A posture adjustment control system and method for a myriapod robot rice field weeding machine

Abstract

This invention discloses a posture adjustment control system and method for a paddy field weeder that mimics a multi-legged robot, belonging to the field of agricultural machinery technology. The system monitors the machine's tilt angle, wheel speed, and terrain height difference in real time using sensors. After preprocessing the data through Kalman filtering and coordinate transformation, an adjustment amount is calculated using a virtual model, PID, and MPC fusion control algorithm to achieve machine leveling and anti-sideslip. The inertial measurement unit (IMU) detects the machine's tilt angle and angular velocity, while the terrain sensor quantifies the height difference of the driving wheels. The controller generates wheel speed difference or suspension adjustment commands based on an inverse kinematics model to simulate multi-legged adjustment. The system ensures rapid response in the dynamic environment of paddy fields through a closed-loop feedback mechanism, with tilt angle error controlled within 3°. This method improves the adaptability of agricultural machinery, reduces the risk of tipping over, optimizes energy consumption, supports wireless communication for remote monitoring, and is suitable for the automation transformation of wheeled agricultural machinery.

Description

Technical Field

[0001] This invention relates to the field of agricultural machinery technology, specifically to a posture adjustment control system and method for a paddy field weeder that mimics a multi-legged robot. Background Technology

[0002] With the continuous improvement of agricultural mechanization, paddy field weeding has become an indispensable and important part of rice production. At present, most paddy field weeders adopt a wheeled chassis structure, and the drive method is mainly a four-wheel / tracked design driven by electric motors or internal combustion engines.

[0003] In existing technologies, the following technical means are mainly adopted to adapt to the complex terrain of paddy fields: First, a passive suspension system is used, which passively adapts to the undulations of the terrain through mechanical springs or hydraulic dampers; second, a simple differential steering control is used, which utilizes the speed difference between the left and right wheels to achieve basic steering and coarse attitude maintenance; third, a basic tilt sensor is equipped, combined with a PID controller for simple speed adjustment to achieve semi-automatic attitude correction. In addition, many self-propelled weeders on the market rely on a fixed wheelbase and basic drive control, and can only maintain approximate balance through manual intervention or simple sensor assistance.

[0004] However, the limitations of the aforementioned traditional techniques are particularly pronounced in the high water content and muddy, soft environment of paddy fields. The wheels are prone to getting stuck in the mud, causing them to become stuck. When one wheel sinks deeper than the other, the machine body experiences a significant roll angle, making it impossible to maintain an overall level posture. At the same time, traditional control methods have a slow response time and cannot achieve independent and precise adjustment of the torque of all four wheels (such as the four-wheel torque monitoring and anti-slip control of the anti-lock braking system (ABS) in industrial machine tools), which can easily lead to skidding, overturning, or work interruption.

[0005] Existing research and practical experience show that in soft paddy fields, the wheels of heavy machinery can sink by 10-20 cm, causing the machine to tilt at an angle exceeding 10°. This reduces operational efficiency to only about 70% of that in dry land, and poses significant safety hazards, frequently requiring manual rescue or machine shutdown for adjustments. Furthermore, current technology lacks dynamic posture adjustment algorithms similar to those used in multi-legged robots, failing to effectively address the instability caused by uneven wheel height. This results in high energy consumption, poor adaptability, and difficulty in meeting the demands of modern precision agriculture in paddy fields for continuous and unmanned operations. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention proposes a posture adjustment control system and method for a paddy field weeder that mimics a multi-legged robot. The system uses sensors to monitor the machine's tilt angle, wheel speed, and terrain height difference in real time. After preprocessing the data through Kalman filtering and coordinate transformation, an adjustment amount is calculated using a virtual model, PID, and MPC fusion control algorithm to achieve leveling and anti-skid operation. The inertial measurement unit (IMU) detects the tilt angle and angular velocity, while the terrain sensor quantifies the height difference of the driving wheels. The controller generates wheel speed differences or suspension adjustment commands based on an inverse kinematics model to simulate multi-legged adjustment. The system employs a closed-loop feedback mechanism to ensure rapid response in the dynamic environment of paddy fields, effectively solving current technical challenges.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0008] A posture adjustment control system for a paddy field weeding machine that mimics a multi-legged robot, the posture adjustment system comprising:

[0009] The wheeled chassis consists of four drive wheels and a frame, and supports differential drive to adapt to uneven terrain.

[0010] The sensor module consists of an inertial measurement unit (IMU), wheel speed sensors, terrain sensors, and force sensors; the inertial measurement unit is used to acquire the fuselage tilt angle in real time. and angular velocity It provides basic attitude data, including Including roll angle and pitch angle The wheel speed sensor is used to monitor the wheel speed of each drive wheel. The force sensor is used to monitor the ground force of each drive wheel. Together, they identify sideslip trends; the terrain sensor uses a laser rangefinder or ultrasonic sensor to detect the ground clearance difference measured individually for each drive wheel. Quantify the unevenness characteristics of paddy fields;

[0011] The controller module consists of a data processing unit, an algorithm execution unit, and a storage unit. The data processing unit receives data from the sensor module, performs preprocessing such as Kalman filtering, coordinate transformation, and median filtering, and calculates the attitude error. The algorithm execution unit, based on a multi-legged robot-like control method, calculates and outputs the final wheel speed adjustment for each drive wheel. Suspension adjustment range The storage unit is used for adjustment to maintain the aircraft's level and prevent sideslip; it is used to record historical data and supports offline analysis.

[0012] The actuator module includes a motor controller; the motor controller is used to adjust the wheel speed of each drive wheel to achieve differential control for compensation. Includes an active suspension system for use according to the aforementioned Fine-tuning the height of the drive wheels simulates multi-legged adjustment, improving adaptability;

[0013] The data communication module is used for real-time data transmission between the sensor module, controller module, and actuator module to ensure closed-loop synchronization of the system and supports Wi-Fi wireless extension for remote monitoring.

[0014] Furthermore, the controller module has an attitude optimization submodule, which performs real-time analysis of the adjustment process and generates reports, including multiple indicators such as tilt error, sideslip risk, adaptability, and energy efficiency. These analysis indicators are used to evaluate the overall performance of the system and support modular upgrades to adapt to different types of agricultural machinery.

[0015] A method for posture adjustment control of a paddy field weeding machine that mimics a multi-legged robot, the method being applied to the posture adjustment control system of the paddy field weeding machine that mimics a multi-legged robot as described in claims 1-2, includes the following steps:

[0016] Step S1: Data acquisition, acquired in real time through the sensor module. , , , and Data supports multi-sensor fusion to improve accuracy;

[0017] Step S2, attitude calculation: Based on the collected data, Kalman filtering is performed for noise reduction and coordinate transformation preprocessing. The inverse kinematics model is used to calculate the body attitude error and expected adjustment amount, and the calculation is optimized for the uneven paddy field environment.

[0018] Step S3: Control algorithm execution, applying VMC (Variable Mechanism Control), PID (PID) algorithm, and MPC (Multi-legged Robot) to fuzzify the input. and Fuzzy logic is used for auxiliary control to generate and The instructions ensure a rapid response; the VMC is Virtual Model Control, and the MPC is Model Predictive Control.

[0019] Step S4, Execution and Feedback: The motor controller executes wheel speed adjustment commands to achieve differential tilt angle compensation, the active suspension system executes suspension adjustment commands to simulate multi-leg adjustment, and provides real-time feedback to monitor the effect, forming a closed-loop control to maintain stability;

[0020] Step S5: Stability assessment and optimization. Based on the feedback data, assess the attitude stability and adaptively optimize the control parameters to adapt to the dynamic changes in the paddy field.

[0021] Step S6: Generate an attitude adjustment report and upload the adjustment process data to the data storage module for subsequent analysis, optimization, and fault diagnosis.

[0022] Furthermore, the specific method for step S1 is as follows:

[0023] Step S1-1, Sensor Installation and Configuration: The IMU is installed in the center of the fuselage for data acquisition. and Wheel speed sensors are installed on the axle of each drive wheel to monitor speed. Force sensors are installed at the suspension of each drive wheel to monitor... Terrain sensors are installed under or on the side of the chassis to monitor... As a basis for identifying the height difference of each drive wheel on uneven terrain, when At that time, and When it is less than a preset small threshold, that is The slight height difference is negligible; when At that time, identify potential sideslip risks and trigger alarm thresholds;

[0024] Step S1-2, Data Synchronization and Transmission: All data is synchronously transmitted to the controller via the CAN bus data communication module, and timestamp alignment is performed to ensure that the data delay is <10ms. Multi-threaded processing is supported to avoid blocking.

[0025] Step S1-3, Preprocessing: Apply Kalman filtering to remove noise, the formula is:

[0026]

[0027] in, This represents the state estimate updated using the current measurements at time k. It is an updated prior prediction estimate based on all information from the observation data from time 1 to k; This represents the Kalman gain at time k. This represents the measurement value at time k. For the measurement matrix, an additional median filter was added to handle outliers and ensure the reliability of the data in the muddy environment of paddy fields.

[0028] Furthermore, the specific method for step S2 is as follows:

[0029] Step S2-1, Model Establishment: Using an inverse kinematics model, the wheeled chassis is treated as a four-legged virtual leg system, with each drive wheel corresponding to a virtual leg, supporting four-leg configurations to match the chassis;

[0030] Step S2-2, Error Quantification: Directly quantify the height difference of the drive wheels using a formula, and preset the desired horizontal tilt angle. and the desired angular velocity of the fuselage ; Calculate tilt angle error ,in ;Calculation based on geometric model ,in The calculation formula is:

[0031]

[0032] in This refers to the distance between the drive wheels; The effective height difference used for tilt angle calculation is calculated as the average difference in the ground contact height between the left and right drive wheels:

[0033]

[0034] in and The difference in ground contact height between the left and right wheels. and The ground contact height difference of the right wheel;

[0035] Step S2-3, Adjustment Calculation: Calculate the desired wheel speed difference. The calculation formula is:

[0036]

[0037] in It is the acceleration due to gravity. This represents the average wheel speed.

[0038] Introducing virtual leg strength balance , The calculation formula is:

[0039]

[0040] in For fuselage weight; This refers to the number of drive wheels; The dynamic force compensation amount is calculated using the following formula:

[0041]

[0042] in, For the force distribution coefficient, the mud resistance coefficient is also considered. When performing optimization, the calculation formula is as follows:

[0043]

[0044] Step S2-4, Boundary Check: If Marked as extremely uneven, with limitations To prevent overload.

[0045] Furthermore, step S3 is specifically as follows:

[0046] Step S3-1: Application of virtual model control;

[0047] Step S3-2: Combination of PID control;

[0048] Step S3-3: Use fuzzy logic for auxiliary control.

[0049] Furthermore, the virtual model control application method in step S3-1 is as follows:

[0050] Step S3-1-1, parameter setting and initialization, is defined as a virtual spring-damping model. The PID parameters in the virtual spring-damping model are set as follows: , , ,in The parameter settings are adaptively adjusted according to the fuselage weight, and the calculation formula is as follows:

[0051]

[0052] Step S3-1-2: Calculate the virtual force. The calculation formula is as follows:

[0053]

[0054] Among them, the integral term is limited to prevent saturation;

[0055] Step S3-1-3: Convert virtual force into wheel speed adjustment, where virtual force... and The mathematical formula for the relationship is as follows:

[0056]

[0057] in For the diameter of the drive wheel, Maximum torque;

[0058] Step S3-1-4: Calculate the desired wheel speed difference scalar. Distribute to each drive wheel, and adjust the speed of the opposite wheel for sideslip: when When tilting to the right, the wheel speed adjustment amount of the left drive wheel is set. Wheel speed adjustment of the right drive wheel Meanwhile, the system supports multi-drive wheel distribution, with front wheels prioritized to improve steering stability;

[0059] The combination method of step S3-2 and PID control is as follows:

[0060] Step S3-2-1, Input The normalized tilt angle error is obtained by normalization. ,in This is the maximum permissible tilt angle threshold;

[0061] Step S3-2-2: Calculate the output control quantity of the PID controller. The calculation formula is as follows:

[0062]

[0063] in , , , The integral is calculated using the trapezoidal rule:

[0064]

[0065] Step S3-2-3: Integrate VMC and PID, the calculation formula is as follows:

[0066]

[0067] in, For the fusion weights, the The core algorithm uses this fusion to resolve fuselage tilt, ensuring... It converges rapidly to 0 in uneven paddy field environments, with a response time of <0.5s;

[0068] Step S3-2-4, Model Predictive Control Optimization, MPC: For the future step, To predict the number of steps, the state equation is:

[0069]

[0070] in , The first Step and the first The system state vector of the step, the state vector is defined as follows: ; For the system matrix, To control the input matrix, where the matrix ,matrix ; For the first The control input vector for the step;

[0071] Construct the optimization objective function: while satisfying the constraints and Under the premise of minimizing the cost function through a quadratic programming solver. Solve for the optimal control sequence; the objective function is:

[0072]

[0073] in, To optimize the objective function; , These are the state error weight matrix and the control input weight matrix, respectively; the control process is under constraints. The following solution is obtained using a quadratic programming solver.

[0074] Step S3-3, using fuzzy logic for auxiliary control: For uncertainties in paddy fields, the input is fuzzified. and The fuzzy sets are defined as small, medium, and large, and the membership function is a triangular membership function; the fuzzy rule is defined as: when large and Negative, then The speed of the opposite wheel is greatly increased, and the corresponding original wheel speed is increased to 1.5 times the standard value; the output defuzzification adopts the center of gravity method to generate the final adjustment amount.

[0075] Furthermore, the specific methods for execution and feedback described in step S4 are as follows:

[0076] Step S4-1, Execution Instruction Application: Applied through the motor controller The response time is <0.5s, and PWM speed control is supported. If a suspension is installed, the suspension height is adjusted using the following formula:

[0077]

[0078] in, For the first The suspension height adjustment range of each drive wheel; For the first The height difference between the drive wheels; is a scaling factor. Set to 0.8; perform compensation to directly counteract the tilt caused by the height difference of the drive wheels, taking into account actuator delay compensation;

[0079] Step S4-2, Feedback Monitoring: Feedback loop every Iteration, monitoring Error and sideslip speed , of which sideslip speed The calculation formula is:

[0080]

[0081] in, This refers to the sideslip speed; and These are the actual rotational speeds of the left and right drive wheels, respectively; when absolute value If this occurs, the attitude adjustment algorithm will be recalculated; a safety threshold will be set. ,if Execute the emergency braking command, causing all drive wheels to rotate at [speed]. And record system logs;

[0082] Step S4-3, Closed-loop verification: Calculate the feedback error The calculation formula is as follows:

[0083]

[0084] in, and These are photos of the posture before and after adjustment. When the absolute value of the feedback error Increase the number of iterations and support online parameter adjustment to improve robustness; the specific method for increasing the number of iterations is as follows:

[0085] Initial iteration count n=1;

[0086] The number of iterations when the error exceeds the limit ;

[0087] Re-execute steps S2 attitude calculation and S3 control algorithm;

[0088] Simultaneously fine-tune the PID proportional gain online: .

[0089] Furthermore, the specific methods for stability evaluation and optimization described in step S5 are as follows:

[0090] Step S5-1, Stability Assessment: Use the Lyapunov function to assess the stability of the system. The function formula is as follows:

[0091]

[0092] in, Let Lyapunov be the value of the function; the stability condition is that when If the system is stable, it is determined that the energy function is stable. An additional energy function is calculated to monitor energy changes and prevent system oscillations. The energy function formula is:

[0093]

[0094] in, The total kinetic energy of the system; The linear velocity of the fuselage; The moment of inertia of the fuselage;

[0095] Step S5-2, Parameter Optimization: Parameters are optimized using a genetic algorithm, with the target... The system iterated for 50 generations, with a population size of 20, a crossover rate of 0.8, and a mutation rate of 0.1. Based on paddy field test data, the proportional gain was adaptively adjusted using the following formula:

[0096]

[0097] in, For proportional gain; It provides the variance of the height difference data for each round and supports the introduction of machine learning training datasets for joint offline and online optimization.

[0098] Step S5-3, Optimization and Verification: Verification is conducted in a simulated paddy field scenario with slope variations. After optimization, the verification is completed. Root mean square error This ensures that the fuselage remains level even in uneven environments over a long period of time.

[0099] Furthermore, the specific method for generating the attitude adjustment report in step S6 is as follows:

[0100] Step S6-1, Report Content Generation: The report includes tilt error statistics, sideslip rate, efficiency improvement and energy consumption analysis. The tilt error statistics include mean and standard deviation. The data format is JSON or CSV, and it supports time-based visualization charts.

[0101] Step S6-2, Data Upload and Storage: Upload data to cloud storage or local database, encrypt transmission to ensure security; analyze trends, such as long-term mud adaptation, and generate optimization suggestions;

[0102] Step S6-3, Report Application: Supports user interface display and integration into the agricultural machinery management system for fault diagnosis and performance evaluation.

[0103] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects:

[0104] (1) This invention provides a posture adjustment control system and method for a paddy field weeder that mimics a multi-legged robot. By analogy of a wheeled chassis to a multi-legged walking structure, a sensor module collects multi-dimensional data such as body tilt angle, angular velocity, and ground force in real time. The controller module decouples and calculates the wheel speed adjustment and suspension adjustment, driving the actuator module to perform coordinated compensation. This process achieves dynamic balance of the weeder in deep muddy and highly subsidence paddy field environments, significantly improving the machine's ability to maintain horizontal position in complex terrain.

[0105] (2) This invention provides a posture adjustment control system and method for a paddy field weeding machine that mimics a multi-legged robot. It achieves accurate quantification of the micro-topographical undulations of the paddy field and the tendency of the drive wheel to slip. Through force-position hybrid monitoring, it effectively solves the problem that traditional weeding machines are prone to tipping over or getting stuck on wet and slippery water surfaces.

[0106] (3) This invention provides a posture adjustment control system and method for a paddy field weeding machine that mimics a multi-legged robot. It achieves deep coupling between body height adjustment and driving power distribution. By simulating the flexibility of multi-legged legs to finely adjust the height of the drive wheels, it maintains constant chassis passability and operation accuracy in uneven paddy fields, which greatly reduces the damage rate of crops caused by weeding operations.

[0107] (4) The present invention provides a posture adjustment control system and method for a paddy field weeding machine that resembles a multi-legged robot. It realizes efficient closed-loop synchronization of sensor data, control strategy and execution feedback, supports real-time transmission and offline optimization of operation status via Wi-Fi, and improves the intelligent operation and maintenance level of the system. Attached Figure Description

[0108] Figure 1 This is a schematic diagram of the overall architecture of a posture adjustment control system for a paddy field weeding machine that mimics a multi-legged robot, according to the present invention.

[0109] Figure 2 This is a schematic diagram of the overall architecture of a posture adjustment control method for a paddy field weeding machine that mimics a multi-legged robot, according to the present invention. Specific implementation methods

[0110] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0111] A posture adjustment control system for a paddy field weeding machine that mimics a multi-legged robot, such as Figure 1 The attitude adjustment system includes:

[0112] The wheeled chassis consists of four drive wheels and a frame, and supports differential drive to adapt to uneven terrain.

[0113] The sensor module consists of an inertial measurement unit (IMU), wheel speed sensors, terrain sensors, and force sensors; the inertial measurement unit is used to acquire the fuselage tilt angle in real time. and angular velocity It provides basic attitude data, including Including roll angle and pitch angle The wheel speed sensor is used to monitor the wheel speed of each drive wheel. The force sensor is used to monitor the ground force of each drive wheel. Together, they identify sideslip trends; the terrain sensor uses a laser rangefinder or ultrasonic sensor to detect the ground clearance difference measured individually for each drive wheel. Quantify the unevenness characteristics of paddy fields;

[0114] The controller module consists of a data processing unit, an algorithm execution unit, and a storage unit. The data processing unit receives data from the sensor module, performs preprocessing such as Kalman filtering, coordinate transformation, and median filtering, and calculates the attitude error. The algorithm execution unit, based on a multi-legged robot-like control method, calculates and outputs the final wheel speed adjustment for each drive wheel. Suspension adjustment range The storage unit is used for adjustment to maintain the aircraft's level and prevent sideslip; it is used to record historical data and supports offline analysis.

[0115] The actuator module includes a motor controller; the motor controller is used to adjust the wheel speed of each drive wheel to achieve differential control for compensation. Includes an active suspension system for use according to the aforementioned Fine-tuning the height of the drive wheels simulates multi-legged adjustment, improving adaptability;

[0116] The data communication module is used for real-time data transmission between the sensor module, controller module, and actuator module to ensure closed-loop synchronization of the system and supports Wi-Fi wireless extension for remote monitoring.

[0117] Furthermore, the controller module has an attitude optimization submodule, which performs real-time analysis of the adjustment process and generates reports, including multiple indicators such as tilt error, sideslip risk, adaptability, and energy efficiency. These analysis indicators are used to evaluate the overall performance of the system and support modular upgrades to adapt to different types of agricultural machinery.

[0118] A method for posture adjustment control of a paddy field weeding machine that mimics a multi-legged robot, such as... Figure 2 The method is applied to the posture adjustment control system of a paddy field weeding machine that mimics a multi-legged robot, as described in claims 1-2, and includes the following steps:

[0119] Step S1: Data acquisition, acquired in real time through the sensor module. , , , and Data supports multi-sensor fusion to improve accuracy;

[0120] Step S2, attitude calculation: Based on the collected data, Kalman filtering is performed for noise reduction and coordinate transformation preprocessing. The inverse kinematics model is used to calculate the body attitude error and expected adjustment amount, and the calculation is optimized for the uneven paddy field environment.

[0121] Step S3: Control algorithm execution, applying VMC (Variable Mechanism Control), PID (PID) algorithm, and MPC (Multi-legged Robot) to fuzzify the input. and Fuzzy logic is used for auxiliary control to generate and The instructions ensure a rapid response; the VMC is Virtual Model Control, and the MPC is Model Predictive Control.

[0122] Step S4, Execution and Feedback: The motor controller executes wheel speed adjustment commands to achieve differential tilt angle compensation, the active suspension system executes suspension adjustment commands to simulate multi-leg adjustment, and provides real-time feedback to monitor the effect, forming a closed-loop control to maintain stability;

[0123] Step S5: Stability assessment and optimization. Based on the feedback data, assess the attitude stability and adaptively optimize the control parameters to adapt to the dynamic changes in the paddy field.

[0124] Step S6: Generate an attitude adjustment report and upload the adjustment process data to the data storage module for subsequent analysis, optimization, and fault diagnosis.

[0125] Furthermore, the specific method for step S1 is as follows:

[0126] Step S1-1, Sensor Installation and Configuration: The IMU is installed in the center of the fuselage for data acquisition. and Wheel speed sensors are installed on the axle of each drive wheel to monitor speed. Force sensors are installed at the suspension of each drive wheel to monitor... Terrain sensors are installed under or on the side of the chassis to monitor... As a basis for identifying the height difference of each drive wheel on uneven terrain, when At that time, and When it is less than a preset small threshold, that is The slight height difference is negligible; when At that time, identify potential sideslip risks and trigger alarm thresholds;

[0127] Step S1-2, Data Synchronization and Transmission: All data is synchronously transmitted to the controller via the CAN bus data communication module, and timestamp alignment is performed to ensure that the data delay is <10ms. Multi-threaded processing is supported to avoid blocking.

[0128] Step S1-3, Preprocessing: Apply Kalman filtering to remove noise, the formula is:

[0129]

[0130] in, This represents the state estimate updated using the current measurements at time k. It is an updated prior prediction estimate based on all information from the observation data from time 1 to k; This represents the Kalman gain at time k. This represents the measurement value at time k. For the measurement matrix, an additional median filter was added to handle outliers and ensure the reliability of the data in the muddy environment of paddy fields.

[0131] Furthermore, the specific method for step S2 is as follows:

[0132] Step S2-1, Model Establishment: Using an inverse kinematics model, the wheeled chassis is treated as a four-legged virtual leg system, with each drive wheel corresponding to a virtual leg, supporting four-leg configurations to match the chassis;

[0133] Step S2-2, Error Quantification: Directly quantify the height difference of the drive wheels using a formula, and preset the desired horizontal tilt angle. and the desired angular velocity of the fuselage ; Calculate tilt angle error ,in ;Calculation based on geometric model ,in The calculation formula is:

[0134]

[0135] in This refers to the distance between the drive wheels; The effective height difference used for tilt angle calculation is calculated as the average difference in the ground contact height between the left and right drive wheels:

[0136]

[0137] in and The difference in ground contact height between the left and right wheels. and The ground contact height difference of the right wheel;

[0138] Step S2-3, Adjustment Calculation: Calculate the desired wheel speed difference. The calculation formula is:

[0139]

[0140] in It is the acceleration due to gravity. This represents the average wheel speed.

[0141] Introducing virtual leg strength balance , The calculation formula is:

[0142]

[0143] in For fuselage weight; This refers to the number of drive wheels; The dynamic force compensation amount is calculated using the following formula:

[0144]

[0145] in, For the force distribution coefficient, the mud resistance coefficient is also considered. When performing optimization, the calculation formula is as follows:

[0146]

[0147] Step S2-4, Boundary Check: If Marked as extremely uneven, with limitations To prevent overload.

[0148] Furthermore, step S3 is specifically as follows:

[0149] Step S3-1: Application of virtual model control;

[0150] Step S3-2: Combination of PID control;

[0151] Step S3-3: Use fuzzy logic for auxiliary control.

[0152] Furthermore, the virtual model control application method in step S3-1 is as follows:

[0153] Step S3-1-1, parameter setting and initialization, is defined as a virtual spring-damping model. The PID parameters in the virtual spring-damping model are set as follows: , , ,in The parameter settings are adaptively adjusted according to the fuselage weight, and the calculation formula is as follows:

[0154]

[0155] Step S3-1-2: Calculate the virtual force. The calculation formula is as follows:

[0156]

[0157] Among them, the integral term is limited to prevent saturation;

[0158] Step S3-1-3: Convert virtual force into wheel speed adjustment, where virtual force... and The mathematical formula for the relationship is as follows:

[0159]

[0160] in For the diameter of the drive wheel, Maximum torque;

[0161] Step S3-1-4: Calculate the desired wheel speed difference scalar. Distribute to each drive wheel, and adjust the speed of the opposite wheel for sideslip: when When tilting to the right, the wheel speed adjustment amount of the left drive wheel is set. Wheel speed adjustment of the right drive wheel Meanwhile, the system supports multi-drive wheel distribution, with front wheels prioritized to improve steering stability;

[0162] The combination method of step S3-2 and PID control is as follows:

[0163] Step S3-2-1, Input The normalized tilt angle error is obtained by normalization. ,in This is the maximum permissible tilt angle threshold;

[0164] Step S3-2-2: Calculate the output control quantity of the PID controller. The calculation formula is as follows:

[0165]

[0166] in , , , The integral is calculated using the trapezoidal rule:

[0167]

[0168] Step S3-2-3: Integrate VMC and PID, the calculation formula is as follows:

[0169]

[0170] in, For the fusion weights, the The core algorithm uses this fusion to resolve fuselage tilt, ensuring... It converges rapidly to 0 in uneven paddy field environments, with a response time of <0.5s;

[0171] Step S3-2-4, Model Predictive Control Optimization, MPC: For the future step, To predict the number of steps, the state equation is:

[0172]

[0173] in , The first Step and the first The system state vector of the step, the state vector is defined as follows: ; For the system matrix, To control the input matrix, where the matrix ,matrix ; For the first The control input vector for the step;

[0174] Construct the optimization objective function: while satisfying the constraints and Under the premise that, among them , Minimize the cost function using a quadratic programming solver Solve for the optimal control sequence; the objective function is:

[0175]

[0176] in, To optimize the objective function; , These are the state error weight matrix and the control input weight matrix, respectively; the control process is under constraints. The following solution is obtained using a quadratic programming solver.

[0177] Step S3-3, using fuzzy logic for auxiliary control: For uncertainties in paddy fields, the input is fuzzified. and The fuzzy sets are defined as small, medium, and large, and the membership function is a triangular membership function; the fuzzy rule is defined as: when large and Negative, then The speed of the opposite wheel is greatly increased, and the corresponding original wheel speed is increased to 1.5 times the standard value; the output defuzzification adopts the center of gravity method to generate the final adjustment amount.

[0178] Furthermore, the specific methods for execution and feedback described in step S4 are as follows:

[0179] Step S4-1, Execution Instruction Application: Applied through the motor controller The response time is <0.5s, and PWM speed control is supported. If a suspension is installed, the suspension height is adjusted using the following formula:

[0180]

[0181] in, For the first The suspension height adjustment range of each drive wheel; For the first The height difference between the drive wheels; is a scaling factor. Set to 0.8; perform compensation to directly counteract the tilt caused by the height difference of the drive wheels, taking into account actuator delay compensation;

[0182] Step S4-2, Feedback Monitoring: Feedback loop every Iteration, monitoring Error and sideslip speed , of which sideslip speed The calculation formula is:

[0183]

[0184] in, This refers to the sideslip speed; and These are the actual rotational speeds of the left and right drive wheels, respectively; when absolute value If this occurs, the attitude adjustment algorithm will be recalculated; a safety threshold will be set. ,if Execute the emergency braking command, causing all drive wheels to rotate at [speed]. And record system logs;

[0185] Step S4-3, Closed-loop verification: Calculate the feedback error The calculation formula is as follows:

[0186]

[0187] in, and These are photos of the posture before and after adjustment. When the absolute value of the feedback error Increase the number of iterations and support online parameter adjustment to improve robustness; the specific method for increasing the number of iterations is as follows:

[0188] Initial iteration count n=1;

[0189] The number of iterations when the error exceeds the limit (Maximum limit n≤5, to prevent computational overload);

[0190] Re-execute steps S2 attitude calculation and S3 control algorithm;

[0191] Simultaneously fine-tune the PID proportional gain online: .

[0192] Furthermore, the specific methods for stability evaluation and optimization described in step S5 are as follows:

[0193] Step S5-1, Stability Assessment: Use the Lyapunov function to assess the stability of the system. The function formula is as follows:

[0194]

[0195] in, Let Lyapunov be the value of the function; the stability condition is that when If the system is stable, it is determined that the energy function is stable. An additional energy function is calculated to monitor energy changes and prevent system oscillations. The energy function formula is:

[0196]

[0197] in, The total kinetic energy of the system; The linear velocity of the fuselage; The moment of inertia of the fuselage;

[0198] Step S5-2, Parameter Optimization: Parameters are optimized using a genetic algorithm, with the target... The system iterated for 50 generations, with a population size of 20, a crossover rate of 0.8, and a mutation rate of 0.1. Based on paddy field test data, the proportional gain was adaptively adjusted using the following formula:

[0199]

[0200] in, For proportional gain; It provides the variance of the height difference data for each round and supports the introduction of machine learning training datasets for joint offline and online optimization.

[0201] Step S5-3, Optimization and Verification: Verification is conducted in a simulated paddy field scenario with slope variations. After optimization, the verification is completed. Root mean square error This ensures that the fuselage remains level even in uneven environments over a long period of time.

[0202] Furthermore, the specific method for generating the attitude adjustment report in step S6 is as follows:

[0203] Step S6-1, Report Content Generation: The report includes tilt error statistics, sideslip rate, efficiency improvement and energy consumption analysis. The tilt error statistics include mean and standard deviation. The data format is JSON or CSV, and it supports time-based visualization charts.

[0204] Step S6-2, Data Upload and Storage: Upload data to cloud storage or local database, encrypt transmission to ensure security; analyze trends, such as long-term mud adaptation, and generate optimization suggestions;

[0205] Step S6-3, Report Application: Supports user interface display and integration into the agricultural machinery management system for fault diagnosis and performance evaluation.

[0206] Those skilled in the art should understand that, unless otherwise specified, the meanings of the technical and scientific terms used herein are consistent with the general understanding of the relevant technical field. Furthermore, terms defined in general dictionaries should be understood in the context of the technical background in this field and should not be interpreted in an overly idealized or formalistic manner divorced from practical application scenarios.

[0207] The above embodiments have described in detail the main concept, technical solution, and technical effects of the present invention. It should be noted that the above content is merely illustrative and not intended to limit the scope of protection of the present invention. Any equivalent modifications, substitutions, or optimizations based on the present invention without departing from its core principles are within the scope of the present invention.

Claims

1. A paddy field weeding machine posture adjustment control system for a myriapod robot, characterized by, The attitude adjustment system includes: The wheeled chassis consists of four drive wheels and a frame, and supports differential drive to adapt to uneven terrain. The sensor module consists of an inertial measurement unit (IMU), wheel speed sensors, terrain sensors, and force sensors; the inertial measurement unit is used to acquire the fuselage tilt angle in real time. and angular velocity It provides basic attitude data, including Including roll angle and pitch angle The wheel speed sensor is used to monitor the wheel speed of each drive wheel. The force sensor is used to monitor the ground force of each drive wheel. Together, they identify sideslip trends; the terrain sensor uses a laser rangefinder or ultrasonic sensor to detect the ground clearance difference measured individually for each drive wheel. Quantify the unevenness characteristics of paddy fields; The controller module consists of a data processing unit, an algorithm execution unit, and a storage unit. The data processing unit receives data from the sensor module, performs preprocessing such as Kalman filtering, coordinate transformation, and median filtering, and calculates the attitude error. The algorithm execution unit, based on a multi-legged robot-like control method, calculates and outputs the final wheel speed adjustment for each drive wheel. Suspension adjustment range The storage unit is used for adjustment to maintain the aircraft's level and prevent sideslip; it is used to record historical data and supports offline analysis. The actuator module includes a motor controller; the motor controller is used to adjust the wheel speed of each drive wheel to achieve differential control for compensation. Includes an active suspension system for use according to the aforementioned Fine-tuning the height of the drive wheels simulates multi-legged adjustment, improving adaptability; The data communication module is used for real-time data transmission between the sensor module, controller module, and actuator module to ensure closed-loop synchronization of the system and supports Wi-Fi wireless extension for remote monitoring.

2. The posture adjustment control system for a paddy field weeding machine mimicking a multi-legged robot according to claim 1, characterized in that, The controller module has an attitude optimization submodule, which performs real-time analysis of the adjustment process and generates reports, including multiple indicators such as tilt error, sideslip risk, adaptability, and energy efficiency. These analysis indicators are used to evaluate the overall performance of the system and support modular upgrades to adapt to different types of agricultural machinery.

3. A method for posture adjustment control of a paddy field weeding machine that mimics a multi-legged robot, characterized in that, The method is applied to the posture adjustment control system of a paddy field weeding machine that resembles a multi-legged robot, as described in claims 1-2, and includes the following steps: Step S1: Data acquisition, acquired in real time through the sensor module. , , , and Data supports multi-sensor fusion to improve accuracy; Step S2, attitude calculation: Based on the collected data, Kalman filtering is performed for noise reduction and coordinate transformation preprocessing. The inverse kinematics model is used to calculate the body attitude error and expected adjustment amount, and the calculation is optimized for the uneven paddy field environment. Step S3: Control algorithm execution, applying VMC (Variable Mechanism Control), PID (PID) algorithm, and MPC (Multi-legged Robot) to fuzzify the input. and Fuzzy logic is used for auxiliary control to generate and The instructions ensure a rapid response; the VMC is Virtual Model Control, and the MPC is Model Predictive Control. Step S4, Execution and Feedback: The motor controller executes wheel speed adjustment commands to achieve differential tilt angle compensation, the active suspension system executes suspension adjustment commands to simulate multi-leg adjustment, and provides real-time feedback to monitor the effect, forming a closed-loop control to maintain stability; Step S5: Stability assessment and optimization. Based on the feedback data, assess the attitude stability and adaptively optimize the control parameters to adapt to the dynamic changes in the paddy field. Step S6: Generate an attitude adjustment report and upload the adjustment process data to the data storage module for subsequent analysis, optimization, and fault diagnosis.

4. The posture adjustment control method for a paddy field weeding machine mimicking a multi-legged robot according to claim 3, characterized in that, The specific method for step S1 is as follows: Step S1-1, Sensor Installation and Configuration: The IMU is installed in the center of the fuselage for data acquisition. and Wheel speed sensors are installed on the axle of each drive wheel to monitor speed. Force sensors are installed at the suspension of each drive wheel to monitor... Terrain sensors are installed under or on the side of the chassis to monitor... As a basis for identifying the height difference of each drive wheel on uneven terrain, when At that time, and When it is less than a preset small threshold, that is The slight height difference is negligible; when At that time, identify potential sideslip risks and trigger alarm thresholds; Step S1-2, Data Synchronization and Transmission: All data is synchronously transmitted to the controller via the CAN bus data communication module, and timestamp alignment is performed to ensure that the data delay is <10ms. Multi-threaded processing is supported to avoid blocking. Step S1-3, Preprocessing: Apply Kalman filtering to remove noise, the formula is: in, This represents the state estimate updated using the current measurements at time k. It is an updated prior prediction estimate based on all information from the observation data from time 1 to k; This represents the Kalman gain at time k. This represents the measurement value at time k. For the measurement matrix, an additional median filter was added to handle outliers and ensure the reliability of the data in the muddy environment of paddy fields.

5. The posture adjustment control method for a paddy field weeding machine mimicking a multi-legged robot according to claim 3, characterized in that, The specific method for step S2 is as follows: Step S2-1, Model Establishment: Using an inverse kinematics model, the wheeled chassis is treated as a four-legged virtual leg system, with each drive wheel corresponding to a virtual leg, supporting four-leg configurations to match the chassis; Step S2-2, Error Quantification: Directly quantify the height difference of the drive wheels using a formula, and preset the desired horizontal tilt angle. and the desired angular velocity of the fuselage ; Calculate tilt angle error ,in ;Calculation based on geometric model ,in The calculation formula is: in This refers to the distance between the drive wheels; The effective height difference used for tilt angle calculation is calculated as the average difference in the ground contact height between the left and right drive wheels: in and The difference in ground contact height between the left and right wheels. and The ground contact height difference of the right wheel; Step S2-3, Adjustment Calculation: Calculate the desired wheel speed difference. The calculation formula is: in It is the acceleration due to gravity. This represents the average wheel speed. Introducing virtual leg strength balance , The calculation formula is: in For fuselage weight; This refers to the number of drive wheels; The dynamic force compensation amount is calculated using the following formula: in, For the force distribution coefficient, the mud resistance coefficient is also considered. When performing optimization, the calculation formula is as follows: Step S2-4, Boundary Check: If Marked as extremely uneven, with limitations To prevent overload.

6. The posture adjustment control method for a paddy field weeding machine mimicking a multi-legged robot according to claim 3, characterized in that, Step S3 is as follows: Step S3-1: Application of virtual model control; Step S3-2: Combination of PID control; Step S3-3: Use fuzzy logic for auxiliary control.

7. The posture adjustment control method for a paddy field weeding machine mimicking a multi-legged robot according to claim 6, characterized in that, The virtual model control application method in step S3-1 is as follows: Step S3-1-1, parameter setting and initialization, is defined as a virtual spring-damping model. The PID parameters in the virtual spring-damping model are set as follows: , , ,in The parameter settings are adaptively adjusted according to the fuselage weight, and the calculation formula is as follows: Step S3-1-2: Calculate the virtual force. The calculation formula is as follows: Among them, the integral term is limited to prevent saturation; Step S3-1-3: Convert virtual force into wheel speed adjustment, where virtual force... and The mathematical formula for the relationship is as follows: in For the diameter of the drive wheel, Maximum torque; Step S3-1-4: Calculate the desired wheel speed difference scalar. Distribute to each drive wheel, and adjust the speed of the opposite wheel for sideslip: when When tilting to the right, the wheel speed adjustment amount of the left drive wheel is set. Wheel speed adjustment of the right drive wheel Meanwhile, the system supports multi-drive wheel distribution, with front wheels prioritized to improve steering stability; The combination method of step S3-2 and PID control is as follows: Step S3-2-1, Input The normalized tilt angle error is obtained by normalization. ,in This is the maximum permissible tilt angle threshold; Step S3-2-2: Calculate the output control quantity of the PID controller. The calculation formula is as follows: in , , , The integral is calculated using the trapezoidal rule: Step S3-2-3: Integrate VMC and PID, the calculation formula is as follows: in, For the fusion weights, the The core algorithm uses this fusion to resolve fuselage tilt, ensuring... It converges rapidly to 0 in uneven paddy field environments, with a response time of <0.5s; Step S3-2-4, Model Predictive Control Optimization, MPC: For the future step, To predict the number of steps, the state equation is: in , The first Step and the first The system state vector of the step, the state vector is defined as follows: ; For the system matrix, To control the input matrix, where the matrix ,matrix ; For the first The control input vector for the step; Construct the optimization objective function: while satisfying the constraints and Under the premise of minimizing the cost function through a quadratic programming solver. Solve for the optimal control sequence; the objective function is: in, To optimize the objective function; , These are the state error weight matrix and the control input weight matrix, respectively; the control process is under constraints. The following solution is obtained using a quadratic programming solver. Step S3-3, using fuzzy logic for auxiliary control: For uncertainties in paddy fields, the input is fuzzified. and The fuzzy sets are defined as small, medium, and large, and the membership function is a triangular membership function; the fuzzy rule is defined as: when large and Negative, then The speed of the opposite wheel is greatly increased, and the corresponding original wheel speed is increased to 1.5 times the standard value; the output defuzzification adopts the center of gravity method to generate the final adjustment amount.

8. The posture adjustment control method for a paddy field weeding machine mimicking a multi-legged robot according to claim 3, characterized in that, The specific methods for execution and feedback described in step S4 are as follows: Step S4-1, Execution Instruction Application: Applied through the motor controller The response time is <0.5s, and PWM speed control is supported. If a suspension is installed, the suspension height is adjusted using the following formula: in, For the first The suspension height adjustment range of each drive wheel; For the first The height difference between the drive wheels; is a scaling factor. Set to 0.8; perform compensation to directly counteract the tilt caused by the height difference of the drive wheels, taking into account actuator delay compensation; Step S4-2, Feedback Monitoring: Feedback loop every Iteration, monitoring Error and sideslip speed , of which sideslip speed The calculation formula is: in, This refers to the sideslip speed; and These are the actual rotational speeds of the left and right drive wheels, respectively; when absolute value If this occurs, the attitude adjustment algorithm will be recalculated; a safety threshold will be set. ,if Execute the emergency braking command, causing all drive wheels to rotate at [speed]. And record in the system log; Step S4-3, Closed-loop verification: Calculate the feedback error The calculation formula is as follows: in, and These are photos of the posture before and after adjustment. When the absolute value of the feedback error Increase the number of iterations and support online parameter adjustment to improve robustness; the specific method for increasing the number of iterations is as follows: Initial iteration count n=1; The number of iterations when the error exceeds the limit ; Re-execute steps S2 attitude calculation and S3 control algorithm; Simultaneously fine-tune the PID proportional gain online: .

9. The posture adjustment control method for a paddy field weeding machine mimicking a multi-legged robot according to claim 3, characterized in that, The specific methods for stability assessment and optimization described in step S5 are as follows: Step S5-1, Stability Assessment: Use the Lyapunov function to assess the stability of the system. The function formula is as follows: in, Let Lyapunov be the value of the function; the stability condition is that when If the system is stable, it is determined that the energy function is stable. An additional energy function is calculated to monitor energy changes and prevent system oscillations. The energy function formula is: in, The total kinetic energy of the system; The linear velocity of the fuselage; The moment of inertia of the fuselage; Step S5-2, Parameter Optimization: Parameters are optimized using a genetic algorithm, with the target... The system iterated for 50 generations, with a population size of 20, a crossover rate of 0.8, and a mutation rate of 0.

1. Based on paddy field test data, the proportional gain was adaptively adjusted using the following formula: in, For proportional gain; It provides the variance of the height difference data for each round and supports the introduction of machine learning training datasets for joint offline and online optimization. Step S5-3, Optimization and Verification: Verification is conducted in a simulated paddy field scenario with slope variations. After optimization, the verification is completed. Root mean square error This ensures that the fuselage remains level even in uneven environments over a long period of time.

10. The posture adjustment control method for a paddy field weeding machine mimicking a multi-legged robot according to claim 3, characterized in that, The specific method for generating the attitude adjustment report in step S6 is as follows: Step S6-1, Report content generation: The report includes tilt error statistics, sideslip rate, efficiency improvement and energy consumption analysis, where tilt error statistics include mean and standard deviation; The data format is JSON or CSV, and it supports time-varying curve visualization charts; Step S6-2, Data Upload and Storage: Upload data to cloud storage or local database, encrypt transmission to ensure security; analyze trends, such as long-term mud adaptation, and generate optimization suggestions; Step S6-3, Report Application: Supports user interface display and integration into the agricultural machinery management system for fault diagnosis and performance evaluation.

Images