Unmanned tractor operation path tracking and machine tool cooperative control system and method

Abstract

The invention relates to an unmanned tractor operation path tracking and machine tool cooperative control system and method, and the system obtains the position and posture of a tractor, machine tool operation parameters and topographic information in real time through fusing the data of multi-source sensors, such as GNSS positioning and laser radar topographic perception. A multi-objective optimization function including a path tracking error, an operation precision error and a ground vibration excitation amount is constructed, and a particle swarm-genetic algorithm (PSO-GA) hybrid optimization algorithm is introduced for result optimization, so that cooperative control of tractor unit operation parameters such as a steering angle, a vehicle speed and a machine tool hydraulic parameter is realized. According to the method, in the operation of the unmanned tractor in the hilly terrain, the path tracking error is effectively reduced, the operation uniformity is improved, the influence of terrain factors on the operation is reduced, and the operation efficiency and the equipment durability under complex working conditions are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural machinery automation control and optimization algorithms, and specifically to an unmanned tractor operation path tracking and implement cooperative control system and method. Background Art

[0002] China is a country with diverse and complex terrains. Among them, hilly and mountainous areas are the most common terrains, accounting for about 70% of the total land area of China. When carrying out agricultural work in hilly and mountainous terrains, it is often necessary to face a slope environment of 6° - 15°. When a hilly and mountainous tractor works on a slope, it is prone to uneven tillage depth, which affects the operation efficiency and the growth of crops.

[0003] With the development of agricultural modernization, the automation level of agricultural machinery has been continuously improved. Especially in the cooperative operation of tractors and implements, control technology has become a key factor in improving operation accuracy and efficiency. Traditional agricultural machinery control systems usually separate tractor path tracking and implement operation control and optimize them independently, making it difficult to fully consider the overall performance of the system. Such control methods ignore the interaction between multiple factors, resulting in the overall performance of the system not reaching the optimal level. An unmanned electric tractor is a modern agricultural equipment integrating autonomous driving, new energy power, and intelligent perception technologies, which can autonomously complete field operations such as tillage, sowing, and fertilization through remote control or preset programs. It has the advanced advantages of high intelligence and high operation efficiency. However, in a complex terrain or an operation environment with large changes in ground conditions, sudden ground loads (such as hard blocks, tree roots, etc.) can have a significant impact on the control of tractors and implements. Therefore, the research and development of unmanned electric tractors are extremely urgent. Summary of the Invention

[0004] To solve the above problems, the purpose of the present invention is to provide an unmanned tractor operation path tracking and implement cooperative control system and method, which can comprehensively optimize path tracking, operation accuracy, and vibration control while considering the influence of sudden ground loads, improve the overall operation effect, effectively improve the accuracy of path tracking and implement operation during the operation of an unmanned tractor in a tractor set, and the anti-interference ability against ground vibration excitation in complex terrains. Based on multi-objective theory and PID control, it realizes the cooperative control of tractor trajectory tracking and implement operation. By optimizing the coordination of path tracking, operation accuracy, and vibration control, it achieves the efficient cooperative operation of tractors and implements in the face of sudden ground loads.

[0005] The core of the present invention is to comprehensively consider the influence of the tractor's path tracking error, machine operation error and vibration excitation through a multi-objective optimization function to achieve the improvement of overall operation efficiency and accuracy. During operation, first, various sensors (such as accelerometers, lidars, visual sensors, etc.) installed on the tractor and the machine are used to monitor the ground conditions and sudden load changes in real time. These sensors continuously collect information about ground unevenness, obstacles, hard blocks, tree roots, etc., and the system transmits this information to the control center for further processing.

[0006] An unmanned tractor operation path tracking and machine tool coordinated control system comprises a sensor system, a computing and control system, an actuator, and a power supply system.

[0007] The sensor system includes a GNSS / RTK positioning device, a laser radar, a visual sensor, an inertial measurement unit (IMU), an acceleration or vibration sensor, and a machine operation sensor.

[0008] Furthermore, the tool operation sensor includes a depth sensor and a torque sensor.

[0009] Furthermore, the computing and control system includes a central control unit CPU and a wireless communication module.

[0010] Furthermore, the actuator specifically includes an electric or hydraulic steering system, a switch or vehicle speed control system, a machine hydraulic control system, and a vibration suppression system. The vibration suppression system specifically includes an active suspension control module and an adjustable damping shock absorber.

[0011] Furthermore, the power supply system specifically includes a tractor's own generator and battery and an independent power supply module UPS.

[0012] Furthermore, each sensor in the sensor system collects the tractor's posture information data, heading, speed, ground environment characteristics and work completion information data in real time, and transmits the information data to the central control unit CPU through the wireless communication module in the computing and control system.

[0013] Furthermore, the central control unit CPU first fuses and pre-processes the raw data from each sensor, then constructs a multi-objective optimization function, and then uses a particle swarm-genetic (PSO-GA) hybrid optimization algorithm to optimize the calculation results of the multi-objective optimization function. The central control unit CPU outputs execution instructions to the electric or hydraulic steering system, power switch or vehicle speed control system, machine hydraulic control system, and vibration suppression system in the actuator according to the optimized data.

[0014] Further, in the power supply system, the on-board generator and battery of the tractor supply power to the actuators, and the independent power supply module UPS supplies power to the sensor system, computing and control system.

[0015] The GNSS / RTK positioning device, lidar, and vision sensor are installed on the front or roof of the tractor, the inertial measurement unit (IMU) is installed on the chassis of the tractor, and multiple acceleration or vibration sensors are respectively installed on the chassis of the tractor and the implement. The GNSS / RTK positioning device, lidar, vision sensor, inertial measurement unit (IMU), and acceleration or vibration sensor collect the pose information data, heading, speed, ground environment characteristics, and operation completion information data of the tractor in real time. The collected data is transmitted to the central control unit CPU through the wireless communication module to ensure the real-time and integrity of data collection.

[0016] During the data processing, the central control unit CPU first fuses and then preprocesses the raw data from different sensors. Fusion refers to integrating multi-source information such as the GNSS / RTK positioning device, lidar, vision sensor, and IMU into a unified environmental model through a data fusion algorithm to generate a comprehensive data stream reflecting the position, attitude, and ground conditions of the tractor; preprocessing is to filter and normalize the fused data and extract key features such as path deviation, operation error, and vibration excitation parameters. These fused and preprocessed data not only provide a basis for control decisions but also lay a foundation for the construction of a multi-objective optimization function.

[0017] During the calculation of the control unit, the central control unit constructs a multi-objective optimization function using the comprehensive data stream obtained after fusion and preprocessing:

[0018] J = ω1·e path 2 + ω2·e work 2 + ω3·F vib 2

[0019] Where: J is the multi-objective optimization function, representing the comprehensive performance of the tractor and implement system; ω1, ω2, ω3 are all weight coefficients, respectively corresponding to adjusting the relative importance of trajectory error, operation error, and vibration excitation in the multi-objective optimization function; e path is the trajectory error, representing the deviation between the actual driving path of the tractor and the predetermined path; e work is the operation error, representing the deviation of the implement relative to the target standard (such as depth, width, etc.) during the operation; F vibIt is force vibration, representing the vibration intensity caused by the uneven ground or operating force of the tractor and the working implement.

[0020] By changing the relative values of the weight coefficients ω1, ω2, and ω3, the optimization priorities of the trajectory error, operation error, and vibration excitation in different operation scenarios are controlled. The specific selection methods include: in high-precision path tracking tasks, increase the weight of ω1 and reduce the weights of ω2 and ω3; in fine operation tasks (such as precision seeding and fertilization), increase the weight of ω2 and appropriately adjust ω1 and ω3; in complex terrains, increase the weight of ω3 to preferentially reduce the impact of vibration on operation accuracy and stability. By selecting appropriate weight coefficients ω1, ω2, and ω3, the importance of each error term in the optimization process is adjusted, where: ω1 + ω2 + ω3 = 1 and ω1, ω2, ω3 ∈ [0, 1]. The value range of the weight coefficients ensures the balance of the multi-objective optimization function and flexibly adjusts the optimization focus according to different operation tasks.

[0021] Furthermore, ω1 + ω2 + ω3 = 1 and ω1 ∈ [0.35, 0.45], ω2 ∈ [0.35, 0.45], ω3 ∈ [0.1, 0.3].

[0022] Furthermore, the trajectory error e path is obtained by calculating the lateral deviation e y between the current position of the tractor and the predetermined path or the heading angle deviation e θ Specifically:

[0023]

[0024] In the calculation of the trajectory error, it is mainly divided into the straight path tracking error and the curved path tracking error according to the operation route. The calculation of both requires considering the lateral deviation e y (the vertical distance between the actual path and the target path) and the heading angle deviation e θ (the error between the driving direction of the tractor on the curved path and the tangent direction of the point closest to the target path on the target path). The weights k1 = 0 or 1, k2 = 0 or 1, but they are not both 0 or 1 at the same time.

[0025] Furthermore, the operation error e work is the deviation of the implement from the predetermined target (such as depth, width, or application amount) during the operation, with the unit of millimeters (mm) or percentage. e work has multiple types, including respectively referring to different types of operation errors such as operation depth and operation width. In actual operation, e in the multi-objective optimization function J workIt is the average value of various types of operation errors. The calculation methods include: in tillage operations, calculating the difference between the actual tillage depth and the set tillage depth; in seeding operations, calculating the differences between the actual seeding depth and width and the set seeding depth and width; in fertilizing or spraying operations, calculating the deviation between the actual application amount and the predetermined application amount. Specifically:

[0026]

[0027] e workn = |h actual - h target |

[0028] Where: the actual operation amount is h actual , and the target operation amount is h target .

[0029] The vibration excitation F vib is divided into the vibration excitation for the tractor and the vibration excitation for the implement The vibration accelerations of the tractor and the implement are measured respectively by acceleration or vibration sensors to represent the intensity of the system vibration, with the unit of meters per second squared (m / s 2 ), and the formulas for F vib and are as follows:

[0030]

[0031] Where: is the vertical acceleration of the tractor; is the pitch angular velocity of the tractor body; is the roll angular velocity of the tractor body; is the vertical acceleration of the implement; is the pitch angular velocity of the implement; is the roll angular velocity of the implement.

[0032] In the multi-objective optimization function, the sum of squares form F vib of the vibration excitation term F vib 2 can effectively reduce the vibration influence caused by uneven ground and operation force fluctuations during the optimization process, ensure the stability and comfort of the tractor and implement operations, reduce the wear of mechanical equipment and extend its service life.

[0033] The calculation and control system uses the particle swarm-genetic (PSO-GA) hybrid optimization algorithm to optimize the calculation results of the multi-objective optimization function, and can achieve rapid screening and convergence of the data output after calculating the multi-objective optimization function, further optimizing control parameters such as the steering angle, vehicle speed, and hydraulic parameters, so as to achieve the optimal balance between various objectives.

[0034] The PSO-GA algorithm is a combined algorithm of the PSO algorithm and the GA algorithm. Specifically, the particle swarm optimization algorithm (PSO algorithm) is first used to quickly converge the data to obtain an effective data set, and then the genetic algorithm (GA algorithm) is used for comparison correction and feedback. In the case of the combination of the two algorithms, the credibility and accuracy of the data can be improved, thereby improving the accuracy in calculations. Through the cyclic iteration method of the PSO-GA algorithm, the multi-objective optimization function J can be minimized, and the control parameters can be adjusted according to the real-time feedback of the system to achieve the optimal comprehensive performance.

[0035] The control parameters optimized by the PSO-GA algorithm are used by the central control unit CPU in the calculation and control system to output control instructions for the next cycle of data input.

[0036] A method for path tracking and implement cooperative control of an autonomous tractor operation, which is applied to the above-mentioned path tracking and implement cooperative control of an autonomous tractor operation, specifically includes the following steps:

[0037] Step1: The path tracking and implement cooperative control system of the autonomous tractor uses multiple sensors installed on the tractor and the implements to monitor the ground conditions and operation parameters in real time. The sensors collect pose information data, heading, speed, ground environment characteristics, and operation completion degree in real time, and transmit the information to the calculation and control system for subsequent processing;

[0038] Step2: The central control unit CPU in the calculation and control system performs data fusion and preprocessing on the data collected by the multiple sensors, and obtains a corresponding digital model. After that, the central control unit CPU obtains a comprehensive evaluation of the operation environment based on the above digital model, and obtains a comprehensive evaluation result. Based on this, path planning and adjustment of operation parameters are carried out. Specifically, the data analysis result is obtained through the operation of the trajectory tracking error, operation error, and ground vibration excitation, and the output is the weight coefficients ω1, ω2, ω3 within a reasonable range;

[0039] Step3: According to the weight coefficients ω1, ω2, ω3, the central control unit CPU optimizes the path tracking of the tractor and the operation accuracy of the implements. The specific implementation is: by changing the relative values of the weight coefficients ω1, ω2, ω3 to control the optimization priorities of the trajectory error, operation error, and vibration excitation in different operation scenarios;

[0040] Step4: After obtaining the appropriate weight coefficients ω1, ω2, ω3, the central control unit CPU outputs to the multi-objective optimization function J and uses the particle swarm-genetic hybrid optimization algorithm to complete the algorithm optimization process;

[0041] Step 5: When the vibration caused by the ground mutation load affects the path tracking and operation accuracy and stability of the tractor, the central control unit CPU adjusts the working state of the vibration suppression system by using the PID control algorithm through real-time monitoring of the vibration acceleration. The central control unit CPU adjusts the damping and stiffness of the shock absorber according to the real-time feedback of the vibration intensity to ensure that the system still maintains high-efficiency operation performance under vibration impact.

[0042] Step 6: The central control unit CPU realizes the control of the actuator after optimization calculation through the PID control algorithm. Specifically, in the process of optimizing the path tracking accuracy by using the PID control algorithm in the control system, the driving direction and speed of the tractor are adjusted in real time. The heading and vehicle speed of the tractor are adjusted through the proportional term, integral term and differential term of the PID controller to ensure that the tractor can quickly respond to the path error and make adjustments according to the change trend of the error. In the process of optimizing the operation accuracy by using the PID control algorithm in the control system, according to the operation error, the operation depth and width of the implement are adjusted through the proportional term, integral term and differential term of the PID controller to ensure the stability of the operation accuracy.

[0043] Step 7: The calculation and control system continuously adjusts and optimizes the operation parameters through the real-time feedback mechanism to ensure that the tractor and the implement can achieve efficient and stable collaborative operation under complex ground conditions. Specifically, in the actual operation of the tractor and the implement, the data collected by the sensor is pre-controlled through the feedforward control unit of the central control unit CPU after being calculated by the system to give the execution target, the change curvature of the actuator variable is calculated, and then the theoretical change amount of the actuator is obtained. Then, an execution instruction is output to the actuator. After the actuator works, the actual change amount of the actuator is obtained. The actual change amount and the theoretical change amount are input into the central control unit CPU at the same time to calculate the ratio, and the output is the real-time error. When the real-time error is not 1, at this time, the actual change amount is different from the theoretical change amount, and feedback optimization is required. At this time, feedback calculation is carried out in the error feedback control unit, and the feedback result is sent to the central control unit CPU to regenerate and output to the actuator for cyclic feedback adjustment; when the real-time error is 1, at this time, the actual change amount is the same as the theoretical change amount, and this execution is completed.

[0044] Through this method, the system can dynamically optimize the operation accuracy, path tracking accuracy and vibration control under the influence of the ground mutation load, and finally maximize the operation efficiency and stability.

[0045] Further, in Step 2, the raw data collected by different sensors are integrated through a data fusion algorithm to obtain an environmental model for the target operation. The model contains all the operation data. Then, through data preprocessing, the data in the environmental model are respectively extracted as: a digital model of the environmental terrain, a digital model of vibration excitation extracted by combining the vibration spectrum, a digital model of the analysis of the uniformity of the operation state generated by combining the operation volume monitoring, and a digital model of the analysis of the machine set attitude generated by combining the inertial measurement unit and the position digital model. After that, the central control unit CPU obtains a comprehensive evaluation of the operation environment based on the above multiple digital models and obtains a comprehensive evaluation result. The comprehensive evaluation result includes information such as the type and intensity of the ground sudden load, whether the operation depth and width meet the predetermined targets, etc.

[0046] Further, the specific selection methods of the weight coefficients ω1, ω2, and ω3 in Step 3 are as follows: in the high-precision path tracking task, increase the weight of ω1 and decrease the weights of ω2 and ω3; in the fine operation tasks (such as precision seeding and fertilization), increase the weight of ω2 and appropriately adjust ω1 and ω3; in the complex terrain, increase the weight of ω3 to preferentially reduce the impact of vibration on the operation accuracy and stability. In actual operation, the change of the weight coefficients is real-time, and the system will adjust in real-time during the operation according to the preset range.

[0047] Further, the specific implementation in Step 4 is as follows: the Particle Swarm Optimization - Genetic Algorithm (PSO - GA) hybrid optimization algorithm is used to solve the multi-objective optimization among path tracking, operation accuracy, and vibration control. The PSO - GA algorithm combines the global search ability of particle swarm optimization and the local fine adjustment ability of genetic algorithm to achieve the collaborative optimization among multiple objectives. The optimal combination of control parameters (such as steering angle, vehicle speed, and hydraulic parameters, etc.) is searched globally through particle swarm optimization, and then these parameters are finely adjusted by using the crossover and mutation processes of genetic algorithm to prevent the system from falling into local optimum. Through this algorithm, the system can optimize and adjust the operation parameters such as the steering angle, vehicle speed of the tractor, and the hydraulic parameters of the implement, so as to achieve the optimization of the comprehensive performance under different operation environments.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] This real-time feedback mechanism ensures that the system can respond and correct quickly in the face of complex farmland environments and ground sudden loads, so as to achieve the continuous and efficient operation of the driverless tractor and the implement in high-precision agricultural operations. In the specific use process, the present invention effectively improves the accuracy of path tracking and machine set operation of the driverless tractor during the machine set operation and the anti-interference ability to ground vibration excitation in complex terrains. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is the overall technical roadmap of an unmanned tractor operation path tracking and machine tool coordinated control system and method of the present invention;

[0051] Figure 2 It is a data acquisition and processing flow chart of an unmanned tractor operation path tracking and machine tool coordinated control system and method of the present invention;

[0052] Figure 3 It is a calculation and optimization adjustment flow chart of an unmanned tractor operation path tracking and machine tool coordinated control system and method of the present invention;

[0053] Figure 4 The present invention discloses an actuator control and real-time feedback flow chart of an unmanned tractor operation path tracking and machine tool coordinated control system and method. DETAILED DESCRIPTION

[0054] In order to better understand the content of the present invention, the present invention will be further described below in conjunction with specific embodiments and drawings. The following embodiments are implemented based on the technology of the present invention, and detailed implementation methods and operating steps are given, but the protection scope of the present invention is not limited to the following embodiments.

[0055] An unmanned tractor operation path tracking and machine tool coordinated control system comprises a sensor system, a computing and control system, an actuator, and a power supply system.

[0056] The sensor system specifically includes a GNSS / RTK positioning device, a laser radar, a visual sensor, an inertial measurement unit (IMU), an acceleration or vibration sensor, and a machine operation sensor. The machine operation sensor specifically includes a depth sensor and a torque sensor.

[0057] The calculation and control system includes a central control unit CPU and a wireless communication module.

[0058] The actuators specifically include electric or hydraulic steering systems, switch or vehicle speed control systems, implement hydraulic control systems, and vibration suppression systems. The vibration suppression system specifically includes an active suspension control module and an adjustable damping shock absorber.

[0059] The power supply system specifically includes a tractor's own generator and battery and an independent power supply module UPS.

[0060] During the sensor data acquisition process, the system obtains the pose information, heading, speed, ground environment characteristics, and operation completion information of the tractor in real time by deploying high-precision GNSS / RTK positioning devices, lidar, vision sensors, inertial measurement units (IMUs), and acceleration or vibration sensors on the tractor and implements. The specific installation method of the sensors is as follows: The GNSS / RTK positioning device, lidar, and vision sensor are installed on the front of the tractor or on the roof. The GNSS / RTK positioning device provides accurate longitude, latitude, and altitude data; the lidar and vision sensor scan the farmland terrain to capture information about the undulations, unevenness, and existing obstacles on the ground; the inertial measurement unit (IMU) is installed on the chassis of the tractor, and multiple acceleration or vibration sensors are respectively installed on the chassis of the tractor and the implement. The IMU detects the vehicle attitude, and the acceleration or vibration sensor detects the vibration data, providing the raw data basis for subsequent dynamic control. The collected data is transmitted to the central control unit through a wireless communication module to ensure the real-time and integrity of data acquisition.

[0061] During the data processing process, the central control unit first fuses and then preprocesses the raw data from different sensors. Fusion refers to integrating multi-source information such as GNSS / RTK positioning devices, lidar, vision sensors, and IMUs into a unified environmental model through a data fusion algorithm to generate a comprehensive data stream reflecting the position, attitude, and ground conditions of the tractor; preprocessing is to filter and normalize the fused data and extract key features such as path deviation, operation error, and vibration excitation parameters. These fused and preprocessed data not only provide a basis for control decisions but also lay the foundation for the construction of a multi-objective optimization function.

[0062] During the control unit calculation process, the central control unit constructs a multi-objective optimization function using the comprehensive data stream obtained after fusion and preprocessing:

[0063] J = ω1·e path 2 + ω2·e work 2 + ω3·F vib 2

[0064] Where: J is the multi-objective optimization function, representing the comprehensive performance of the tractor and implement system; ω1, ω2, ω3 are all weight coefficients, respectively corresponding to adjusting the relative importance of trajectory error, operation error, and vibration excitation in the multi-objective optimization function; e path is the trajectory error, representing the deviation between the actual driving path of the tractor and the predetermined path; e workis the operation error, representing the deviation of the machine tool relative to the target standard (such as depth, width, etc.) during operation; F vib is the vibration excitation (Force vibration), representing the vibration intensity caused by the uneven ground or operation force of the tractor and the operation implement.

[0065] By changing the relative values of the weight coefficients ω1, ω2, and ω3, the optimization priorities of the trajectory error, operation error, and vibration excitation in different operation scenarios are controlled. The specific selection methods include: in high-precision path tracking tasks, increase the weight of ω1 and decrease the weights of ω2 and ω3; in fine operation tasks (such as precision seeding and fertilization), increase the weight of ω2 and appropriately adjust ω1 and ω3; in complex terrains, increase the weight of ω3 to preferentially reduce the impact of vibration on operation accuracy and stability. By selecting appropriate weight coefficients ω1, ω2, and ω3, the importance of each error term in the optimization process is adjusted, where: ω1 + ω2 + ω3 = 1 and ω1, ω2, ω3 ∈ [0, 1]. The value range of the weight coefficients ensures the balance of the multi-objective optimization function and flexibly adjusts the optimization focus according to different operation tasks. For example, in operations on relatively flat terrains, the fineness of the operation is mainly considered, and the distribution method of ω1 = 0.4, ω2 = 0.4, ω3 = 0.2 can be adopted to ensure the operation efficiency, while in operations on complex terrains, the impact of the terrain on the tractor and the implement is mainly considered, and the distribution method of ω1 = 0.3, ω2 = 0.3, ω3 = 0.4 can be adopted to ensure the minimum interference degree of the terrain on the operation process. In actual operations, the change of the weight coefficients is real-time, and the system will adjust in real-time according to the preset range during operation. For example, when operating in a large site, the main goal is to obtain better operation quality. At this time, the value ranges of ω1 and ω2 are both large, and the set weight coefficient ranges are ω1 ∈ [0.35, 0.45], ω2 ∈ [0.35, 0.45], ω3 ∈ [0.1, 0.3]. During operation, the system obtains values within the range through the collected real-time data for efficient operation.

[0066] The trajectory error e path is obtained by calculating the lateral deviation e y or the heading angle deviation e θ between the current position of the tractor and the predetermined path, specifically:

[0067]

[0068] The specific method is as follows: In the calculation of the trajectory error, it is mainly divided into the straight-line path tracking error and the curved-line path tracking error according to the operation route. The calculation of both requires considering the lateral deviation e y (the vertical distance between the actual path and the target path) and the heading angle deviation e θ(The error between the driving direction of the tractor on the curved path and the tangent direction of the point closest to the target path). However, in actual operations, the tractor needs to continuously adjust its driving direction to make the actual path close to the target path. At this time, the weights k1 and k2 of the heading angle deviation need to be dynamically adjusted, where k1 = 0 or 1, k2 = 0 or 1, but they cannot be 0 or 1 at the same time. Specifically, when tracking a straight path, theoretically, the heading angle deviation e also needs to be considered θ , although the heading angle deviation e θ is small or 0 at this time, but the heading angle deviation e θ needs to be actively increased during path correction to return to the target path. Therefore, when calculating the straight path error, mainly the lateral deviation e y is considered. At this time, by setting the parameter k2 = 0, the weight of the heading angle deviation e θ is reduced, thereby reducing the influence of the heading angle deviation e θ on the system. Under curved path tracking, the lateral deviation e y will change with the change of the heading angle deviation e θ . Therefore, at this time, k1 = 0 needs to be set to reduce the weight of the lateral deviation e y , thereby reducing the influence of the lateral deviation e y on the system.

[0069] The operation error e work is the deviation of the implement from the predetermined target (such as depth, width, or application rate) during the operation, in millimeters (mm) or percentage. e work is of multiple types, including respectively referring to different types of operation errors such as operation depth, operation width, etc. In actual operations, e work in the multi-objective optimization function J is the average value of multiple types of operation errors. Its calculation methods include: in tillage operations, calculating the difference between the actual tillage depth and the set tillage depth; in seeding operations, calculating the differences between the actual seeding depth and width and the set seeding depth and width; in fertilization or spraying operations, calculating the deviation between the actual application rate and the predetermined application rate. Specifically:

[0070]

[0071] e workn = |h actual - h target |

[0072] where: the actual operation amount is h actual , and the target operation amount is h target .

[0073] The vibration excitation F vib is divided into the vibration excitation of the tractor and the vibration excitation of the implement The vibration accelerations of the tractor and the implement are measured separately by an acceleration or vibration sensor, representing the intensity of the system vibration, with the unit of meters per second squared (m / s 2 ), F vib and The formula is as follows:

[0074]

[0075] Where: is the vertical acceleration of the tractor; is the pitch angular velocity of the tractor body; is the roll angular velocity of the tractor body; is the vertical acceleration of the implement; is the pitch angular velocity of the implement; is the roll angular velocity of the implement.

[0076] In the multi-objective optimization function, the vibration excitation term F vib in the form of the sum of squares F vib 2 can effectively reduce the vibration influence caused by uneven ground and operation force fluctuation during the optimization process, ensure the stability and comfort of the tractor and implement operation, reduce the loss of mechanical equipment and extend the service life.

[0077] In the optimization of the calculation results of the multi-objective optimization function, the system adopts a particle swarm-genetic (PSO-GA) hybrid optimization algorithm, which combines the global optimization characteristics of the particle swarm algorithm (PSO) and the local fine optimization characteristics of the genetic algorithm (GA). It can realize the rapid screening and convergence of the output data by the control system after the calculation of the multi-objective optimization function, and further optimize the control parameters such as the steering angle, vehicle speed and hydraulic parameters, so as to achieve the optimal balance among various objectives.

[0078] The PSO-GA algorithm is the combined algorithm of the PSO algorithm and the GA algorithm. Specifically, first use the particle swarm algorithm (PSO algorithm) to make the data converge quickly to obtain an effective data set, and then use the genetic algorithm (GA algorithm) for comparison, correction and feedback. In the case of the combination of the two algorithms, the credibility and accuracy of the data can be improved, and further the accuracy in the calculation can be improved. The multi-objective optimization function J can be minimized through the cyclic iteration method of the PSO-GA algorithm, and the control parameters can be adjusted according to the real-time feedback of the system to achieve the optimal comprehensive performance.

[0079] Particle Swarm Optimization (PSO) finds the optimal solution by simulating the movement of a particle swarm, adjusts the tractor control parameters, and optimizes the comprehensive balance of trajectory tracking accuracy, operation error, and vibration excitation. The particles in the particle swarm algorithm refer to the data sets for which the results need to be optimized. This algorithm can exclude invalid or less-referential data and achieve the rapid acquisition and collation of available data. The movement mode of the particles is as follows:

[0080] v i (t + 1) = ω·v i (t) + c1·r1·(pbest i -x i (t)) + c2·r2·(gbest - x i (t))

[0081] x i (t + 1) = x i (t) + v i (t + 1)

[0082] Where: v i (t) is the velocity of the particle, representing the rate of change of the particle at the current moment; pbest i is the historical optimal position of particle i, and gbest is the optimal position of the entire particle swarm; c1 and c2 are acceleration constants, controlling the moving speed of the particle towards the individual optimal and global optimal positions; ω is the inertia weight, controlling the influence of the previous velocity of the particle; r1 and r2 are random factors, increasing the randomness and exploratory nature of the algorithm; x i (t) is the position of the particle, representing the position of the particle in the control parameter space (such as steering angle, vehicle speed, etc.). In this system, the particle swarm rapidly approaches the optimal control parameters (such as steering angle, vehicle speed, hydraulic pressure) within the solution space through a two-way learning mechanism of individual optimal and group optimal, meeting the real-time requirements of farmland operations. In this system, the PSO algorithm provides rapid data convergence for the system and can also guide the selection of the optimization weight target during the operation of the system.

[0083] The Genetic Algorithm (GA) simulates the genetic mechanism of organisms and finds the optimal solution through selection, crossover, and mutation operations to ensure the optimization effect under complex constraint conditions. The basic steps include selection, crossover, and mutation. The Gradient Descent Method calculates the gradient of the multi-objective optimization function with respect to the control parameters and gradually adjusts the parameter values to quickly find the local or global minimum. Its basic formula is as follows:

[0084] Selection process:

[0085] Crossover process:

[0086] Mutation process: x mut= x + Δx

[0087] Where: P select is the selection probability, representing the probability that a certain solution is selected to enter the next generation, which is determined according to the fitness f(x) of the solution; the crossover process generates offspring by mixing the genes of two parents, x1 and x2 are two parent solutions, and x new is the offspring solution after crossover; the mutation process randomly changes the solution in the solution space, x is the current solution, and Δx is the change amount.

[0088] In this system, the PSO algorithm may fall into a local optimal solution. Therefore, a genetic algorithm (GA) is introduced to form a PSO-GA combined algorithm to perform local refinement adjustment on the candidate control parameters initially obtained by the PSO algorithm, thereby further improving the optimization quality of the control parameters of the driverless tractor and implement system. After the improvement of the genetic algorithm, a set of optimal control parameters that can not only meet the requirements of path tracking and operation accuracy but also effectively suppress vibration excitation is finally obtained. This algorithm ensures that the system can quickly respond and correct when facing complex farmland environments and ground mutation loads, so as to achieve the continuous and efficient operation of the driverless tractor and implement in high-precision agricultural operations.

[0089] The PSO-GA algorithm adopted by the present invention can not only search for the optimal solution globally and avoid falling into a local optimum, but also further improve the adaptive performance and response speed of the entire control system through the local adjustment ability of the genetic algorithm. The central control unit CPU realizes the dynamic adjustment of the steering, vehicle speed, and hydraulic parameters by obtaining the data of each sensor in real time and continuously updating the calculation result of the multi-objective optimization function J using the above algorithm, ensuring that the system reaches the best balance between high-precision operation and vibration suppression. The combination of the PSO-GA hybrid optimization strategy and PID control adopted by the present invention enables the driverless tractor to achieve efficient, stable, and precise collaborative operations in various complex environments.

[0090] During the operation and optimization of the actuator, the optimized parameters output by the central control unit CPU are transmitted to the actuator, which includes an electric or hydraulic steering system, a throttle or vehicle speed control system, a implement hydraulic control system, and a vibration suppression device. The electric or hydraulic steering system adjusts the driving direction of the tractor according to the optimized steering angle; the throttle or vehicle speed control system adjusts the engine output to keep the tractor at the most suitable operating speed; the implement hydraulic control system adjusts the tillage depth and operation width according to the control instruction to ensure the operation accuracy; the vibration suppression device effectively reduces the vibration caused by ground mutation loads by adjusting the damping and stiffness. Each actuator works together to ensure that the entire mechanical system maintains high precision and high stability during the actual operation process.

[0091] During the real-time feedback optimization process, the system establishes a closed-loop control structure, and the status data of the actuator and the operation feedback are transmitted back to the central control unit CPU in real time. The central control unit CPU continuously updates the environmental model and error data, and uses the PID control algorithm to recalculate the parameters according to the newly feedback error signal to achieve dynamic adjustment of the steering, speed, and hydraulic systems. The specific implementation is as follows: The control system uses the PID control algorithm to adjust the driving direction and speed of the tractor in real time during the process of optimizing the path tracking accuracy. The proportional term, integral term, and derivative term of the PID controller are used to adjust the tractor's heading and vehicle speed to ensure that the tractor can quickly respond to the path error and make adjustments according to the changing trend of the error. The control system uses the PID control algorithm to adjust the operation depth and width of the implement during the process of optimizing the operation accuracy. The operation control system adjusts the operation depth and width of the implement through the proportional term, integral term, and derivative term of the PID controller according to the operation error to ensure the stability of the operation accuracy.

[0092] The present invention can build a system model based on MATLAB / Simulink and verify it in combination with a hardware-in-the-loop (HIL) simulation test platform. The entire system is verified and debugged through HIL simulation and a field test platform to ensure that the theoretical model is consistent with the actual application.

[0093] Combined with Figures 1-4 , the method for path tracking and implement cooperative control of the driverless tractor of the present invention is as follows:

[0094] Step1: The path tracking and implement cooperative control system of the driverless tractor monitors the ground conditions and operation parameters in real time through sensors installed on the tractor and the operation implement. The sensors include GNSS / RTK positioning devices, implement operation sensors, acceleration or vibration sensors, lidar, vision sensors, and inertial measurement units. The functions of the sensors are to collect information such as path error, operation depth, operation width, and ground unevenness in real time and transmit the information to the calculation and control system for subsequent processing.

[0095] Step2: The central control unit CPU in the calculation and control system performs data fusion and preprocessing on the data collected by the sensors. The raw data collected by different sensors are integrated through a data fusion algorithm to obtain an environmental model under the target operation. All the data are included in the model. Then, through data preprocessing, the data in the environmental model are respectively extracted as: a digital model of the environmental terrain, a digital model of vibration excitation extracted by combining the vibration spectrum, a digital model of operation state uniformity analysis generated by combining the operation amount monitoring, and a digital model of unit attitude analysis generated by combining the inertial measurement unit and the position digital model, as Figure 2As shown. Subsequently, the central control unit CPU obtains a comprehensive evaluation of the operating environment based on the above-mentioned multiple digital models, and obtains a comprehensive evaluation result. The comprehensive evaluation result includes information such as the type and intensity of the ground mutation load, and whether the operating depth and width meet the predetermined goals. Based on this, path planning and adjustment of operating parameters are carried out. Specifically, data analysis results are obtained through operations on the trajectory tracking error, operating error, and ground vibration excitation, and the output is the weight coefficients ω1, ω2, ω3 within a reasonable range.

[0096] Step3: According to the weight coefficients ω1, ω2, ω3, the central control unit CPU optimizes the operation path tracking of the tractor and the operation accuracy of the implement. The specific implementation is as follows: By changing the relative values of the weight coefficients ω1, ω2, ω3, the optimization priorities of the trajectory error, operating error, and vibration excitation in different operating scenarios are controlled. The specific selection methods include: In high-precision path tracking tasks, increase the weight of ω1 and reduce the weights of ω2 and ω3; In fine operation tasks (such as precision seeding and fertilization), increase the weight of ω2 and appropriately adjust ω1 and ω3; In complex terrains, increase the weight of ω3 to preferentially reduce the impact of vibration on the operation accuracy and stability. In actual operations, the change of the weight coefficients is real-time, and the system will adjust in real-time according to the range set in advance during the operation.

[0097] Step4: After obtaining the appropriate weight coefficients ω1, ω2, ω3, the central control unit CPU outputs the multi-objective optimization function J and completes the algorithm optimization process. The specific implementation is as follows: The Particle Swarm Optimization - Genetic Algorithm (PSO - GA) hybrid optimization algorithm is used to solve the multi-objective optimization among path tracking, operation accuracy, and vibration control. The PSO - GA algorithm combines the global search ability of particle swarm optimization and the local fine adjustment ability of genetic algorithm to achieve collaborative optimization among multiple objectives. Through particle swarm optimization, the optimal combination of control parameters (such as steering angle, vehicle speed, and hydraulic parameters, etc.) is searched globally, and then the genetic algorithm's crossover and mutation processes are used to finely adjust these parameters to prevent the system from falling into local optima. Through this algorithm, the system can optimize and adjust the operation parameters such as the steering angle, vehicle speed, and hydraulic parameters of the implement of the tractor, so as to achieve the optimization of the comprehensive performance under different operating environments, such as Figure 3 shown.

[0098] Step 5: To further ensure that the system can achieve the optimal control effect under different operating environments, the central control unit CPU further introduces a vibration suppression mechanism. The specific implementation is as follows: When the vibration caused by the sudden ground load affects the path tracking and operation accuracy and stability of the tractor, the system monitors the vibration acceleration in real time and uses the PID control algorithm to adjust the working state of the vibration suppression device. According to the real-time feedback of the vibration intensity, the system adjusts the damping and stiffness of the shock absorber to ensure that the system still maintains high-efficiency operation performance under vibration impact.

[0099] Step 6: The central control unit CPU uses the PID control algorithm to control the actuator after optimization calculation. The specific implementation is that the control system uses the PID control algorithm to adjust the driving direction and speed of the tractor in real time during the process of optimizing the path tracking accuracy. The heading and vehicle speed of the tractor are adjusted through the proportional term, integral term, and differential term of the PID controller to ensure that the tractor can quickly respond to the path error and adjust according to the change trend of the error. When the control system uses the PID control algorithm to optimize the operation accuracy, the operation control system adjusts the operation depth and width of the implement through the proportional term, integral term, and differential term of the PID controller according to the operation error to ensure the stability of the operation accuracy.

[0100] Step 7: The calculation and control system continuously adjusts and optimizes the operation parameters through the real-time feedback mechanism to ensure that the tractor and the implement can achieve efficient and stable cooperative operation under complex ground conditions. The specific implementation is as follows: During the actual operation of the tractor and the implement, the data collected by the sensor is pre-controlled through the feedforward control unit of the central control unit CPU after being calculated by the system to give the execution target, calculate the change curvature of the actuator variable, and then obtain the theoretical change amount of the actuator, and then output the execution instruction to the actuator. After the actuator works, the actual change amount of the actuator is obtained. The actual change amount and the theoretical change amount are input into the central control unit CPU at the same time to calculate the ratio, and the output is the real-time error. When the real-time error is not 1, at this time the actual change amount and the theoretical change amount are different, and feedback optimization is required. At this time, feedback calculation is performed in the error feedback control unit, and the feedback result is sent to the central control unit CPU to regenerate and output to the actuator for cyclic feedback adjustment; when the real-time error is 1, at this time the actual change amount and the theoretical change amount are the same, and this execution is completed, as Figure 4 shown. Through this method, the system can dynamically optimize the operation accuracy, path tracking accuracy, and vibration control under the influence of sudden ground loads, and finally maximize the operation efficiency and stability.

[0101] The method further includes the following steps: continuously monitor the dynamic states of the tractor and the working implement, and obtain real-time error values through sensor data (such as path deviation, working depth, vibration acceleration, etc.); adjust control parameters such as steering angle, vehicle speed, implement depth, etc. according to the real-time error values to ensure that the system can adapt to environmental changes in real time, optimize the working accuracy and control stability; adopt an adaptive adjustment strategy to automatically optimize the weight coefficients ω1, ω2, ω3 according to changes in working conditions to meet different task requirements.

[0102] The method is applicable to the optimization control of different working types in agricultural production, including but not limited to operations such as precision seeding, precision fertilization, precision tillage, field spraying, etc., and can dynamically adjust the optimization objectives according to changes in working requirements.

[0103] The method can be integrated into the agricultural machinery automation control system as the core part of the system to achieve full-process automation control, and provide real-time feedback and adjustment through a remote monitoring system.

[0104] The method can, under uneven or complex terrains, dynamically adjust control parameters by measuring real-time path tracking errors, working errors, and vibration acceleration, ensure that the system can continuously optimize path accuracy, working accuracy, and vibration control during the execution process, and finally optimize the multi-objective optimization function J to achieve the best working effect.

[0105] The method of the present invention can flexibly adjust the optimization strategy according to different working requirements to ensure optimal control in complex terrains or special working tasks. This flexible control method makes the collaborative operation of the tractor and the implement more efficient, with higher working quality, and effectively controls the maintenance cost and energy consumption of the system. In addition, the vibration suppression system introduced in the present invention can respond in real time to the vibration impact caused by sudden ground loads, greatly improving the stability and working accuracy of the system.

[0106] In practical applications, the control system will be closely linked with the driving control system of the tractor, the working system of the implement, and the vibration suppression system. By optimizing the coordination of path tracking, working accuracy, and vibration control, ensure the efficient operation of the system in complex environments. This system can be monitored and adjusted in real time through a central control unit. The operator can obtain working status information through a remote control interface and make necessary adjustments to the system. In addition, the system can also automatically select a suitable optimization strategy according to different working requirements. For example, in high-precision working tasks, the system will give priority to optimizing working accuracy and path tracking accuracy; in complex terrains or environments where vibration needs to be suppressed, the system will increase the weight of vibration control to achieve the optimal comprehensive performance.

[0107] In summary, the present invention provides a new control system and method, which improves the stability, accuracy and efficiency of the operation of an autonomous tractor.

[0108] The above are only embodiments of the present invention, and do not impose any form of limitation on the present invention. The present invention can also have other forms of embodiments according to the above structure and function, which will not be listed one by one. Therefore, any person skilled in the relevant art, without departing from the scope of the technical solution of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. An unmanned tractor operation path tracking and implement collaborative control system, characterized in that, It includes a sensor system, a computing and control system, an actuator, and a power supply system; The sensor system includes a GNSS / RTK positioning device, a lidar, a vision sensor, an inertial measurement unit (IMU), an acceleration or vibration sensor, and a implement operation sensor; The computing and control system includes a central control unit CPU and a wireless communication module; The actuator includes an electric or hydraulic steering system, a throttle or vehicle speed control system, a implement hydraulic control system, and a vibration suppression system; The power supply system includes a tractor-mounted generator, a battery, and an independent power supply module UPS; Each sensor in the sensor system collects in real time the pose information data, heading, speed, ground environment characteristics, and operation completion degree information data of the tractor, and transmits the information data to the central control unit CPU through the wireless communication module in the computing and control system; The central control unit CPU first fuses and then preprocesses the raw data from each sensor, then constructs a multi-objective optimization function, and then uses a particle swarm-genetic hybrid optimization algorithm to optimize the calculation results of the multi-objective optimization function. The central control unit CPU outputs execution instructions to the electric or hydraulic steering system, throttle or vehicle speed control system, implement hydraulic control system, and vibration suppression system in the actuator according to the optimized data; The tractor-mounted generator and battery in the power supply system supply power to the actuator, and the independent power supply module UPS supplies power to the sensor system and the computing and control system.

2. The unmanned tractor operation path tracking and implement collaborative control system according to claim 1, characterized in that The implement operation sensor includes a depth sensor and a torque sensor; The vibration suppression system is specifically an active suspension control module and an adjustable damping shock absorber.

3. The unmanned tractor operation path tracking and implement cooperative control system according to claim 1, characterized in that The GNSS / RTK positioning device, lidar, and vision sensor are installed on the tractor head or roof, and the inertial measurement unit is installed on the tractor chassis, and multiple acceleration or vibration sensors are respectively installed on the tractor chassis and the implement.

4. The unmanned tractor operation path tracking and implement collaborative control system according to claim 1, characterized in that, The multi-objective optimization function is: J = ω1·e path 2 + ω2·e work 2 + ω3·F vib 2 Where: J is a multi-objective optimization function representing the comprehensive performance of the tractor and implement system; ω1, ω2, and ω3 are all weight coefficients, corresponding to the relative importance of the adjustment trajectory error, operation error, and vibration excitation in the multi-objective optimization function; e path is the trajectory error, representing the deviation between the actual driving path of the tractor and the predetermined path; e work is the operation error, representing the deviation of the implement relative to the target standard (such as depth, width, etc.) during operation; F vib is the vibration excitation, representing the vibration intensity caused by the uneven ground or operation force of the tractor and implement; Where: ω1 + ω2 + ω3 = 1 and ω1, ω2, ω3 ∈ [0, 1] 5. The unmanned tractor operation path tracking and implement cooperative control system according to claim 4, characterized in that, By changing the relative values of the weight coefficients ω1, ω2, ω3, the optimization priorities of the trajectory error, operation error, and vibration excitation in different operation scenarios are controlled. Among them, the trajectory error e path is obtained by calculating the lateral deviation e y of the current position of the tractor from the predetermined path or the heading angle deviation e θ Specifically, it is as follows: Operation error work It is the deviation of the machine from the predetermined target during operation, in millimeters (mm) or percentage. Refers to various types of operational errors, e work is the average value of various types of operation errors, calculated using the following formula: e workn = |h actual - h target | where: the actual workload is h actual , the target workload is h target ; Vibration excitation F vib Divided into the vibration excitation of the tractor and the vibration excitation of the implement The unit is meters per second squared (m / s 2 ), F vib and The formula is as follows: Wherein: is the vertical acceleration of the tractor; is the pitching angular velocity of the tractor body; is the roll angular velocity of the tractor body; is the vertical acceleration of the implement; is the pitching angular velocity of the implement; is the roll angular velocity of the implement.

6. The unmanned tractor operation path tracking and implement collaborative control system according to claim 5, wherein, The weights k1 = 0 or 1, k2 = 0 or 1, but the two are not both 0 or 1 at the same time; ω1 + ω2 + ω3 = 1 and ω1 ∈ [0.35, 0.45], ω2 ∈ [0.35, 0.45], ω3 ∈ [0.1, 0.3].

7. A method for operating path tracking and implement cooperative control of an autonomous tractor, characterized in that, Applied to an unmanned tractor operation path tracking and implement collaborative control system according to any one of claims 1-6, the control method includes: Step1: The unmanned tractor operation path tracking and implement collaborative control system monitors the ground conditions and operation parameters in real time through multiple sensors installed on the tractor and the operation implement. The sensors collect in real time the pose information data, heading, speed, ground environment characteristics, and operation completion degree, and transmit the information to the computing and control system for subsequent processing; Step 2: The central control unit CPU in the computing and control system performs data fusion and preprocessing on the data collected by the multiple sensors, and obtains corresponding digital models. Then, the central control unit CPU obtains a comprehensive evaluation of the operating environment based on the above digital models, and obtains a comprehensive evaluation result. Based on this, path planning and adjustment of operating parameters are carried out. Specifically, data analysis results are obtained through operations on trajectory tracking errors, operating errors, and ground vibration excitations, and the output is weight coefficients ω1, ω2, ω3 within a reasonable range; Step 3: According to the weight coefficients ω1, ω2, ω3, the central control unit CPU optimizes the operation path tracking of the tractor and the operation accuracy of the implement. The specific implementation is: by changing the relative values of the weight coefficients ω1, ω2, ω3 to control the optimization priorities of trajectory errors, operating errors, and vibration excitations in different operating scenarios; Step 4: After obtaining the appropriate weight coefficients ω1, ω2, ω3, the central control unit CPU outputs the multi-objective optimization function J and completes the algorithm optimization process using the particle swarm-genetic hybrid optimization algorithm; Step 5: When the vibration caused by sudden ground loads affects the path tracking and operation accuracy and stability of the tractor, the central control unit CPU adjusts the working state of the vibration suppression system by real-time monitoring of vibration acceleration using the PID control algorithm. The central control unit CPU adjusts the damping and stiffness of the shock absorber according to the real-time feedback of the vibration intensity to ensure that the system still maintains high-efficiency operation performance under vibration shocks; Step 6: The central control unit CPU controls the actuator after optimization calculation through the PID control algorithm; Step 7: The computing and control system continuously adjusts and optimizes the operating parameters through a real-time feedback mechanism to ensure that the tractor and the implement can achieve efficient and stable collaborative operation under complex ground conditions. Specifically: during the actual operation of the tractor and the implement, the data collected by the sensors is calculated by the system and given an execution target, and then pre-control is achieved through the feed-forward control unit of the central control unit CPU. The change curvature of the actuator variable is calculated, and then the theoretical change amount of the actuator is obtained. Then, an execution instruction is output to the actuator. After the actuator works, the actual change amount of the actuator is obtained. The actual change amount and the theoretical change amount are simultaneously input into the central control unit CPU for ratio calculation, and the output is the real-time error. When the real-time error is not 1, at this time the actual change amount and the theoretical change amount are different, and feedback optimization is required. At this time, feedback calculation is performed in the error feedback control unit, and the feedback result is sent to the central control unit CPU to regenerate and output to the actuator for cyclic feedback adjustment; when the real-time error is 1, at this time the actual change amount and the theoretical change amount are the same, and this execution is completed.

8. The method for operating path tracking and implement cooperative control of an autonomous tractor according to claim 7, wherein, In Step 2, the original data collected by different sensors are integrated through a data fusion algorithm to obtain an environmental model for the target operation. The model contains all the calculated data. Then, through data preprocessing, the data in the environmental model are respectively extracted as: a digital model of the environmental terrain, a digital model of vibration excitation extracted by combining the vibration spectrum, a digital model of the analysis of the uniformity of the operation state generated by combining the operation volume monitoring, and a digital model of the analysis of the machine set attitude generated by combining the inertial measurement unit and the position digital model. After that, the central control unit CPU obtains a comprehensive evaluation of the operation environment based on the above multiple digital models and obtains a comprehensive evaluation result. The comprehensive evaluation result includes the type and intensity of the ground sudden load, and whether the operation depth and width meet the predetermined target information.

9. The method for unmanned tractor operation path tracking and implement cooperative control according to claim 7, characterized in that, In Step 3, the specific selection methods of the weight coefficients ω1, ω2, and ω3 include: in the high-precision path tracking task, increase the weight of ω1 and reduce the weights of ω2 and ω3; in the fine operation tasks (such as precision seeding and fertilization), increase the weight of ω2 and appropriately adjust ω1 and ω3; in complex terrains, increase the weight of ω3 to preferentially reduce the impact of vibration on the operation accuracy and stability. In actual operations, the change of the weight coefficients is real-time, and the system will adjust them in real-time during the operation according to the pre-set range.

10. The method for operating path tracking and implement cooperative control of an autonomous tractor according to claim 7, characterized in that, The specific implementation in Step 4 is as follows: The particle swarm-genetic hybrid optimization algorithm is used to solve the multi-objective optimization among path tracking, operation accuracy, and vibration control. The PSO-GA algorithm combines the global search ability of the particle swarm optimization and the local fine adjustment ability of the genetic algorithm to achieve the collaborative optimization among multiple objectives. The particle swarm optimization is used to search for the optimal combination of control parameters globally, and then the genetic algorithm's crossover and mutation processes are used to finely adjust these parameters to prevent the system from falling into a local optimum. Through this algorithm, the system can optimize and adjust the steering angle, vehicle speed of the tractor, and the hydraulic parameter operation parameters of the implement, so as to achieve the optimization of the comprehensive performance under different operation environments.

Images