A TSN-based teleoperation inspection robot control system and control method

Abstract

The present invention belongs to the fields of robots and industrial Internet, and specifically relates to a teleoperation inspection robot control system based on TSN, including: a host computer, a controller module, a communication module, a motor drive module, an information acquisition module, a navigation and positioning module, an image recognition module, and a storage module provided on the robot body; the host computer is used for wirelessly communicating with the controller module through the communication module to realize wireless control of the robot body and receive the environmental data transmitted back by the controller module; the controller module controls the pose of the robot body and the actions of the robotic arm of the robot body through the motor drive module; at the same time, it acquires the environmental data in the target area collected in real time by the information acquisition module and transmits the environmental data back to the host computer through the communication module; in the process of controlling the robot, the present invention can improve the traffic congestion problem generated by the data streams collected by multiple sensors and the data stream for controlling the movement of the robot.

Description

Technical Field

[0001] The present invention belongs to the fields of robots and industrial Internet, and specifically relates to a teleoperation inspection robot control system and control method based on TSN. Background Art

[0002] As the lifeblood of the national economy, the exploitation of mineral resources in China has always been carried out in an orderly manner. These mineral resources provide a stable energy supply for industrial production, provide strong guarantees for scientific and technological development and progress, and are also important materials for improving the living standards of the people. Mineral resource exploitation refers to the exploitation of solid metal and non-metal ore deposits, including open-pit mining of surface ore bodies and shallow ore bodies, and underground mining of blind ore bodies and deep ore bodies. Underground mining usually requires miners to go deep into the deep mine tunnels to carry out operation tasks. However, frequent mine tunnel safety accidents have brought huge losses of life and property. Strengthening the inspection of mine tunnels, monitoring the status of mine tunnels in real time and transmitting it to the monitoring center, and timely discovering and eliminating potential mine tunnel safety hazards will greatly reduce the occurrence of mine tunnel safety accidents. Considering that "the fewer people, the safer", we hand over the inspection work of mine tunnels to remotely controlled robots. Remotely controlled robots are a very effective solution. By remotely controlling the robots in real time and at a distance, the safety of employees can be largely guaranteed and productivity can be effectively liberated. The key issues of remotely controlled robots are the consistency and real-time issues of the master-slave system operations of the robots, including technical issues such as the real-time performance, stability of signals, and conflicts that occur when transmitting multiple signals. At the same time, it is necessary to ensure that the transmission delay is controllable in an extreme network environment.

[0003] In the prior art, a network-based robot control system with the patent application number 201810984133.5 applies a network communication module to the remote control of robots, which is convenient for use and development. The mine survey robot control system based on STM32 with the patent application number 201911173667.0 develops the microprocessor STM32F103ZET6 and uses the method of adding various sensors by virtue of the rich interfaces of STM32 to realize the remote control of mine robots, so as to achieve the purpose of monitoring. A control method and system for an outdoor survey robot based on a 5G network with the patent application number 202011064929.2 utilizes the advantages of the 5G network to realize the operation of large-flow applications on the survey robot, solves the problem of single remote control function, and improves the working process of the survey robot.

[0004] However, in terms of communication protocols, some traditional industrial Ethernet communication protocols such as PROFINET, Ethernet / IP, EtherCAT, SERCOSIII, CC-Link, POWERLINK, etc. cannot provide services that simultaneously possess real-time performance, bounded low latency, reliability, and good compatibility. Summary of the Invention

[0005] To solve the above problems, the present invention provides a TSN-based teleoperation inspection robot control system, which makes full use of the deterministic transmission characteristics of the TSN time-sensitive network, such as low latency and latency boundedness. Thus, while ensuring the consistency of the master-slave end control of the mine inspection robot, the operation process has deterministic low latency. And due to the multi-service unified bearing characteristic of the TSN technology, the cost of on-site network layout is saved, providing an excellent, practical and low-cost solution for the inspection work in the mine.

[0006] The technical solution adopted by the present invention to achieve the above object is: a TSN-based teleoperation inspection robot control system, including: a host computer, a controller module, a communication module, a motor drive module, an information acquisition module, a navigation and positioning module, an image recognition module, and a storage module provided on the robot body;

[0007] The host computer is used to wirelessly communicate with the controller module through the communication module to realize wireless control of the robot body and receive the environmental data transmitted back by the controller module;

[0008] The controller module is used to receive the manipulation instructions sent by the host computer and control the pose of the robot body and the actions of the robot body's manipulator arm through the motor drive module; at the same time, acquire the environmental data in the target area collected in real time by the information acquisition module and transmit the environmental data back to the host computer through the communication module;

[0009] The information acquisition module is used to collect the environmental data in the target area and feedback it to the controller module;

[0010] The navigation and positioning module is a Beidou navigation system, which is used to position the robot body and send the positioning data to the controller module;

[0011] The image recognition module is used to take real-time images of the target area and send the images to the host computer to realize real-time monitoring of the target area; at the same time, send the current image to the storage module for storage, so as to be called by the controller module for map construction.

[0012] The information acquisition module includes: a temperature and humidity sensor, a gas concentration sensor, a wind sensor, a distance sensor, a sound sensor, and an infrared thermal imager, all of which are connected to the controller module;

[0013] The temperature and humidity sensor, gas concentration sensor, wind sensor, distance sensor, sound sensor, and infrared thermal imager all support the TSN protocol;

[0014] The temperature and humidity sensor is a device for measuring the environmental temperature in the mine and the equipment temperature of the main mine fan shaft;

[0015] The gas concentration sensor is used to measure the concentration of methane gas in the mine;

[0016] The wind sensor is used to measure the wind speed and direction in the mine;

[0017] The distance sensor sends distance data to the image recognition module for map construction;

[0018] The sound sensor has a pair of capacitance electret microphones sensitive to sound built in. The sound wave causes the electret film in the microphone to vibrate, resulting in a change in capacitance and then generating a corresponding tiny voltage change;

[0019] The infrared thermal imager is used to detect heat sources in the mine tunnel.

[0020] The communication module is a wireless communication module or a TSN switch that supports the TSN protocol.

[0021] The upper computer includes: a remote operation module, a data display module, a data storage module, and a data diagnosis module;

[0022] The remote operation module is used to send operation instructions to the controller module to achieve remote control of the pose of the robot body and the actions of the robot body's manipulator arm;

[0023] The data display module is used to display the environmental data in the target area received from the controller module in real time;

[0024] The data storage module is used to store the environmental data in the target area sent by the controller module in real time;

[0025] The data diagnosis module is used to determine whether the environmental data in the target area sent by the controller module meets the preset standard values.

[0026] A control method for a remotely operated inspection robot control system based on TSN includes the following steps:

[0027] 1) The operator creates a new task project through the upper computer located in the monitoring center and establishes a remote connection with the robot in the mine tunnel through the communication module;

[0028] 2) The controller module collects the environmental data at the current position through the information collection module and transmits the environmental data back to the data display module of the host computer. If the data display module shows the environmental data collected by various sensors, the host computer successfully establishes a remote connection with the robot through the communication module;

[0029] 3) The remote operation module of the host computer sends operation instructions to the controller module through the communication module, and the controller module remotely controls the robot body according to the operation instructions;

[0030] 4) During the inspection process of the robot body, the controller module uses a genetic algorithm that combines the SA algorithm and the GA algorithm through the communication module to send real-time data to the host computer to prevent data transmission delay.

[0031] For step 3) above, the following steps are executed:

[0032] 3-1) The entry into the mine cave on the inspection robot controls the pose of the robot body and the actions of the robot body's manipulator through the motor drive module according to the operation instructions;

[0033] 3-2) The navigation and positioning module real-time locates the robot body and transmits the positioning data to the host computer through the controller module;

[0034] 3-3) The information collection module real-time collects the environmental data in the target area and transmits the environmental data back to the host computer through the communication module;

[0035] 3-4) The image recognition module real-time captures the real-time image of the target area and sends the images to the host computer respectively to achieve real-time monitoring of the target area; the image recognition module also sends the current image to the storage module for storage to be used for map construction through the controller module;

[0036] 3-5) The distance sensor in the information collection module real-time monitors the distance data between the inspection robot and the cave wall and stores it in the storage module;

[0037] 3-6) After the inspection of the robot body is completed, the robot is controlled to return to the origin. The controller module retrieves the distance data and environmental data in the storage module for map construction to obtain a complete mine cave modeling diagram;

[0038] 3-7) The host computer obtains the corresponding data at various locations in the mine cave and the complete mine cave modeling diagram and stores them in the data storage module. At the same time, the environmental data is compared with the preset standard values through the data diagnosis module, and the marks that do not meet the preset standard values are marked in the mine cave modeling diagram.

[0039] In step 3-6), the controller module retrieves the distance data and environmental data from the storage module to construct a map, obtaining a complete mine cave modeling map. Specifically:

[0040] (1) Collect distance data: The controller module obtains the distance data inside the mine cave through the distance sensors installed on the mobile robot.

[0041] (2) Collect environmental data: Use various environmental sensors to obtain the environmental data inside the mine cave; the environmental data includes: temperature, humidity, methane gas concentration information.

[0042] (3) Store distance data and environmental data: Store the collected distance data and environmental data in the storage module.

[0043] (4) Initialize the map: Create an initial map in the controller module, and the map uses the representation of a three-dimensional point cloud map.

[0044] (5) Extract distance data and environmental data: The controller module extracts the distance data and environmental data from the storage module for map construction, and according to specific requirements, selects to extract all the data or only part of the data.

[0045] (6) Data preprocessing: Preprocess the extracted distance data and environmental data; including steps such as noise removal, filtering, calibration, etc., to improve the accuracy and reliability of the data.

[0046] (7) Map update: Use the preprocessed distance data and environmental data to fuse or update them with the initial map; complete the map update through an extended Kalman filter, a particle filter using a probabilistic mapping algorithm or other map construction methods.

[0047] (8) Map optimization: Optimize the constructed map to improve the quality and accuracy of the map.

[0048] The map optimization includes techniques such as loop detection and loop closure optimization to correct map drift or errors caused by sensor errors or incomplete data.

[0049] (9) Map storage: Store the constructed and optimized map in the storage module for subsequent transmission to the host computer.

[0050] During the inspection process of the robot body, the controller module uses a genetic algorithm that combines the SA algorithm and the GA algorithm through the communication module to send the real-time data to the host computer. Specifically:

[0051] 4-1) Schedule the time-triggered flow using real-number coding that represents individual information better than binary coding; randomly initialize M data strings and use them as M individuals in the initial population, and set the maximum number of iterations to G;

[0052] In the GA algorithm, evaluation is based on fitness. By taking the reciprocal form of the objective function values of different individuals and adding a constant term to control the gap between individuals of the data strings, that is:

[0053]

[0054] where s is the constant term, t n is the objective function value, f(x) is the fitness function. The smaller the constant term, the smaller the fitness gap between individuals in this term, so that individual data strings with limited differences can survive and diversify the population;

[0055] 4-2) Select N elite individuals from the population formed by M data strings during the operation and directly add them to the next-generation population;

[0056] 4-3) Select individuals from the current population for crossover with a probability of P C ; the offspring are placed in the buffer pool; repeat step 4-3) until there are (M - N) individuals in the buffer pool;

[0057] 4-4) Select individuals with a mutation probability of P M from the buffer pool, introduce the SA algorithm under the Metropolis criterion, and accept the new solution with a set probability to obtain the new generation population P(t + 1);

[0058] In step 4-4), the process of introducing the SA algorithm under the Metropolis criterion and accepting the new solution with a set probability to obtain the new generation population P(t + 1) includes the following steps:

[0059] (1) Initialize parameters: generation gen = 0, maximum number of iterations G, population size M, crossover probability P C , mutation probability P M , starting temperature t0, termination temperature t t , cooling rate Δt;

[0060] (2) Generate the initial population P t using real coding;

[0061] (3) Calculate the fitness f(x t ) of each individual in the population P i and determine the optimal individual;

[0062] (4) Select the N elite individuals with the highest fitness values and directly add them to the next-generation population;

[0063] (5) For the remaining (M - N) individuals, perform crossover and mutation with probabilities of P C and P M respectively, and calculate and compare the fitness values f(x i+1 );

[0064] If f(x i+1 ) > f(x i ), accept the new individual and jump to step (7).

[0065] (6) t i = t i - Δt, where t i is the current temperature and Δt is the cooling rate. Update the temperature t i ; If t i > t t , execute step (5).

[0066] (7) Combine the (M - N) individuals after crossover and mutation with the N elite individuals in step (4) to form a new population P(t + 1).

[0067] (8) Evaluate the fitness of the individuals in the new population P(t + 1) and save the optimal individual;

[0068] (9) Determine whether the algorithm meets the convergence criterion. If it does, output the optimal individual and terminate the algorithm; otherwise, repeat steps (4) to (9).

[0069] During data transmission, add a specific identifier or label to the optimal data string so that the host computer can recognize and correctly process it. Define a field or marker in the data transmission format to indicate that the data is the optimal data string;

[0070] To reflect the queuing delay during transmission, set up a data queue at both the sending end and the receiving end to store data to be transmitted or processed; when the data is ready, add it to the sending queue and retrieve the data from the queue at a set rate for transmission. The receiving end also sets up a queue to receive and process the transmitted data in sequence.

[0071] The present invention has the following beneficial effects and advantages:

[0072] 1. In the present invention, multiple modules all adopt the TSN technology, which can improve the traffic congestion problem generated by the data streams collected by various sensors and the data streams for controlling the movement of the robot during the process of controlling the robot, achieving unified bearing of multiple services, ensuring the real-time, stability and reliability of data transmission, and saving the costs of network deployment and operation at the same time.

[0073] 2. In the data transmission process based on the TSN protocol in the present invention, an improved genetic algorithm is adopted, which is superior to the genetic algorithm in aspects such as local search. Before the new individuals generated by the genetic operation enter the next generation population, the improved genetic algorithm can avoid local optimal traps.

[0074] 3. The improved genetic algorithm of the present invention combines the advantages of the SA algorithm in global parallel search and the advantages of the GA algorithm in local serial search, providing an effective tool for solving the TSN traffic scheduling problem.

[0075] 4. By setting up an information collection module in the present invention, the environment in the mine tunnel can be monitored in real time from aspects such as the temperature and humidity, gas concentration, wind direction and wind force in the mine tunnel, and the data obtained through the monitoring is transmitted to the upper computer software, so as to obtain the real-time information in the mine tunnel and help the operators adjust the operations at the same time. By applying the TSN technology, while transmitting the sensor traffic with high bandwidth and large flow, it is ensured that the robot can move in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 Schematic diagram of the control system structure of the mine tunnel inspection robot of the present invention;

[0077] Figure 2 Specific process diagram of the improved genetic algorithm in the TSN data traffic scheduling process of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0078] The present invention will be further described in detail below with reference to the drawings and embodiments.

[0079] Traditional Ethernet has uncertainty, which limits the scalability and performance of modular robots. By applying the TSN technology, the time-sensitive traffic is completely isolated from the low-priority traffic through the TSN time-aware shaper, ensuring the real-time motion control of the robot while transmitting the high-bandwidth sensor traffic.

[0080] As Figure 1 shown, it is a schematic diagram of the control system structure of the mine tunnel inspection robot of the present invention. The present invention is based on a teleoperated inspection robot. The body of the inspection robot includes: a crawler-type mobile structure and a grabber manipulator structure.

[0081] The control system of a TSN-based teleoperation inspection robot of the present invention includes a software part and a hardware part;

[0082] (1) Software part

[0083] The software part is a host computer developed based on Linux. The host computer is used to wirelessly communicate with the controller module through the communication module to realize wireless control of the robot body and receive the environmental data transmitted back by the controller module; the movement of the robot can be operated by touch screen or connected handle, and a display screen is set to display the state of the robot and the collected data. It mainly includes: a remote control module, a data display module, a data storage module, and a data diagnosis module;

[0084] The remote control module is used to send control instructions to the controller module to realize remote control of the pose of the robot body and the actions of the manipulator of the robot body;

[0085] The data display module is used to display the environmental data in the target area transmitted by the controller module in real time;

[0086] The data storage module is used to store the environmental data in the target area transmitted by the controller module in real time;

[0087] The data diagnosis module is used to judge whether the environmental data in the target area transmitted by the controller module meets the preset standard value.

[0088] (2) Hardware part

[0089] In terms of hardware, it includes: a controller module, a communication module, a motor drive module, an information collection module, a navigation and positioning module, an image recognition module, and a storage module provided on the robot body; the movement and sensors of the robot need to transmit traffic with different characteristics, and the controller and various sensors need to support the TSN protocol. To ensure the motion control of the robot, the actuator requires hard real-time and low-latency traffic, while the modules responsible for collecting data, such as cameras and temperature and humidity sensors, can generate a large amount of data, as Figure 1 shown, the controller module is connected to the motor module responsible for the movement of the robot through the motor drive module, and then the motor drive module is connected to various data collection modules to transmit the motor parameters to the controller module.

[0090] The controller module is used to receive the control instructions sent by the host computer and control the pose of the robot body and the actions of the manipulator of the robot body through the motor drive module; at the same time, it obtains the environmental data in the target area collected by the information collection module in real time and transmits the environmental data back to the host computer through the communication module;

[0091] The communication module is a wireless communication module or a TSN switch supporting the TSN protocol.

[0092] An information acquisition module, which is used to collect environmental data in the target area and feedback it to the controller module; it includes: a temperature and humidity sensor, a gas concentration sensor, a wind sensor, a distance sensor, a sound sensor, and an infrared thermal imager, all of which are connected to the controller module;

[0093] The temperature and humidity sensor, the gas concentration sensor, the wind sensor, the distance sensor, the sound sensor, and the infrared thermal imager all support the TSN protocol;

[0094] The temperature and humidity sensor is a device for measuring the environmental temperature in the mine and the equipment temperature of the main mine fan shaft;

[0095] The gas concentration sensor is used to measure the concentration of methane gas in the mine;

[0096] The wind sensor is used to measure the wind speed and direction in the mine;

[0097] The distance sensor sends distance data to the image recognition module for map construction;

[0098] The sound sensor has a pair of condenser electret microphones sensitive to sound built in. The sound wave causes the electret film in the microphone to vibrate, resulting in a change in capacitance, and then generating a corresponding tiny voltage change;

[0099] The infrared thermal imager is used to detect heat sources in the mine tunnel.

[0100] The navigation and positioning module is a Beidou navigation system, which is used to position the robot body and send the positioning data to the controller module. At the same time, through built-in map software such as Amap, the driving path of the robot is displayed in real time;

[0101] The image recognition module is used to take real-time images of the target area and send the images to the upper computer to realize real-time monitoring of the target area; at the same time, the current image is sent to the storage module for storage for map construction through the controller module.

[0102] The hardware part also includes a power supply module for powering each module. The power supply module uses a 48V 50Ah lithium battery pack;

[0103] As Figure 1 shown, a control method for a TSN-based remote operation inspection robot control system of the present invention includes the following steps:

[0104] 1) The operator creates a new task item through the upper computer located in the monitoring center and establishes a remote connection with the robot in the mine tunnel through the communication module;

[0105] 2) The controller module collects the environmental data of the current location through the information collection module and transmits the environmental data back to the data display module of the host computer. If the data display module displays the environmental data collected by various sensors, the host computer successfully establishes a remote connection with the robot through the communication module;

[0106] 3) The remote operation module of the host computer sends operation instructions to the controller module through the communication module, and the controller module remotely controls the robot body according to the operation instructions;

[0107] 3-1) The inspection robot enters the mine cave and controls the pose of the robot body and the actions of the robot body's manipulator through the motor drive module according to the operation instructions;

[0108] 3-2) The navigation and positioning module real-time locates the robot body and sends the positioning data to the host computer through the controller module;

[0109] 3-3) The information collection module real-time collects the environmental data in the target area and transmits the environmental data back to the host computer through the communication module;

[0110] 3-4) The image recognition module real-time captures the real-time image of the target area and sends the images to the host computer respectively to achieve real-time monitoring of the target area; the image recognition module also sends the current image to the storage module for storage to be used for map construction by calling through the controller module;

[0111] 3-5) The distance sensor in the information collection module real-time monitors the distance data between the inspection robot and the cave wall and stores it in the storage module;

[0112] 3-6) After the inspection of the robot body is completed, control the robot to return to the origin. The controller module retrieves the distance data and environmental data in the storage module for map construction to obtain a complete mine cave modeling diagram;

[0113] (1) Collect distance data: The controller module obtains the distance data inside the mine cave through the distance sensor installed on the mobile robot;

[0114] (2) Collect environmental data: Use various environmental sensors to obtain the environmental data inside the mine cave; the environmental data includes: temperature, humidity, methane gas concentration information;

[0115] (3) Store distance data and environmental data: Store the collected distance data and environmental data in the storage module;

[0116] (4) Initialize the map: Create an initial map in the controller module, and the map uses the representation method of a three-dimensional point cloud map;

[0117] (5) Extract distance data and environmental data: The controller module extracts distance data and environmental data from the storage module for map construction. According to specific requirements, all data or only part of the data can be selected for extraction;

[0118] (6) Data preprocessing: Preprocess the extracted distance data and environmental data; including steps such as noise removal, filtering, and calibration to improve the accuracy and reliability of the data;

[0119] (7) Map update: Use the preprocessed distance data and environmental data to fuse or update them with the initial map; complete the map update through an extended Kalman filter, a particle filter using a probabilistic mapping algorithm, or other map construction methods;

[0120] (8) Map optimization: Optimize the constructed map to improve the quality and accuracy of the map;

[0121] The map optimization includes techniques such as loop detection and loop closure optimization to correct map drift or errors caused by sensor errors or incomplete data;

[0122] (9) Map storage: Store the constructed and optimized map in the storage module for subsequent transmission to the host computer.

[0123] 3 - 7) The host computer obtains the corresponding data at various locations in the mine cave and the complete mine cave modeling diagram, and stores them in the data storage module. At the same time, the environmental data is compared with the preset standard values through the data diagnosis module, and the marks that do not meet the preset standard values are marked on the mine cave modeling diagram.

[0124] 4) During the inspection process of the robot body, the controller module uses a genetic algorithm that combines the SA algorithm and the GA algorithm through the communication module to send real - time data to the host computer to prevent data transmission delay.

[0125] As Figure 2 shown, it is the specific process diagram of the improved genetic algorithm in the TSN - based data traffic scheduling process of the present invention. Among them, different types of data are transmitted in the same network, and the slow response speed of the device and network delay are inevitable. Possible network delays include propagation delay, processing delay, transmission delay, and queuing delay. The present invention proposes an algorithm aiming to achieve deterministic network delay and jitter, laying a foundation for real - time communication. In TSN, the propagation delay is bounded and thus deterministic; the processing delay can also be considered deterministic; the transmission delay is also bounded and deterministic as long as the traffic has a constant bit rate. Therefore, this algorithm only needs to reduce the uncertain queuing delay of time - triggered services.

[0126] The SA algorithm is combined with the GA algorithm to form an improved genetic algorithm, which is superior to the genetic algorithm in aspects such as local search. Before the new individuals generated by the genetic operation enter the next-generation population, the SA operation is superior to the new individuals generated by the genetic operation in avoiding local optimal traps.

[0127] During the inspection process of the robot body, the controller module uses a genetic algorithm that combines the SA algorithm and the GA algorithm through the communication module to send real-time data to the upper computer, including the following steps:

[0128] 4-1) Schedule the time-triggered flow using real-number coding that reflects individual information rather than binary coding; randomly initialize M data strings and use them as M individuals in the initial population, and set the maximum number of iterations to G;

[0129] In GA, it is evaluated according to fitness. This requires a good fitness function. Since our optimization is a minimization problem, the reciprocal form of the objective function values of different individuals can be taken, and a constant term is added to control the gap between them, that is:

[0130]

[0131] where s is the constant term, t n is the objective function value, f(x) is the fitness function, the smaller the constant term, the smaller the fitness gap between individuals in this term. In this case, individuals with limited differences can survive, making the population more diverse; therefore, this algorithm can effectively avoid local optimal traps.

[0132] 4-2) Select N elite individuals from the population formed by M data strings during the operation and directly add them to the next-generation population;

[0133] 4-3) Select individuals from the current population for crossover with a probability of P C ; the offspring are put into the buffer pool; repeat step 4-3) until there are (M - N) individuals in the buffer pool;

[0134] 4-4) Select individuals with a mutation probability of P M from the buffer pool, introduce the SA algorithm under the Metropolis criterion, and accept the new solution with a set probability to obtain a new generation of population P(t + 1);

[0135] In step 4-4), introducing the SA algorithm under the Metropolis criterion and accepting the new solution with a set probability to obtain a new generation of population P(t + 1) includes the following steps:

[0136] (1) Initialize parameters: the evolution generation gen = 0, the maximum number of iterations G, the population size M, the crossover probability PC , mutation probability P M , initial temperature t0, termination temperature t t , cooling rate Δt;

[0137] (2) Generate the initial population P through real coding t ;

[0138] (3) Calculate the fitness f(x t ) of each individual in the population P i and determine the optimal individual;

[0139] (4) Select the N elite individuals with the highest fitness values and directly add them to the next-generation population;

[0140] (5) For the remaining (M - N) individuals, perform crossover and mutation with probabilities P C and P M respectively, and calculate and compare the fitness values f(x i+1 ) of all individuals;

[0141] If f(x i+1 ) > f(x i ), then accept the new individual and jump to step (7).

[0142] (6) t i = t i - Δt, where t i is the current temperature and Δt is the cooling rate, update the temperature t i ; If t i > t t , execute step (5).

[0143] (7) Combine the (M - N) individuals after crossover and mutation with the N elite individuals in step (4) to form a new population P(t + 1).

[0144] (8) Evaluate the fitness of the individuals in the new population P(t + 1) and save the optimal individual;

[0145] (9) Determine whether the algorithm meets the convergence criterion. If it does, output the optimal individual and terminate the algorithm; otherwise, repeat steps (4) to (9).

[0146] The improved genetic algorithm in the present invention combines the advantages of the SA algorithm in global parallel search and the advantages of the GA algorithm in local serial search, providing an effective tool for solving the TSN traffic scheduling problem.

[0147] Finally, during the data transmission process, add specific identifiers or tags to the optimal data string so that the host computer can recognize and process it correctly. Define a field or marker in the data transmission format to indicate that the data is the optimal data string;

[0148] To reflect the queuing delay during the transmission process, set up a data queue at both the sending end and the receiving end to store data to be transmitted or processed; when the data is ready, add it to the sending queue and retrieve the data from the queue at a set rate for transmission. The receiving end also sets up a queue to receive and process the transmitted data in sequence.

[0149] Although the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, it is not a limitation of the protection scope of the present invention. Those skilled in the art should understand that based on the technical solution of the present invention, various modifications or deformations that can be made without creative labor by those skilled in the art are still within the protection scope of the present invention.

Claims

1. A control method for a teleoperation inspection robot control system based on TSN, characterized in that, Its control system includes: a host computer, a controller module, a communication module, a motor drive module, an information acquisition module, a navigation and positioning module, an image recognition module, and a storage module provided on the robot body; The host computer is used to wirelessly communicate with the controller module through the communication module to realize wireless control of the robot body and receive the environmental data transmitted back by the controller module; The controller module is used to receive the manipulation instructions sent by the host computer and control the pose of the robot body and the actions of the robot body's manipulator arm through the motor drive module; at the same time, obtain the environmental data in the target area collected in real time by the information acquisition module and transmit the environmental data back to the host computer through the communication module; The information acquisition module is used to collect the environmental data in the target area and feedback it to the controller module; The navigation and positioning module is a Beidou navigation system, which is used to position the robot body and send the positioning data to the controller module; The image recognition module is used to capture the real-time image of the target area and send the image to the host computer to realize the real-time monitoring of the target area; at the same time, send the current image to the storage module for storage to be used for map construction through the controller module; The control method includes the following steps: 1) The operator creates a new task project through the host computer located in the monitoring center and establishes a remote connection with the robot in the mine through the communication module; 2) The controller module collects the environmental data at the current position through the information acquisition module and transmits the environmental data back to the data display module of the host computer. If the data display module displays the environmental data collected by various sensors, the host computer successfully establishes a remote connection with the robot through the communication module; 3) The remote manipulation module of the host computer sends manipulation instructions to the controller module through the communication module, and the controller module remotely controls the robot body according to the manipulation instructions; 4) During the inspection process of the robot body, the controller module uses a genetic algorithm that combines the SA algorithm and the GA algorithm through the communication module to send the real-time data to the host computer to prevent data transmission delay; During the inspection process of the robot body, the controller module uses a genetic algorithm that combines the SA algorithm and the GA algorithm through the communication module to send the real-time data to the host computer. Specifically: 4-1) Use real number coding that reflects individual information better than binary coding to schedule the time-triggered flow; randomly initialize M data strings and use them as M individuals in the initial population, and set the maximum number of iterations to G; In the GA algorithm, the evaluation is based on fitness. By taking the reciprocal form of the objective function values of different individuals and adding a constant term to control the gap between individuals of the data strings, that is: where s is a constant term, and t n is the objective function value, f(x) is the fitness function. The smaller the constant term, the smaller the fitness gap between individuals in this term, so that individual data strings with limited differences can survive and diversify the population; 4-2) In the operation, select N elite individuals from the population formed by M data strings and directly add them to the next generation population; (4-3) Select individuals from the current population for crossover with a probability of P C ; The offspring are placed in the buffer pool; Repeat step (4-3) until there are (M - N) individuals in the buffer pool; 4-4) Select individuals with a mutation probability of P from the buffer pool M Introduce the SA algorithm under the Metropolis criterion, accept the new solution with a set probability, and obtain a new generation of population P(t+1); In step 4-4), the SA algorithm is introduced under the Metropolis criterion, and new solutions are accepted under the set probability to obtain a new generation population P(t+1), including the following steps: (1) Initialize parameters: evolutionary generation gen = 0, maximum number of iterations G, population size M, crossover probability P C , mutation probability P M , initial temperature t0, termination temperature t t , cooling rate Δt; (2) Generate the initial population P through real coding t ; (3) Calculate the population P t and determine the fitness f(x i ) of each individual in it, and determine the optimal individual; (4) Select the N elite individuals with the highest fitness values and directly add them to the next generation population; (5) Respectively perform crossover and mutation on the remaining (M - N) individuals with probabilities of P C and P M , and calculate and compare the fitness values f(x i+1 ) of all individuals; If f(x i+1 ) > f(x i ), then accept the new individual and jump to step (7); (6)t i = t i - Δt, t i is the current temperature, Δt is the cooling rate, update the temperature t i ; if t i > t t , perform step (5); (7) The (M - N) individuals after cross - mutation are combined with the N elite individuals in step (4) to form a new population P(t + 1); (8) Evaluate the fitness of the individuals in the new population P(t + 1) and save the optimal individual; (9) Determine whether the algorithm meets the convergence criterion. If it meets, output the optimal individual and terminate the algorithm; otherwise, repeat steps (4) - (9).

2. The control method of a TSN-based teleoperation inspection robot control system according to claim 1, characterized in that, The information acquisition module includes: a temperature - humidity sensor, a gas concentration sensor, a wind sensor, a distance sensor, a sound sensor, and an infrared thermal imager, all of which are connected to the controller module; The temperature - humidity sensor, gas concentration sensor, wind sensor, distance sensor, sound sensor, and infrared thermal imager all support the TSN protocol; The temperature - humidity sensor is used to measure the environmental temperature in the mine and the equipment temperature of the main mine fan shaft; The gas concentration sensor is used to measure the concentration of methane gas in the mine; The wind sensor is used to measure the wind speed and direction in the mine; The distance sensor sends distance data to the image recognition module for map construction; The sound sensor has a pair of capacitance electret microphones sensitive to sound built - in. The electret film in the microphone vibrates through sound waves, causing a change in capacitance, and then generating a corresponding tiny voltage change; The infrared thermal imager is used to detect heat sources in the mine tunnel; 3. The control method of a TSN-based teleoperation inspection robot control system according to claim 1, characterized in that The communication module is a wireless communication module or a TSN switch that supports the TSN protocol; 4. The control method of a TSN-based teleoperation inspection robot control system according to claim 1, wherein, The upper computer includes: a remote control module, a data display module, a data storage module, and a data diagnosis module; The remote control module is used to send control instructions to the controller module to remotely control the pose of the robot body and the actions of the robot body's manipulator arm; The data display module is used to display the environmental data in the target area received from the controller module in real - time; The data storage module is used to store the environmental data in the target area sent by the controller module in real - time; The data diagnosis module is used to determine whether the environmental data in the target area sent by the controller module meets the preset standard values; 5. The control method of a TSN-based teleoperation inspection robot control system according to claim 1, characterized in that, In step 3), the following steps are executed: 3 - 1) The inspection robot enters the mine tunnel and controls the pose of the robot body and the actions of the robot body's manipulator arm through the motor drive module according to the control instructions; 3 - 2) The navigation and positioning module locates the robot body in real - time and sends the positioning data to the upper computer through the controller module; 3 - 3) The information acquisition module collects the environmental data in the target area in real - time and transmits the environmental data back to the upper computer through the communication module; 3 - 4) The image recognition module takes real - time images of the target area in real - time and sends the images to the upper computer respectively to realize real - time monitoring of the target area; the image recognition module also sends the current image to the storage module for storage, which is used for map construction through the controller module; 3 - 5) The distance sensor in the information acquisition module monitors the distance data between the inspection robot and the cave wall in real - time and stores it in the storage module; After the inspection of the robot body in 3-6 is completed, the robot is controlled to return to the origin. The controller module retrieves the distance data and environmental data in the storage module to construct a map and obtains a complete mine tunnel modeling diagram. In 3-7, the host computer obtains the corresponding data at various locations in the mine tunnel and the complete mine tunnel modeling diagram, and stores them in the data storage module. At the same time, through the data diagnosis module, the environmental data is compared with the preset standard values, and the marks that do not meet the preset standard values are marked on the mine tunnel modeling diagram.

6. The control method of a TSN-based teleoperation inspection robot control system according to claim 5, characterized in that, In step 3-6, the controller module retrieves the distance data and environmental data in the storage module to construct a map and obtains a complete mine tunnel modeling diagram. Specifically: (1) Collect distance data: The controller module obtains the distance data inside the mine tunnel through the distance sensors installed on the mobile robot. (2) Collect environmental data: Use various environmental sensors to obtain the environmental data inside the mine tunnel; the environmental data includes: temperature, humidity, methane gas concentration information. (3) Store the distance data and environmental data: Store the collected distance data and environmental data in the storage module. (4) Initialize the map: Create an initial map in the controller module, and the map uses the representation of a three-dimensional point cloud map. (5) Extract the distance data and environmental data: The controller module extracts the distance data and environmental data from the storage module for map construction, and according to specific requirements, selects to extract all the data or only part of the data. (6) Data preprocessing: Preprocess the extracted distance data and environmental data; including noise removal, filtering, calibration steps to improve the accuracy and reliability of the data. (7) Map update: Use the preprocessed distance data and environmental data to fuse or update them with the initial map; complete the map update through an extended Kalman filter, particle filter using a probabilistic mapping algorithm or other map construction methods. (8) Map optimization: Optimize the constructed map to improve the quality and accuracy of the map. The map optimization includes loop detection and loop closure optimization techniques to correct map drift or errors caused by sensor errors or incomplete data. (9) Map storage: Store the constructed and optimized map in the storage module for subsequent transmission to the host computer.

7. The control method of a TSN-based teleoperation inspection robot control system according to claim 1, characterized in that, During data transmission, add a specific identifier or label to the optimal data string so that the host computer can recognize and process it correctly. Define a field or mark in the data transmission format to indicate that the data is the optimal data string. To reflect the queuing delay during the transmission process, data queues are set at both the sending end and the receiving end for storing data to be transmitted or processed; when the data is ready, add it to the sending end queue and retrieve the data from the queue at a set rate for transmission. A queue is also set at the receiving end for receiving and processing the transmitted data in order.

Images