Hydraulic tillage depth control method, device and system based on tillage depth prediction results

Abstract

A hydraulic tillage depth control method, device, and system based on tillage depth prediction results belongs to the field of intelligent agricultural machinery technology, particularly to adaptive adjustment of tillage depth. This method addresses the problem that existing hydraulic tillage depth control systems, which mostly rely on fixed control parameters and cannot dynamically adjust to real-time changes such as tractor body posture and terrain, limit their performance and effectiveness in practical applications. The method includes obtaining real-time, collected tillage depth data; and using a control module to adjust control instructions based on the deviation between the real tillage depth data and a given tillage depth, thereby achieving adaptive adjustment of the given tillage depth. The hydraulic tillage depth control method, device, and system based on tillage depth prediction results are suitable for predicting tillage depth and dynamically adjusting according to the predicted results, achieving precise control of tillage depth.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent agricultural machinery, and in particular to adaptive adjustment of tillage depth. Background Art

[0002] In modern agriculture, tractors are the primary tillage machines, and their tillage depth and uniformity have a significant impact on crop growth and yield. However, traditional tractor tillage depth control relies heavily on manual adjustment, a method with numerous drawbacks. First, manual adjustment requires the operator to manually adjust the tillage depth handle based on experience, which not only increases labor intensity but can also lead to unstable and uneven tillage depth. Second, manual adjustment often makes it difficult to achieve the ideal tillage depth, thus compromising tillage quality and efficiency.

[0003] To address these issues, those skilled in the art have conducted extensive research and exploration. Hydraulic tillage depth control technology has garnered significant attention due to its advantages, including fast response and high control precision. However, existing hydraulic tillage depth control systems often rely on fixed control parameters and are unable to dynamically adjust to real-time changes in information, such as the tractor's posture and terrain. This limits their performance and effectiveness in practical applications. Summary of the Invention

[0004] The present invention proposes a hydraulic tillage depth control method, device and system based on tillage depth prediction results, which solves the problem that most existing hydraulic tillage depth control systems rely on fixed control parameters and cannot be dynamically adjusted according to real-time changing information such as tractor body posture and terrain, thereby limiting their performance and effectiveness in practical applications.

[0005] The hydraulic tillage depth control method based on tillage depth prediction results of the present invention comprises the following steps:

[0006] Step S1: Acquire real-time collected tractor tillage data; the tractor tillage data includes tractor lower drawbar angle data, tractor body posture data, and terrain data; pre-process the tractor tillage data to obtain multi-dimensional time series data;

[0007] Step S2: Input the multidimensional time series data into the tillage depth prediction model to obtain the tillage depth prediction result;

[0008] Step S3: Using the predicted tillage depth as a given tillage depth, a control module is used to calculate a control instruction; the control instruction is used to adjust the angle of the tillage implement and the lifting and lowering of the suspension system, thereby performing the tillage operation according to the given tillage depth;

[0009] Step S4: Acquire the real tillage depth data collected in real time; use the control module to adjust the control instructions according to the deviation between the real tillage depth data and the given tillage depth, and complete the adaptive adjustment of the given tillage depth.

[0010] Furthermore, a preferred embodiment is provided, wherein the tillage depth prediction model is a BOA-TCN-Transformer-LSTM model;

[0011] The BOA-TCN-Transformer-LSTM model is obtained by automatically adjusting the parameters of the TCN-Transformer-LSTM model using the Bayesian optimization algorithm and searching for the optimal hyperparameter configuration.

[0012] Furthermore, a preferred embodiment is provided, wherein the control module includes a PID control module integrating an RL strategy network and a BP neural network;

[0013] The PID control module that integrates the RL strategy network and the BP neural network is a two-layer control architecture, including a bottom control layer and an upper optimization layer;

[0014] The bottom control layer is a PID controller based on a BP neural network, which is used to calculate the control quantity of the hydraulic system in real time and output control instructions to the hydraulic system's actuators to ensure rapid response to tillage depth deviations.

[0015] The upper layer is a reinforcement learning (RL) strategy network, which is used to analyze the plowing depth error of the hydraulic system, dynamically optimize the PID controller parameters, and adjust the BP neural network weights to improve global adaptability.

[0016] The present invention also proposes a hydraulic tillage depth control device based on the tillage depth prediction result, which includes the following modules:

[0017] Module S1: Acquire real-time collected tractor tillage data; the tractor tillage data includes tractor lower drawbar angle data, tractor body posture data, and terrain data; pre-process the tractor tillage data to obtain multi-dimensional time series data;

[0018] Module S2: Input multidimensional time series data into the tillage depth prediction model to obtain the tillage depth prediction results;

[0019] Module S3: Using the predicted tillage depth as a given tillage depth, the control module calculates control instructions. The control instructions are used to adjust the angle of the tillage implement and the lifting and lowering of the suspension system to perform tillage operations according to the given tillage depth.

[0020] Module S4: Acquire the real tillage depth data collected in real time; use the control module to adjust the control instructions according to the deviation between the real tillage depth data and the given tillage depth to complete the adaptive adjustment of the given tillage depth.

[0021] The present invention also proposes a hydraulic tillage depth control system based on tillage depth prediction results, the system comprising a collection device, a hydraulic system and the hydraulic tillage depth control device based on tillage depth prediction results;

[0022] The collecting device is used to collect tractor tillage data and actual tillage depth data in real time;

[0023] The hydraulic tillage depth control device based on the tillage depth prediction result is used to output a control instruction based on the real-time collected tractor tillage data and the actual tillage depth data;

[0024] The hydraulic system is used to adjust the angle of the tillage implement and the lifting and lowering of the suspension system according to the control instructions, so as to perform tillage operations according to a given tillage depth.

[0025] Furthermore, a preferred embodiment is provided, wherein the system further comprises: a mode switching module, an alarm prompt module and an operation interface module;

[0026] The mode switching module is used to switch between manual mode and automatic mode; in manual mode, the driver manually adjusts the tillage depth; in automatic mode, the hydraulic tillage depth control device based on the tillage depth prediction results completes the adaptive adjustment of the given tillage depth;

[0027] The alarm prompt module is used to issue an alarm message when an abnormality occurs during the tractor's tillage operation, reminding the driver to take corresponding measures;

[0028] The operation interface module is used for human-computer interaction with the driver, including displaying the current tillage depth, displaying alarm information, and setting the target value of the tillage depth.

[0029] The present invention also provides a tractor, comprising any one of the hydraulic tillage depth control systems based on the tillage depth prediction results described above.

[0030] The present invention also proposes a computer device, comprising: a processor and a memory, wherein the memory is used to store executable instructions of the processor, and the processor is configured to execute any one of the above-mentioned hydraulic tillage depth control methods based on tillage depth prediction results by executing the executable instructions.

[0031] The present invention also proposes a computer storage medium, in which a computer program is stored. When the computer program is run, any one of the above-mentioned hydraulic tillage depth control methods based on tillage depth prediction results is executed.

[0032] The present invention also proposes a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of any one of the above-mentioned hydraulic tillage depth control methods based on tillage depth prediction results.

[0033] The present invention has the following beneficial effects:

[0034] 1. The hydraulic tillage depth control method based on tillage depth prediction results described in this invention collects multi-dimensional information such as the tractor's lower drawbar angle, tractor body posture, and terrain, and uses advanced deep learning algorithms to construct a high-precision tillage depth prediction model. The tillage depth prediction model is a BOA-TCN-Transformer-LSTM model. The TCN-Transformer-LSTM model is a multi-layered hybrid architecture that combines the advantages of a temporal convolutional network (TCN), a Transformer encoder, and a long short-term memory network (LSTM). It is designed to efficiently process multi-dimensional time series data and improve prediction performance. The core concept of the model is to extract local temporal features of the input data using a TCN (Temporal Convolutional Network), capture global dependencies using a Transformer encoder, and finally use an LSTM (Long Short-Term Memory) to predict the depth of the tillage depth. Multi-scale feature fusion is performed to generate prediction results. The model's key hyperparameters and model structure are crucial for optimizing the performance of the TCN-Transformer-LSTM model. Fine-tuning these parameters is often the key to achieving optimal model performance. Given the tedious and inefficient manual adjustment of these parameters, the Bayesian optimization algorithm (BOA) was innovatively introduced to intelligently search for the optimal parameter configuration.

[0035] 2. The hydraulic tillage depth control method based on the tillage depth prediction results described in the present invention adopts a PID control module that integrates the RL strategy network and the BP neural network to receive and analyze the output results of the tillage depth prediction model, calculate the parameters required for adjusting the flow rate, pressure, etc., generate precise control instructions, and send them to the hydraulic system; the hydraulic system immediately responds by fine-tuning key components such as the flow valve, pressure valve, and electro-hydraulic proportional reversing valve to accurately control the flow and pressure of the hydraulic oil and drive the suspension system to rise and fall to achieve the preset tillage depth; during the tillage process, the tillage depth changes are monitored in real time, and once a deviation is found, a secondary adjustment is immediately made to ensure that the tillage depth is accurate; as the operation progresses, the tillage depth prediction model continuously updates the prediction results, and the PID control module that integrates the RL strategy network and the BP neural network continuously issues control instructions accordingly to achieve dynamic adjustment and optimization of the tillage depth.

[0036] 3. The hydraulic tillage depth control method based on tillage depth prediction results described in the present invention integrates tillage depth prediction and hydraulic control to form a closed-loop intelligent control system, overcomes the shortcomings of traditional tillage depth control technology, and realizes accurate and automatic control of tillage depth.

[0037] 4. The hydraulic tillage depth control system based on tillage depth prediction results described in the present invention provides a user-friendly operation interface, allowing the driver to easily set the tillage depth target value, monitor the current tillage depth, and receive system alarm information.

[0038] 5. The hydraulic tillage depth control system based on tillage depth prediction results described in the present invention supports switching between manual mode and automatic mode to adapt to different working environments and needs.

[0039] The hydraulic tillage depth control method, device and system based on the tillage depth prediction results described in the present invention are suitable for predicting the tillage depth by real-time monitoring and analysis of information such as the tractor lower drawbar angle, tractor body posture and terrain, and dynamically adjusting the control parameters of the hydraulic system according to the prediction results to achieve precise control of the tillage depth. Its application scenarios cover multi-dimensional complex tillage environments and are suitable for multi-scale slope terrains. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0041] Figure 1 A schematic flow chart of a hydraulic tillage depth control method based on tillage depth prediction results in one embodiment of the present invention;

[0042] Figure 2 A schematic diagram of tractor farming data in one embodiment of the present invention;

[0043] Figure 3 This is a schematic diagram of the structure of the BOA-TCN-Transformer-LSTM model in one embodiment of the present invention;

[0044] Figure 4This is a simulation diagram of a PID controller based on a BP neural network in one embodiment of the present invention. In the figure, zeros(s) represents the zero point of the system, that is, the value of s when the numerator of the transfer function is zero. The zero point has a significant impact on the dynamic response of the system, and they can affect the stability and response speed of the system. Poles(s) represents the poles of the system, that is, the value of s when the denominator of the transfer function is zero. The poles determine the stability and dynamic characteristics of the system, such as oscillation and damping ratio. The distribution of the pole positions on the complex plane directly affects the stability and response characteristics of the system. 1 / z: In a discrete-time system, z is a complex frequency domain variable. 1 / z represents the delay of one sampling period. This module is used to simulate the transmission delay of the signal in the system and is a common delay link in discrete-time systems. BPPID represents a PID controller based on a BP neural network.

[0045] Figure 5 Schematic diagram of simulation results of a PID controller based on a BP neural network in one embodiment of the present invention;

[0046] Figure 6 This is a schematic diagram of the structure of a PID controller based on a BP neural network in one embodiment of the present invention;

[0047] Figure 7 This is a structural diagram of a PID control module that integrates an RL strategy network and a BP neural network in one embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to make the technical solutions and advantages of the present invention more clearly described, the specific embodiments of the present invention will be further described in detail and completely in conjunction with the accompanying drawings. The various embodiments described below are only part of the preferred embodiments of the present invention, rather than all implementation plans; the various embodiments described below are intended to explain the present invention and cannot be understood as limiting the present invention; the reasonable combination of the technical features defined in the various embodiments of the present invention, as well as all other implementation plans obtained by ordinary technicians in this field based on the embodiments of the present invention without making creative work, all fall within the scope of protection of the present invention.

[0049] In a first embodiment, a hydraulic tillage depth control method based on a tillage depth prediction result is provided, the method comprising the following steps:

[0050] Step S1: Acquire real-time collected tractor tillage data; the tractor tillage data includes tractor lower drawbar angle data, tractor body posture data, and terrain data; pre-process the tractor tillage data to obtain multi-dimensional time series data;

[0051] Step S2: Input the multidimensional time series data into the tillage depth prediction model to obtain the tillage depth prediction result;

[0052] Step S3: Using the predicted tillage depth as a given tillage depth, a control module is used to calculate a control instruction; the control instruction is used to adjust the angle of the tillage implement and the lifting and lowering of the suspension system, thereby performing the tillage operation according to the given tillage depth;

[0053] Step S4: Acquire the real tillage depth data collected in real time; use the control module to adjust the control instructions according to the deviation between the real tillage depth data and the given tillage depth, and complete the adaptive adjustment of the given tillage depth.

[0054] In this embodiment, the tractor farming data is collected in real time using sensors and drones.

[0055] In this embodiment, the control module uses PID control or fuzzy control to calculate the control instructions.

[0056] In this embodiment, the terrain data is DEM data, which includes slope information at different scales.

[0057] DEM data, or Digital Elevation Model, is a raster dataset that represents surface elevations through an ordered array of numerical values. Its core function is to construct a continuous terrain surface using discretized elevation points (e.g., intervals of 3 meters or 1 meter), reflecting the topographical characteristics of a region. In this implementation, after data collection, the acquired DEM data was precisely processed to extract slope information for the study area at five different scales.

[0058] Five different scales: flight altitude of 25 meters, image resolution of 0.68 cm; flight altitude of 30 meters, image resolution of 0.82 cm; flight altitude of 35 meters, image resolution of 0.96 cm; flight altitude of 40 meters, image resolution of 1.12 cm; flight altitude of 45 meters, image resolution of 1.23 cm.

[0059] In this embodiment, the tractor body posture data includes the horizontal tilt angle and pitch angle of the tractor body during deep plowing operation.

[0060] In this embodiment, the tractor lower pull rod angle data is the lifting angle of the pull rod of the tractor connected to the hydraulic reversible plow relative to the horizontal plane.

[0061] In this embodiment, the actual tillage depth data can be collected and obtained using the "A system and method for online monitoring of tillage depth for a tractor-type tillage machine" described in the patent document with publication number "CN119289888A" and name "A system and method for online monitoring of tillage depth for a tractor-type tillage machine".

[0062] In addition, in one embodiment, in step S3, the angle of the tillage implement is adjusted:

[0063] Based on the control strategy, the displacement of the hydraulic cylinder, the flow valve (of the hydraulic system) and the pressure valve are adjusted to optimize the (tillage) angle of the tillage implement.

[0064] In addition, in one embodiment, in step S3, the lifting and lowering of the suspension system is adjusted:

[0065] Electro-hydraulic multi-objective optimization control is used to adjust the lifting and lowering of the suspension system:

[0066] A control module (such as PID control) is used to adjust the electro-hydraulic proportional reversing valve, thereby adjusting the flow of hydraulic oil to achieve adjustment of the lifting and lowering of the suspension system.

[0067] In the second embodiment, the tillage depth prediction model is a BOA-TCN-Transformer-LSTM model;

[0068] The BOA-TCN-Transformer-LSTM model is obtained by automatically adjusting the parameters of the TCN-Transformer-LSTM model using the Bayesian optimization algorithm and searching for the optimal hyperparameter configuration.

[0069] In this embodiment, the Bayesian optimization algorithm is called “Bayesian Optimization Algorithm” in English, and is abbreviated as “BOA”.

[0070] In this implementation, the Bayesian Optimization Algorithm (BOA) is used to intelligently search for the optimal hyperparameter configuration (automatically adjust hyperparameters) of the TCN-Transformer-LSTM model to improve prediction performance.

[0071] Deep learning models (such as BOA-TCN-Transformer-LSTM) typically contain a large number of hyperparameters (such as learning rate, convolution kernel size of TCN, number of attention heads of Transformer, hidden layer dimension of LSTM, etc.). Manually adjusting these parameters is time-consuming and inefficient.

[0072] The Bayesian optimization algorithm automatically searches for the optimal hyperparameter combination for model performance (such as minimizing prediction error) through mathematical modeling and iterative sampling, avoiding the inefficiency of exhaustive search (such as grid search) or random search.

[0073] Bayesian optimization process:

[0074] Surrogate Model: Use Gaussian Process or Random Forest to model the relationship between hyperparameters and model performance.

[0075] Acquisition Function: Based on the "potential benefits" predicted by the surrogate model, select the next set of hyperparameters to be evaluated (such as expected improvement EI, upper confidence bound UCB).

[0076] Iterative optimization: Repeat "sampling hyperparameters → training model → evaluating performance → updating proxy model" until the preset number of iterations is reached or convergence occurs.

[0077] Advantages of Bayesian optimization:

[0078] Efficiency: Through surrogate models and acquisition functions, it prioritizes exploring hyperparameter areas with better performance and reduces invalid attempts.

[0079] Adapt to complex objective functions: Applicable to nonlinear and non-convex complex relationships between model performance and hyperparameters.

[0080] Automation: No manual intervention is required, making it suitable for large-scale hyperparameter tuning.

[0081] Bayesian optimization does not modify the model structure, but only adjusts hyperparameters. Through the optimized parameters, the feature extraction capabilities of the TCN, Transformer, and LSTM modules are maximized and coordinated, improving the final prediction accuracy.

[0082] In summary, Bayesian optimization is an automated parameter tuning tool that, through mathematical modeling and intelligent sampling, finds the optimal hyperparameter configuration for the TCN-Transformer-LSTM combination model in deep-plow prediction tasks. This method significantly reduces the cost of manual parameter tuning while improving the model's generalization and prediction accuracy.

[0083] The optimized difference parameters are shown in the following table:

[0084]

[0085] In this implementation, the BOA-TCN-Transformer-LSTM model combines the advantages of temporal convolutional networks (TCNs), Transformer encoders, and long short-term memory networks (LSTMs) to process multidimensional time series data.

[0086] The BOA-TCN-Transformer-LSTM model relies on TCN (Temporal Convolutional Network) to extract local temporal features of input data, uses the Transformer encoder to capture global dependencies, and finally uses LSTM (Long Short-Term Memory) to perform multi-scale feature fusion and generate deep learning prediction results.

[0087] In this embodiment, the TCN-Transformer-LSTM model includes a TCN module, a Transformer encoder, and an LSTM module;

[0088] The TCN module, Transformer encoder, and LSTM module are connected in series.

[0089] The TCN module is used to extract local temporal features of multidimensional time series data;

[0090] The Transformer encoder is used to capture the global dependencies of multidimensional time series data;

[0091] The LSTM module is used to perform multi-scale feature fusion to obtain the cultivation depth prediction result.

[0092] In this implementation, the TCN module, Temporal Convolutional Network, is:

[0093] The TCN module consists of multiple stacked temporal convolution residual blocks (Temporal Block);

[0094] Each residual block extracts local temporal patterns through causal convolution and dilated convolution to enhance the perception of short-term dynamic changes.

[0095] The output of this residual block is added to the output of the previous residual block and used as the input of the next residual block.

[0096] Each residual block consists of two repeated series structures; each series structure consists of a dilated causal convolution layer, a weight normalization layer, a ReLU layer, and a Dropout layer in series.

[0097] In this implementation, the Transformer encoder includes a multi-head attention layer (i.e., a multi-head self-attention mechanism) and a feedforward layer.

[0098] Through the multi-head self-attention mechanism, it is possible to model long-term dependencies in time series and improve the expressiveness of global features.

[0099] In this implementation, the LSTM module (Long Short-Term Memory) is:

[0100] Time series information is further integrated to capture long-term temporal dependencies and generate the final tillage depth prediction results.

[0101] In this embodiment, the LSTM module includes an LSTM layer and a fully connected layer; the fully connected layer is used to map the hidden state output by the LSTM layer to the target monitoring space to complete the final regression task.

[0102] In this implementation, the training method of the BOA-TCN-Transformer-LSTM model is as follows:

[0103] The dataset is divided into a training set and a test set with a ratio of 4:1. The training set is used for training, and the test set is used to verify the model.

[0104] In this implementation, the BOA-TCN-Transformer-LSTM model uses the Huber loss function to fine-tune the output prediction value.

[0105] Huber loss function:

[0106]

[0107] In the third embodiment, the control module includes a PID controller based on a BP neural network:

[0108] The PID controller based on the BP neural network includes a BP neural network and a PID controller;

[0109] The PID controller is used to perform closed-loop feedback control on the hydraulic system, calculate the control quantity of the hydraulic system in real time, and output control instructions to the actuator of the hydraulic system;

[0110] The BP neural network is used to take the system state (i.e. the state of the hydraulic system) as input and the system performance index (i.e. the performance index of the hydraulic system) as the optimization target, and adjust the neural network weights and bias through the back propagation algorithm to adaptively optimize the parameters of the PID controller.

[0111] In this embodiment, BP neural network (Back Propagation Neural Network) is:

[0112] "BP" stands for "Back Propagation," referring to a multi-layer feedforward neural network trained using the backpropagation algorithm. Its core structure consists of an input layer, hidden layers, and an output layer, minimizing prediction errors by iteratively adjusting weights and biases.

[0113] In this embodiment, the PID controller is used to perform closed-loop feedback control on the controlled object (ie, the hydraulic system) to adjust the flow rate and pressure of the hydraulic oil in the hydraulic system.

[0114] In this embodiment, the PID controller based on the BP neural network can be referred to as the BP-PID control system. It is a control strategy that combines the back propagation (BP) neural network with the PID controller. It dynamically optimizes the PID parameters (proportional, integral, and differential coefficients) through the self-learning ability of the neural network, so that the control algorithm has both the real-time performance of PID and the adaptability of the neural network, solving the problems of fixed parameters and poor adaptability of traditional PID control, as well as insufficient stability of pure neural network control.

[0115] In this embodiment, the system performance indicators include overshoot, adjustment time and steady-state error.

[0116] In this embodiment, the parameters of the PID controller include proportional, integral and differential coefficients.

[0117] The three parameters of PID control, namely proportional, integral and differential, have the characteristics of mutual coordination and mutual restriction.

[0118] In this embodiment, by utilizing the mapping capability of the BP neural network to nonlinear functions, the optimal solutions of various nonlinear combinations of the three parameters in the PID controller can be obtained through self-learning.

[0119] In this embodiment, the BP neural network dynamically optimizes the parameters of the PID controller according to the system state and the learning algorithm, where:

[0120] (1) System status:

[0121] Refers to the real-time feedback variables of the BP neural network input layer, including: error (e) and error change rate (Δe)

[0122] Error (e): The deviation between the reference input r (given tillage depth) and the system output y (true tillage depth data), i.e., e = r - y;

[0123] Error change rate (Δe): The change in error between adjacent moments, usually expressed as Δe(k) = e(k) - e(k-1)

[0124] These state variables form the neural network's input feature vector, reflecting the system's current dynamic characteristics. For example, when a large error occurs, the network prioritizes adjusting the proportional term Kp. When the error persists (with a pronounced integral characteristic), it prioritizes optimizing the Ki parameter.

[0125] (2) Learning algorithm:

[0126] Refers to the backpropagation algorithm that drives neural network parameter optimization. Its operating mechanism is as follows:

[0127] Forward propagation: input layer → hidden layer → output layer, generating PID parameter adjustment (ΔKp, ΔKi, ΔKd);

[0128] Error back propagation: Based on the output layer error function (e.g., E = ½(ry)²), the gradient is calculated layer by layer using the chain rule.

[0129] Weight update: Gradient descent method is used to correct network weights and biases. The update formula is:

[0130]

[0131] Where η is the learning rate, It is calculated by back propagation of the error.

[0132] (3) The process of BP neural network dynamically optimizing PID parameters according to system status and learning algorithm includes:

[0133] Initialize PID parameters and neural network weights → Collect system output y and reference input r → Calculate PID controller output → Adjust parameters through neural network → Iterate and update until the system is stable.

[0134] During the implementation process, the input layer of the BP neural network receives the system state variables (error and error change rate), the hidden layer processes the signal through the activation function, and the output layer generates the adjustment amount of the PID parameters; through the error back propagation algorithm, the neural network continuously optimizes the weights and biases, thereby realizing dynamic adjustment of the PID parameters.

[0135] The combination of back propagation (BP) neural network and PID controller enables traditional PID to break through the limitations of linear combination and effectively cope with complex working conditions such as time-varying and nonlinear conditions.

[0136] In this embodiment, the advantages of the PID controller based on the BP neural network compared with the traditional PID control module (or traditional PID controller) are shown in the following table:

[0137]

[0138] It should be noted that in the prior art, real-time tillage depth control is generally achieved by combining real-time tillage depth data collection with traditional PID control. In this embodiment, real-time tillage depth control is achieved by combining a tillage depth prediction model with a PID controller based on a BP neural network.

[0139] Compared to the traditional approach of combining real-time tillage depth data collection with traditional PID control, the approach of combining a tillage depth prediction model with a PID controller based on a BP neural network achieves dynamic adaptive decision-making through deep learning. This approach integrates multi-source data to predict tillage depth trends and proactively optimizes control parameters (such as hydraulic oil pressure). Dynamic adjustments are made based on real-time changes in tractor body posture and terrain, significantly reducing hysteresis and mechanical losses. Its advantages are:

[0140] (1) Adapts to complex environments (such as alternating soft and hard soils) without manual parameter adjustment;

[0141] (2) Automatically optimize strategies during long-term operations to improve energy efficiency, and ultimately achieve a farming process with higher precision and lower energy consumption through intelligent decision-making.

[0142] (3) The combination of the tillage depth prediction model and the PID controller based on the BP neural network has a certain degree of robustness and fault tolerance, and can resist data noise and interference to a certain extent. This means that even if the real tillage depth data collected in real time has certain errors or fluctuations, the neural network model can still output relatively accurate results. The traditional combination of "real-time tillage depth data collection and traditional PID control" relies entirely on the real-time data collection. If there are errors or fluctuations in the data, it may lead to a decline in control effect.

[0143] In this embodiment, if Figure 5 The figure shows the simulation results of the PID controller based on the BP neural network. The controller output is the output of the PID controller based on the BP neural network. It can be seen from the figure that:

[0144] The large fluctuations in the controller output during the initial phase demonstrate its high flexibility. It can quickly adjust its output based on the system's real-time status and requirements, effectively controlling system behavior. The system output, however, is more influenced by the controller output and is relatively less flexible, varying more in response to changes in the controller output. During the stabilization phase, the controller output gradually stabilizes, its fluctuations significantly decreasing, and ultimately stabilizing within a narrow range. This stability helps maintain the system output near the desired stable value, avoiding unnecessary oscillation or overshoot. The stability of the controller output significantly influences the stability of the system output and is a key factor in ensuring stable system operation.

[0145] In the fourth embodiment, the control module includes a PID control module integrating an RL strategy network and a BP neural network;

[0146] The PID control module that integrates the RL strategy network and the BP neural network is a two-layer control architecture, including a bottom control layer and an upper optimization layer;

[0147] The bottom control layer is a PID controller based on a BP neural network, which is used to calculate the control quantity of the hydraulic system in real time and output control instructions to the hydraulic system's actuators to ensure rapid response to tillage depth deviations.

[0148] The upper layer is a reinforcement learning (RL) strategy network, which is used to analyze the plowing depth error of the hydraulic system, dynamically optimize the PID controller parameters, and adjust the BP neural network weights to improve global adaptability.

[0149] In this embodiment, the control variable of the hydraulic system is, for example, the displacement or flow of the hydraulic cylinder.

[0150] Furthermore, in one embodiment, the PID controller based on the BP neural network as the bottom control layer includes a BP neural network and a PID controller;

[0151] The PID controller is used to perform closed-loop feedback control on the hydraulic system, calculate the control quantity of the hydraulic system in real time, and output control instructions to the actuator of the hydraulic system;

[0152] The BP neural network is used to take the system state (i.e. the state of the hydraulic system) as input and the system performance index (i.e. the performance index of the hydraulic system) as the optimization target, and adjust the neural network weights and bias through the back propagation algorithm to adaptively optimize the parameters of the PID controller.

[0153] Furthermore, in one embodiment, the reinforcement learning RL strategy network (or RL algorithm) as the upper optimization layer adopts an Actor-Critic structure and a multi-objective optimization mechanism to enable the control strategy to maintain stability while exploring optimal parameters.

[0154] Multi-objective optimization mechanism: Design a reward function that takes into account tillage depth tracking error, response speed, and hydraulic system energy consumption.

[0155] In this implementation, a hierarchical intelligent control framework integrating reinforcement learning (RL strategy network) and PID controller based on BP neural network is proposed to address the nonlinearity, time-varying and insufficient anti-interference capabilities of the hydraulic (control) system for tillage depth in farmland. A two-layer optimization structure is used to achieve adaptive adjustment of tillage depth to meet the precise control requirements in complex farmland environments.

[0156] It should be noted that traditional PID control relies on manual parameter adjustment in tillage depth adjustment, which makes it difficult to adapt to changes in terrain; and although the BP neural network PID has the ability to self-tune parameters, it is prone to fall into local optimality under complex interference.

[0157] In this implementation, in order to solve the above problems, a two-layer control architecture is constructed:

[0158] The bottom layer is a PID controller based on a BP neural network, which outputs the control quantity (control instruction) to the hydraulic actuator (hydraulic system) in real time; the upper layer introduces an RL strategy network to optimize the PID controller parameters or adjust the neural network weights online by analyzing the system status (such as pressure, flow, and error signal).

[0159] In this embodiment, a two-layer control architecture is adopted to achieve high-precision and adaptive control of tillage depth in farmland operations, adapt to changes in different terrain and landforms, and improve the intelligence level of agricultural machinery operations.

[0160] In addition, in one embodiment, in step S3, the lifting and lowering of the suspension system is adjusted:

[0161] Electro-hydraulic multi-objective optimization control is used to adjust the lifting and lowering of the suspension system:

[0162] The PID controller based on BP neural network is used to adjust the electro-hydraulic proportional reversing valve, thereby adjusting the hydraulic oil flow to achieve the adjustment of the suspension system's lifting and lowering.

[0163] In addition, in one embodiment, a field experiment was conducted to verify the accuracy and robustness of the method described in this embodiment (BOA-TCN-Transformer-LSTM model outputs tillage depth prediction results), which can meet the requirements of finishing depth consistency.

[0164] Field experiment details:

[0165] The experimental plot covers a total area of ​​178 mu, with relatively flat terrain and typical black soil.

[0166] The agricultural machinery used in the experiment included the Deutz-Fahr CD2004 (G4) wheeled tractor and the Lovol 1LFTT-4455R hydraulic tipping plow.

[0167] The tractor posture, GPS position and pull rod angle information are recorded every 5 seconds and matched with the terrain elevation data (collected by RTK drone) as the input of the model (ploughing depth prediction model), which outputs the deep plowing depth prediction result.

[0168] To evaluate the model performance, four comparison models were set up, including linear regression, K-nearest neighbor regression, random forest regression, and support vector regression.

[0169] The comparison results are as follows:

[0170] Traditional model: The coefficient of determination (R²) ranged from 0.53 to 0.65; the root mean square error (RMSE) ranged from 3.71 to 4.02 cm; and the mean absolute error (MAE) ranged from 2.85 to 3.24 cm.

[0171] BOA-TCN-Transformer-LSTM model: R² of 0.765, RMSE of 1.978 cm, and MAE of 1.489 cm.

[0172] In summary, compared with the traditional model, the proposed BOA-TCN-Transformer-LSTM model achieves the highest monitoring accuracy.

[0173] In a fifth embodiment, a hydraulic tillage depth control device based on tillage depth prediction results is provided, the device comprising the following modules:

[0174] Module S1: Acquire real-time collected tractor tillage data; the tractor tillage data includes tractor lower drawbar angle data, tractor body posture data, and terrain data; pre-process the tractor tillage data to obtain multi-dimensional time series data;

[0175] Module S2: Input multidimensional time series data into the tillage depth prediction model to obtain the tillage depth prediction results;

[0176] Module S3: Using the predicted tillage depth as a given tillage depth, the control module calculates control instructions. The control instructions are used to adjust the angle of the tillage implement and the lifting and lowering of the suspension system to perform tillage operations according to the given tillage depth.

[0177] Module S4: Acquire the real tillage depth data collected in real time; use the control module to adjust the control instructions according to the deviation between the real tillage depth data and the given tillage depth to complete the adaptive adjustment of the given tillage depth.

[0178] In a sixth embodiment, a hydraulic tillage depth control system based on a tillage depth prediction result is provided, the system comprising a data acquisition device, a hydraulic system, and the hydraulic tillage depth control device based on a tillage depth prediction result according to the fourth embodiment;

[0179] The collecting device is used to collect tractor tillage data and actual tillage depth data in real time;

[0180] The hydraulic tillage depth control device based on the tillage depth prediction result is used to output a control instruction based on the real-time collected tractor tillage data and the actual tillage depth data;

[0181] The hydraulic system is used to adjust the angle of the tillage implement and the lifting and lowering of the suspension system according to the control instructions, so as to perform tillage operations according to a given tillage depth.

[0182] In the seventh embodiment, the system further includes: a mode switching module, an alarm prompt module, and an operation interface module;

[0183] The mode switching module is used to switch between manual mode and automatic mode; in manual mode, the driver manually adjusts the tillage depth; in automatic mode, the hydraulic tillage depth control device based on the tillage depth prediction results completes the adaptive adjustment of the given tillage depth;

[0184] The alarm prompt module is used to issue an alarm message when an abnormality occurs during the tractor's tillage operation, reminding the driver to take corresponding measures;

[0185] The operation interface module is used for human-computer interaction with the driver, including displaying the current tillage depth, displaying alarm information, and setting the target value of the tillage depth.

[0186] In this embodiment, the abnormality occurring when the tractor performs the tillage operation includes the tillage depth deviating from a predetermined range.

[0187] Embodiment 8 provides a tractor, comprising any one of the hydraulic tillage depth control systems based on tillage depth prediction results described above.

[0188] Embodiment nine provides a computer device, comprising: a processor and a memory, wherein the memory is used to store executable instructions of the processor, and the processor is configured to execute any one of the above-mentioned hydraulic tillage depth control methods based on tillage depth prediction results by executing the executable instructions.

[0189] In a tenth embodiment, a computer storage medium is provided, wherein a computer program is stored in the storage medium. When the computer program is run, any one of the above-mentioned hydraulic tillage depth control methods based on tillage depth prediction results is executed.

[0190] In an eleventh embodiment, a computer program product is provided, comprising a computer program / instruction, which, when executed by a processor, implements the steps of any one of the above-mentioned methods for hydraulic tillage depth control based on tillage depth prediction results.

[0191] This embodiment provides a computer device or system, the hardware device of this part is a general model and is not shown in the form of a diagram. The system includes a processor and a memory, wherein the processor and the memory can be connected through a bus or other means. The memory is a non-transient computer-readable storage medium that can be used to store non-transient software programs, non-transient computer executable programs and modules, and corresponding program instructions / modules. The processor executes various functional applications and data processing of the processor by running the non-transient software programs, instructions and modules stored in the memory, so as to realize the data space entity resolution data quality enhancement method in the above method embodiment.

[0192] The memory may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created by the processor, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory may optionally include a memory remotely located relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, an intranet, a mobile communication network, and combinations thereof.

[0193] One or more modules are stored in the memory. When the processor executes, the method steps in the embodiment are executed. In this way, the purpose of the invention can be achieved through the method, device and process of the present invention. The specific details of the above-mentioned computer equipment can be understood by referring to the corresponding descriptions and effects in the embodiment, and will not be repeated here.

[0194] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware through a computer program. The program can be stored in a computer-readable storage medium, and when executed, the program can include the processes in the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (HDD), or a solid-state drive (SSD). The storage medium can also include a combination of the above-mentioned types of memory.

[0195] The above further describes the technical solution provided by the present invention in detail through several specific embodiments in order to highlight the advantages and benefits of the technical solution provided by the present invention. However, the several specific embodiments described above are not intended to limit the present invention. Any reasonable changes and improvements to the present invention, reasonable combinations of implementation methods and equivalent replacements based on the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A hydraulic tillage depth control method based on tillage depth prediction results, characterized in that: The method comprises the following steps: Step S1: Acquire real-time collected tractor tillage data; the tractor tillage data includes tractor lower drawbar angle data, tractor body posture data, and terrain data; pre-process the tractor tillage data to obtain multi-dimensional time series data; Step S2: Input the multidimensional time series data into the tillage depth prediction model to obtain the tillage depth prediction result; Step S3: Using the predicted tillage depth as a given tillage depth, a control module is used to calculate a control instruction; the control instruction is used to adjust the angle of the tillage implement and the lifting and lowering of the suspension system, thereby performing the tillage operation according to the given tillage depth; Step S4: Acquire real tillage depth data collected in real time; use the control module to adjust the control instructions according to the deviation between the real tillage depth data and the given tillage depth to complete the adaptive adjustment of the given tillage depth; The tillage depth prediction model is a BOA-TCN-Transformer-LSTM model; The BOA-TCN-Transformer-LSTM model is obtained by automatically adjusting the parameters of the TCN-Transformer-LSTM model using the Bayesian optimization algorithm and searching for the optimal hyperparameter configuration; The TCN-Transformer-LSTM model includes a TCN module, a Transformer encoder, and an LSTM module; The TCN module, Transformer encoder, and LSTM module are connected in series. The TCN module is used to extract local temporal features of multidimensional time series data; The Transformer encoder is used to capture the global dependencies of multidimensional time series data; The LSTM module is used to perform multi-scale feature fusion to obtain the cultivation depth prediction result.

2. The hydraulic tillage depth control method based on tillage depth prediction results according to claim 1, characterized in that: The control module includes a PID control module that integrates an RL strategy network and a BP neural network; The PID control module that integrates the RL strategy network and the BP neural network is a two-layer control architecture, including a bottom control layer and an upper optimization layer; The bottom control layer is a PID controller based on a BP neural network, which is used to calculate the control quantity of the hydraulic system in real time and output control instructions to the hydraulic system's actuators to ensure rapid response to tillage depth deviations. The upper layer is a reinforcement learning (RL) strategy network, which is used to analyze the plowing depth error of the hydraulic system, dynamically optimize the PID controller parameters, and adjust the BP neural network weights to improve global adaptability.

3. A hydraulic tillage depth control device based on tillage depth prediction results, characterized in that: The device comprises the following modules: Module S1: Acquire real-time collected tractor tillage data; the tractor tillage data includes tractor lower drawbar angle data, tractor body posture data, and terrain data; pre-process the tractor tillage data to obtain multi-dimensional time series data; Module S2: Input multidimensional time series data into the tillage depth prediction model to obtain the tillage depth prediction results; Module S3: Using the predicted tillage depth as a given tillage depth, the control module calculates control instructions. The control instructions are used to adjust the angle of the tillage implement and the lifting and lowering of the suspension system to perform tillage operations according to the given tillage depth. Module S4: Acquiring real-time tillage depth data; The control module is used to adjust the control instructions according to the deviation between the actual tillage depth data and the given tillage depth, thereby completing the adaptive adjustment of the given tillage depth; The tillage depth prediction model is a BOA-TCN-Transformer-LSTM model; The BOA-TCN-Transformer-LSTM model is obtained by automatically adjusting the parameters of the TCN-Transformer-LSTM model using the Bayesian optimization algorithm and searching for the optimal hyperparameter configuration; The TCN-Transformer-LSTM model includes a TCN module, a Transformer encoder, and an LSTM module; The TCN module, Transformer encoder, and LSTM module are connected in series. The TCN module is used to extract local temporal features of multidimensional time series data; The Transformer encoder is used to capture the global dependencies of multidimensional time series data; The LSTM module is used to perform multi-scale feature fusion to obtain the cultivation depth prediction result.

4. A hydraulic tillage depth control system based on tillage depth prediction results, characterized in that: The system includes a collection device, a hydraulic system, and a hydraulic tillage depth control device based on tillage depth prediction results as described in claim 3; The collecting device is used to collect tractor tillage data and actual tillage depth data in real time; The hydraulic tillage depth control device based on the tillage depth prediction result is used to output a control instruction based on the real-time collected tractor tillage data and the actual tillage depth data; The hydraulic system is used to adjust the angle of the tillage implement and the lifting and lowering of the suspension system according to the control instructions, so as to perform tillage operations according to a given tillage depth.

5. The hydraulic tillage depth control system based on tillage depth prediction results according to claim 4, characterized in that: The system also includes: a mode switching module, an alarm prompt module and an operation interface module; The mode switching module is used to switch between manual mode and automatic mode; in manual mode, the driver manually adjusts the tillage depth; in automatic mode, the hydraulic tillage depth control device based on the tillage depth prediction results completes the adaptive adjustment of the given tillage depth; The alarm prompt module is used to issue an alarm message when an abnormality occurs during the tractor's tillage operation, reminding the driver to take corresponding measures; The operation interface module is used for human-computer interaction with the driver, including displaying the current tillage depth, displaying alarm information, and setting the target value of the tillage depth.

6. A tractor, characterized in that: The tractor includes the hydraulic plowing depth control system based on the plowing depth prediction result according to claim 4 or 5.

7. A computer device comprising: A processor and a memory, characterized in that the memory is used to store executable instructions of the processor, and the processor is configured to execute the hydraulic tillage depth control method based on tillage depth prediction results as described in any one of claims 1 or 2 by executing the executable instructions.

8. A computer storage medium, characterized in that The storage medium stores a computer program, and when the computer program is run, the hydraulic tillage depth control method based on tillage depth prediction results according to any one of claims 1 or 2 is executed.

9. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the steps of the hydraulic tillage depth control method based on the tillage depth prediction result as described in any one of claims 1 or 2 are implemented.

Images