight.

[0066] Selection, crossover, and mutation: In genetic algorithms, selection, crossover, and mutation are the three core genetic operations, which together determine the algorithm's search efficiency and the quality of the solution.

[0067] Selection: After each individual has its own corresponding fitness value, the genetic algorithm operation is entered for selection. In the iterative process, a combination of roulette and tournament selection is used. First, 40 individuals are randomly selected from the population, and then a certain number of individuals are randomly selected from this generation as contestants through tournament selection. Then, an individual with the smallest fitness value is selected from these contestants to enter the next generation of the population, and the execution is repeated until a sufficient number of individuals are selected to form a new generation of the population. Roulette selection can allocate selection probabilities according to fitness, ensuring that individuals with high fitness have a greater chance of being selected, but also retains the selection opportunities for individuals with low fitness, maintaining population diversity. Tournament selection further screens out individuals with higher fitness through group competition, enhancing the stability and effect of selection. Combining the two selection methods can effectively avoid the premature convergence problem caused by a single selection method.

[0068] Roulette and Tournament Selection:

[0069] ① Calculate the total fitness value F:

[0070] ②Calculate the selection probability:

[0071] ③Construct cumulative probability distribution P i : Generate a random number r and filter out P i ≥r individuals, the M individuals selected form a set S;

[0072] ④ Randomly select T individuals from the set S to form a tournament T i , select the individual with the smallest fitness value from each tournament group as the winner: Repeat the above process until a new population is formed. i is the fitness value of each individual, p i is the selection probability of each individual, f ij represents the fitness value of the jth individual in the i-th tournament group.

[0073] Crossover: The crossover operation uses mixed crossover, which is more compatible with real number coding. The offspring gene values ​​generated by this crossover method are not limited to the gene range of the parent generation, which means that the offspring has a larger search space, increasing the diversity of the population during the optimization process, making it more effective in solving problems.

[0074]

[0075] Here, x1 and x2 are two parents, y1 and y2 are two generated children, and α is a random number between 0 and 1, usually drawn from a uniform distribution.

[0076] Mutation: The mutation operation used in this paper is Gaussian mutation, which changes the original gene value by adding a Gaussian distributed random number with the current parameter value as the mean to the individual's gene value. By setting a suitable standard deviation σ in the algorithm, the algorithm can jump out of the local optimum and avoid too large a search space, making it easier for the algorithm to find offspring with better fitness during the mutation process, thereby enabling the algorithm to find the optimal solution to the problem more comprehensively and quickly.

[0077] x′=x+N(0,σ 2 )

[0078] Among them: x is the original gene value, x′ is the mutated gene value, N(0,σ 2 ) means the mean is 0 and the variance is σ 2 Gaussian distributed random numbers.

[0079] The above embodiments are not limited to the technical solutions of the embodiments themselves, and the embodiments can be combined with each other to form new embodiments. The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Any modification or equivalent replacement that does not deviate from the spirit and scope of the present invention shall be included in the scope of the technical solutions of the present invention.

Claims

1. A bionic snake-like robot based on a pneumatic soft structure, characterized in that: The invention comprises a head structure (1) and a trunk structure (2), wherein the rear end of the head structure (1) is fixed to the front end of the trunk structure (2), the trunk structure (2) is connected to an air supply system (9) via a pipeline, and the air supply system (9) is electrically connected to an air pressure control system (10), thereby realizing motion control of a bionic snake-like robot.

2. The bionic snake-like robot based on a pneumatic soft structure according to claim 1, characterized in that: The trunk mechanism (2) is formed by connecting multiple body mechanisms (3) end to end, the body mechanism (3) includes a hexagonal body structure (303), the front end of the hexagonal body structure (303) is provided with a snap-on end of a sliding connection structure (301), and the rear end is provided with a fixed end of the sliding connection structure (301), the fixed end of the sliding connection structure (301) in the front body mechanism (3) cooperates with the snap-on end of the sliding connection structure (301) in the rear body mechanism (3) to realize the connection of the two body mechanisms (3), and flexible air bags (8) are installed on both sides of the hexagonal body structure (303), and the flexible air bags (8) are connected to an air supply system (9) through a pipeline, so as to realize the movement of the bionic snake-like robot.

3. The bionic snake-like robot based on a pneumatic soft structure according to claim 2, characterized in that: The bottom of the hexagonal body structure (303) is provided with a plurality of snap-fitting grooves (302), and the snap-fitting grooves (302) are inserted with scale gripping structures (304) for increasing the friction between the bionic snake-like robot and the ground, thereby improving the smoothness of movement.

4. The bionic snake-like robot based on a pneumatic soft structure according to claim 3, characterized in that: The flexible air bag (8) comprises a plurality of folded air cavity structures (801), the adjacent surfaces of two air cavity structures (801) are sealed and connected by plastic sealing, an air inlet pipe (802) is arranged on one side of the air cavity structure (801), and the air inlet pipes (802) of two adjacent air cavity structures (801) are connected to each other through an air guide pipe (803), and an air vent (804) is arranged on the outer surface of the air cavity structure (801) at one end of the flexible air bag (8), and the air vent (804) is connected to the air source supply system (9) through a pipeline, so as to realize the expansion or contraction of the flexible air bag (8).

5. The bionic snake-like robot based on a pneumatic soft structure according to claim 4, characterized in that: The air source supply system (9) comprises an air compressor (901) and a pressure reducing valve (902), wherein the air compressor (901) is electrically connected to a power source, the air compressor (901) is connected to the pressure reducing valve (802), the pressure reducing valve (902) is connected to the vents in the plurality of flexible air bags (8) via a multi-channel air supply pipeline, and an electrical proportional valve (903) is provided on each of the multi-channel air supply pipelines.

6. The bionic snake-like robot based on a pneumatic soft structure according to claim 5, characterized in that: The air pressure control system (10) comprises a 24V constant voltage power supply (1000), a host computer (1001), a PLC control cabinet (1002), a data acquisition card (1003), a pressure sensor (7), a bending sensor (6) and a camera (4); the 24V constant voltage power supply (1000) and the camera (4) are both electrically connected to the power supply; the 24V constant voltage power supply (1000) is electrically connected to the electrical proportional valve (903) and the PLC control cabinet (1002) for stable power supply; the PLC control cabinet (1002) is electrically connected to the electrical proportional valve (903) for controlling the electrical proportional valve (903); the host computer (1001) and the PLC control cabinet are electrically connected to the electrical proportional valve (903) for controlling the electrical proportional valve (903); (1002) are electrically connected to transmit action instructions. The bending sensor (6) is installed on the front side of the head structure (1) to collect the bending angle signal of the bionic snake-like robot in real time. The camera (4) is symmetrically arranged above the head structure (1). The pressure sensor (7) is installed at the tail of the trunk mechanism (2) to detect the gas pressure value output by the electrical proportional valve (903) in real time. The camera (4), the bending sensor (6) and the pressure sensor (7) are all electrically connected to the data acquisition card (1003). The data acquisition card (1003) is electrically connected to the host computer (1001), so as to transmit the collected information to the host computer (1001) for analysis.

7. The bionic snake-like robot based on a pneumatic soft structure according to claim 6, characterized in that: The bending sensor (6) comprises a circuit board (601) and a strain gauge (603); the circuit board (601) and the strain gauge (603) are connected via a connecting line (602); the circuit board (601) is electrically connected to a data acquisition card (1003); and the strain gauge (603) is adhered to the side of the front end of the trunk mechanism (3).

8. The bionic snake-like robot based on a pneumatic soft structure according to claim 7, characterized in that: A hollow cavity (5) is provided in the center of the hexagonal structure (303) of the body, which is used to accommodate pipelines and wires, thereby improving the mobility of the bionic snake-like robot.

9. A control method for a bionic snake-like robot based on the bionic snake-like robot according to claim 8, characterized in that: The front end of the body mechanism (3) of the trunk mechanism (2) is divided into a C area and a D area, and the remaining body mechanisms (3) in the trunk mechanism (2) are divided into an A area and a B area. The A area is divided into an a1 half section and an a2 half section, and the B area is divided into a b1 half section and a b2 half section. The specific control process is as follows: S1, forward S1.1, when air is inflated into the region A of the body structure (3) in the trunk structure (2), the a1 half of the region A of the body structure (3) moves forward, and the a2 half has a tendency to move backward. Due to the friction between the scale gripping structure (304) and the tube wall, the region A of the body structure (3) moves forward; S1.2, when the area B of the body structure (3) in the trunk structure (2) is inflated, the b1 half of the area B of the body structure (3) moves forward, and the b2 half has a tendency to move backward. At the same time, when the area A of the body structure (3) in the trunk structure (2) is deflated, the a1 half of the area A of the body structure (3) has a tendency to move backward, and the a2 half moves forward. Due to the friction between the scale gripping structure (304) and the tube wall, the area B of the body structure (3) moves forward. S1.3, when air is inflated into the A region of the body structure (3) in the trunk structure (2), the a1 half of the A region of the body structure (3) moves forward, and the a2 half has a tendency to move backward. At the same time, when air is deflated into the B region of the body structure (3) in the trunk structure (2), the b1 half of the B region of the body structure (3) has a tendency to move backward, and the b2 half moves forward. Due to the friction between the scale gripping structure (304) and the tube wall, the A region of the body structure (3) moves forward. S1.4, continuously executing the above S1.2 and S1.3, thus achieving continuous forward movement; S2, Turn S2.1, a pipeline centerline model is established through the host computer (1001) and a set value is input, the camera (4) transmits the image in the pipeline to the host computer (1001), and the host computer (1001) monitors in real time; S2.2, left turn: when the host computer (1001) detects that the actual pipeline centerline is deviated to the left and the deviation value from the pipeline centerline model is far from the set value, the host computer (1001) sends a signal to the electric proportional valve (903) through the PLC control cabinet (1002), thereby inflating the D area of ​​the frontmost body structure (3) of the trunk structure (2), so that the head structure (1) bends to the left, the camera (4) transmits the image in the pipeline in real time, and the bending sensor (6) transmits the bending angle signal as a process value to the host computer (1001) in real time. The host computer (1001) uses the Python program of the genetic algorithm to optimize the PID to perform calculations, and transmits the three parameters KP, Ki, and Kd generated by the calculation to the PID controller in the host computer (1001). After the PID controller calculates the control signal, it sends it to the PLC control cabinet (1002) through the TCP connection. The PLC control cabinet (1002) then sends an adjustment signal to the electric proportional valve (903), thereby adjusting the electric proportional valve (903). The opening size of the inflation valve in the middle of the upper computer (1001) is finally used to gradually complete the left turn action of the head structure (1); when the upper computer (1001) detects that the deviation value between the actual pipeline centerline and the pipeline centerline model is close to the set value, the camera (4) transmits the image in the pipeline in real time and the bending sensor (6) transmits the bending angle signal as a process value to the upper computer (1001) in real time; the upper computer (1001) uses the Python program of the genetic algorithm to optimize the PID to perform calculations, and transmits the three parameters KP, Ki and Kd generated by the calculations to the PID controller in the upper computer (1001); the PID controller calculates the control signal and sends it to the PLC control cabinet (1002) through the TCP connection; the PLC control cabinet (1002) controls the opening size of the inflation valve in the electric proportional valve (903), thereby deflating the D area of ​​the body mechanism (3) at the front end of the trunk mechanism (2); when the bending angle signal transmitted by the bending sensor (6) to the upper computer (1001) is 0, the bionic snake-like robot resumes the forward state; S2.3, right turn: when the host computer (1001) detects that the actual pipeline centerline is biased to the right and the deviation value from the pipeline centerline model is far from the set value, the host computer (1001) sends a signal to the electric proportional valve (903) through the PLC control cabinet (1002), thereby inflating the C area of ​​the front end body structure (3) of the trunk structure (2), so that the head structure (1) bends to the right, the camera (4) transmits the image in the pipeline in real time and the bending sensor (6) transmits the bending angle signal as a process value to the host computer (1001) in real time, the host computer (1001) uses the genetic algorithm to optimize the PID Python program to perform calculations, and transmits the three parameters KP, Ki, and Kd generated by the calculation to the PID controller in the host computer (1001), the PID controller calculates the control signal and sends it to the PLC control cabinet (1002) through the TCP connection, and the PLC control cabinet (1002) controls the opening size of the inflation valve in the electric proportional valve (903) , and finally the head structure (1) gradually completes the right turn action; when the host computer (1001) detects that the deviation value between the actual pipeline centerline and the pipeline centerline model is close to the set value, the camera (4) transmits the image in the pipeline in real time and the bending sensor (6) transmits the bending angle signal as a process value to the host computer (1001) in real time, and the host computer (1001) uses the genetic algorithm to optimize the Python program of PID to perform calculations, and transmits the three parameters KP, Ki, and Kd generated by the calculation to the PID controller in the host computer (1001). After the PID controller calculates the control signal, it sends it to the PLC control cabinet (1002) through the TCP connection. The PLC control cabinet (1002) controls the opening size of the inflation valve in the electric proportional valve (903), so that the C area of ​​the front end body mechanism (3) of the trunk mechanism (2) is deflated; when the bending angle signal transmitted by the bending sensor (6) to the host computer (1001) is 0, the bionic snake-like robot resumes the forward state.

10. The control method of a bionic snake-like robot according to claim 9, characterized in that: The genetic algorithm is optimized through chromosome encoding, calculation of individual fitness function values, selection, crossover and mutation and other genetic operations; each individual represents a solution, and by calculating the fitness function value and performing selection, crossover and mutation operations, the population evolves from generation to generation and gradually approaches the optimal solution. After multiple generations of survival of the fittest, the population continues to multiply and evolve, and finally finds the best solution to the problem.

Citation Information

Patent Citations

  • Crawling type micro pipeline robot

    CN101144558A

  • Module based on paper folding structure and bionic soft motion robot

    CN116690533A

  • Bionic robot

    CN116833987A

  • Bionic biped robot based on pneumatic soft body structure and control method

    CN118493415A

  • Inchiworm robot

    KR1020110011892A

Cited By

  • Three-degree-of-freedom snakelike robot control method of variable telescopic structure and robot

    CN119772865A

  • Flexible robot motion control fault diagnosis method and system

    CN120516688A