Big Data-Based Agricultural Machinery Tillage Depth Control and Improvement System

Abstract

This invention discloses a big data-based agricultural machinery tillage depth control and improvement system, belonging to the field of agricultural machinery intelligent control technology. It includes a tillage data sensing module, a soil hardness spatial prediction modeling module, and a tillage depth control module. The invention employs a multi-source spatiotemporal fusion-based soil hardness spatial prediction modeling method to predict the soil hardness change trend in the target area ahead of the agricultural machinery and generate a reference operating depth corresponding to the soil hardness. This allows the agricultural machinery to adjust its depth before entering hard soil areas. The invention also employs a tillage depth control method based on soil hardness prediction to analyze the soil hardness change trend ahead of the tillage area in advance and dynamically generate the target operating depth based on this trend. Simultaneously, it combines mechanical load, operating speed, and slip ratio to smoothly adjust and adaptively optimize the depth change process, achieving continuous and stable depth adjustment of the agricultural machinery in different soil regions.

Description

Technical Field

[0001] This invention belongs to the field of agricultural machinery intelligent control technology, specifically referring to an agricultural machinery tillage depth control and improvement system based on big data. Background Technology

[0002] The depth of tillage by agricultural machinery directly affects soil loosening, crop growth, and operational efficiency. However, current agricultural machinery tillage methods largely rely on mechanical depth limits or passive load feedback, making it difficult to detect changes in soil hardness in advance. When encountering locally hard soil, this can easily lead to sudden increases in load, tillage depth deviations, and hydraulic shocks, resulting in insufficient tillage uniformity and operational stability. Furthermore, existing methods do not adequately utilize historical operational data, soil type differences, and real-time operational status, making it difficult to achieve differentiated and adaptive depth adjustment for different soil regions.

[0003] Therefore, there is an urgent need for a big data-based system for improving farmland depth control. This system can predict the trend of soil hardness changes in front of the farmland by combining historical farmland operation data, basic soil information, and real-time operation status of agricultural machinery. It can also dynamically generate operation depth adjustment strategies based on soil hardness in different areas, thereby achieving predictive, adaptive, and continuous stable control of the farmland depth of agricultural machinery. This would reduce mechanical impact and traction load fluctuations, improve the stability of agricultural machinery operations, and enhance the adaptive operation capability in complex farmland environments. Summary of the Invention

[0004] In view of the above situation and to overcome the shortcomings of the prior art, the present invention provides an agricultural machinery tillage depth control and improvement system based on big data, including a tillage data sensing module, a soil hardness spatial prediction modeling module, and a tillage depth control module.

[0005] The cultivated land data sensing module is used to sense cultivated land data, obtain cultivated land operation data, and send the cultivated land operation data to the soil hardness spatial modeling module and the cultivated land depth control module.

[0006] The soil hardness spatial modeling module is used for soil hardness spatial modeling. Based on farmland operation data, it adopts a soil hardness spatial prediction modeling method based on multi-source spatiotemporal fusion to construct farmland spatiotemporal association, perform multi-source sensor fusion modeling, spatial interpolation prediction and depth reference value mapping, and obtain a soil hardness spatial map and depth reference value mapping results covering the target area in front of the agricultural machinery. The soil hardness spatial map and depth reference value mapping results are then sent to the farmland depth control module.

[0007] The cultivated land depth control module is used for cultivated land depth control. Based on cultivated land operation data, soil hardness spatial map and depth reference value mapping results, it adopts a cultivated land depth control method based on soil hardness prediction to generate target operation depth, control depth change smoothing, construct state space, adaptively optimize operation depth and execute depth control, so as to obtain adaptive continuous cultivated land depth control results for different soil regions.

[0008] Furthermore, the aforementioned farmland data sensing specifically involves sampling the location information and mechanical operation parameters of agricultural machinery during the farmland tillage process, and then combining this data with basic soil information to form farmland operation data.

[0009] The location information includes continuous latitude and longitude coordinates, timestamp information, and the agricultural machinery's travel trajectory;

[0010] The mechanical operating parameters include mechanical load, operating depth, operating speed, acoustic and vibration feedback signals, fuel consumption, and slip ratio status parameters.

[0011] The historical operation data includes the location information of historical sampling, mechanical load, operation depth, and operation speed.

[0012] Furthermore, the soil hardness spatial prediction modeling is used to predict the soil hardness in front of the cultivated land and generate an operable reference operation depth; wherein, the construction of the cultivated land spatiotemporal correlation specifically involves constructing standardized time series data by performing time alignment, outlier processing and normalization on the cultivated land operation data, then dividing the cultivated land area into grids according to GPS spatial location, mapping the standardized time series data to the corresponding grid cells, and performing spatiotemporal correlation encoding to obtain gridded spatiotemporal correlation data;

[0013] The multi-source sensor fusion modeling is used to combine historical operation data and real-time sensor feedback to predict soil hardness. Specifically, based on gridded spatiotemporal correlation data, the historical baseline of soil hardness is obtained by calculating the relationship between historical mechanical load and operation depth of each grid cell. The historical baseline of soil hardness is then dynamically corrected by combining real-time collected mechanical load and acoustic vibration feedback signals to obtain the predicted value of soil hardness for the current grid cell.

[0014] The spatial interpolation prediction specifically involves obtaining the predicted soil hardness values ​​of the sampled grid cells and, in conjunction with the tractor's current direction of travel, extracting the target area ahead. Then, a spatial interpolation algorithm is used to continuously estimate the soil hardness of the unsampled grid cells within the target area, thereby reconstructing the continuous hardness field of the target area and obtaining a continuous soil hardness spatial map covering the target area in front of the agricultural machinery.

[0015] The depth reference value mapping is used to map the predicted soil hardness value to an operable reference working depth. Specifically, by using a preset mapping function between soil hardness and working depth, the depth mapping calculation is performed on the predicted soil hardness value in the continuous soil hardness space map to obtain the reference working depth corresponding to each grid cell and generate a reference working depth map.

[0016] Furthermore, the tillage depth control is used to predictively and dynamically adjust the tillage depth of agricultural machinery; wherein, the target operating depth generation is used to generate a target operating depth based on the soil hardness change trend, specifically by calculating the soil hardness gradient between adjacent grid cells based on the soil hardness spatial map, then generating a depth adjustment increment based on the soil hardness gradient, superimposing the depth adjustment increment with the reference tillage depth to obtain the initial target operating depth, and then applying upper and lower limit constraints to the initial target operating depth to obtain the target operating depth;

[0017] The depth change smoothing control specifically involves dynamically adjusting the smoothing filter parameters based on real-time collected operating speed and mechanical load, and continuously filtering the target operating depth using a one-dimensional smoothing filter to achieve depth change smoothing control and obtain a smooth target operating depth sequence.

[0018] The state space construction specifically involves acquiring the predicted soil hardness value and smoothed target working depth corresponding to each grid cell, and combining it with real-time collected mechanical load, working speed and slip ratio state parameters to perform multi-dimensional working state feature fusion to construct a depth control state space, providing state input for subsequent adaptive optimization of working depth;

[0019] The adaptive optimization of the working depth is used to achieve adaptive and differentiated depth control for different soil regions. Specifically, based on the depth control state space, the adjustment amount of the agricultural machinery working depth is constructed as the action space, and a reward function is constructed according to the mechanical load deviation and the working depth deviation. The depth control strategy is trained and optimized using a reinforcement learning algorithm to obtain an adaptive depth control strategy that can adapt to different soil regions, and the optimal adjustment amount of the agricultural machinery working depth is output based on the adaptive depth control strategy.

[0020] The depth control execution specifically involves generating a corresponding hydraulic execution control signal based on the optimal agricultural machinery operating depth adjustment amount, and outputting the hydraulic execution control signal to the agricultural machinery hydraulic control system to drive the plow lifting mechanism to achieve real-time depth adjustment; at the same time, the operating depth and mechanical load are collected in real time during the execution process, the execution error is monitored and feedback is performed, and the hydraulic execution signal is dynamically corrected to achieve stable tracking control of the target operating depth;

[0021] Through the aforementioned depth control execution process, the accuracy of agricultural machinery tillage depth control is improved, the stability of the operation process is enhanced, and the adaptive adjustment capability for different soil regions is improved, thereby achieving overall optimization and improvement of agricultural machinery tillage depth control performance.

[0022] The beneficial effects achieved by the present invention using the above solution are as follows:

[0023] (1) This scheme creatively adopts a spatial prediction modeling method for soil hardness based on multi-source spatiotemporal fusion to realize the prediction of the soil hardness change trend in the target area in front of the agricultural machinery in advance, and generates a reference working depth corresponding to the soil hardness, so that the agricultural machinery can adjust the depth in advance before entering the hard soil area, reduce mechanical impact and traction load fluctuation, and improve the uniformity of working depth and the stability of agricultural machinery operation.

[0024] (2) This scheme creatively adopts a method for controlling the depth of cultivated land based on soil hardness prediction. It analyzes the trend of soil hardness change in front of the cultivated land in advance and dynamically generates the target operating depth based on the trend of soil hardness change. At the same time, it combines mechanical load, operating speed and slip ratio to smoothly adjust and adaptively optimize the depth change process, so as to realize the continuous and stable depth adjustment of agricultural machinery in different soil areas, thereby improving the adaptability of depth control in complex cultivated land environment and the continuity of cultivated land operation, and improving the crop root growth environment. Attached Figure Description

[0025] Figure 1 A schematic diagram of the modules of the agricultural machinery tillage depth control and improvement system based on big data provided by the present invention;

[0026] Figure 2 A schematic diagram of the soil hardness spatial prediction modeling module provided by the present invention;

[0027] Figure 3 This is a flowchart illustrating the process of the farmland depth control module.

[0028] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0029] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0030] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0031] Example 1, see Figure 1 The present invention provides a big data-based agricultural machinery tillage depth control and improvement system, which includes a tillage data sensing module, a soil hardness spatial prediction modeling module, and a tillage depth control module.

[0032] The cultivated land data sensing module is used to sense cultivated land data, obtain cultivated land operation data, and send the cultivated land operation data to the soil hardness spatial modeling module and the cultivated land depth control module.

[0033] The soil hardness spatial modeling module is used for soil hardness spatial modeling. Based on farmland operation data, it adopts a soil hardness spatial prediction modeling method based on multi-source spatiotemporal fusion to construct farmland spatiotemporal association, perform multi-source sensor fusion modeling, spatial interpolation prediction and depth reference value mapping, and obtain a soil hardness spatial map and depth reference value mapping results covering the target area in front of the agricultural machinery. The soil hardness spatial map and depth reference value mapping results are then sent to the farmland depth control module.

[0034] The cultivated land depth control module is used for cultivated land depth control. Based on cultivated land operation data, soil hardness spatial map and depth reference value mapping results, it adopts a cultivated land depth control method based on soil hardness prediction to generate target operation depth, control depth change smoothing, construct state space, adaptively optimize operation depth and execute depth control, so as to obtain adaptive continuous cultivated land depth control results for different soil regions.

[0035] Example 2, see Figure 1 This embodiment is based on the above embodiment. The farmland data perception specifically involves sampling the real-time operation data of agricultural machinery during the farmland tillage process to obtain location information and mechanical operation parameters, and combining the data with historical operation data to form farmland operation data.

[0036] The location information is specifically obtained through a GPS receiver installed on the agricultural machinery, including continuous latitude and longitude coordinates, timestamp information, and the agricultural machinery's travel trajectory, which is used to provide tillage path and spatial positioning reference;

[0037] The mechanical operation parameters are specifically collected through the agricultural machinery CAN bus or operation monitoring system, including mechanical load, operation depth, operation speed, sound and vibration feedback signals, fuel consumption and slip ratio status parameters, which are used to reflect the operation status and indirectly characterize soil resistance.

[0038] The historical operation data includes historical sampling location information, mechanical load, operation depth, and operation speed;

[0039] The data is uniformly organized, specifically by classifying and archiving real-time collected location information and mechanical operation parameters, as well as historical operation data, according to preset data fields and standardizing their formats.

[0040] Example 3, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. The spatial prediction modeling of soil hardness is used to predict the soil hardness in front of the cultivated land and generate an operable reference operation depth. Specifically, the construction of the spatiotemporal correlation of cultivated land involves constructing standardized time-series data by performing time alignment, outlier processing, and normalization on the cultivated land operation data. Then, the cultivated land area is divided into grids according to the GPS spatial location, and the standardized time-series data is mapped to the corresponding grid cells and spatiotemporal correlation encoding is performed to obtain gridded spatiotemporal correlation data.

[0041] The outlier processing includes, but is not limited to, moving mean filtering, median filtering, sliding window smoothing, or outlier removal based on statistical thresholds.

[0042] The multi-source sensor fusion modeling is used to combine historical operation data and real-time sensor feedback to predict soil hardness. Specifically, based on gridded spatiotemporal correlation data, the historical baseline of soil hardness is obtained by calculating the relationship between historical mechanical load and operation depth of each grid cell. The historical baseline of soil hardness is then dynamically corrected by combining real-time collected mechanical load and acoustic vibration feedback signals to obtain the predicted value of soil hardness for the current grid cell.

[0043] Optionally, nonlinear hardness changes can be captured by multi-parameter regression models or lightweight neural networks, and the nonlinear characteristics of soil hardness changes can be learned to achieve soil hardness prediction.

[0044] The formula for calculating the historical baseline of soil hardness is:

[0045] ;

[0046] In the formula, H hist (i) represents the historical soil hardness baseline of the i-th grid cell, where i is the grid cell index, and N is the number of grid cells. i T is the number of sampling points in the i-th grid cell, t is the sampling time index, and T iIt is the set of historical sampling times for the i-th grid cell, f(L) t D t ) is a linear relationship function between historical mechanical load and working depth, L t The mechanical load D is time t. t It is the depth of the operation at time t;

[0047] The calculation formula for the dynamic correction is as follows:

[0048] ;

[0049] In the formula, H pred (i) is the predicted soil hardness value of the i-th grid cell, and w1 and w2 are sensor weight coefficients. The sensor weight coefficients are set by least squares fitting or regression training based on historical data. c This is the current real-time mechanical load. It is the historical average mechanical load, S c It is the current real-time acoustic vibration feedback signal;

[0050] The spatial interpolation prediction specifically involves obtaining the predicted soil hardness values ​​of the sampled grid cells and, in conjunction with the tractor's current direction of travel, extracting the target area within a 10-20 meter range ahead. Then, a spatial interpolation algorithm is used to continuously estimate the soil hardness of the unsampled grid cells within the target area, thereby reconstructing the continuous hardness field of the target area and obtaining a continuous soil hardness spatial map covering the target area in front of the agricultural machinery.

[0051] The spatial interpolation algorithms include, but are not limited to, Kriging interpolation, inverse distance weighted interpolation, spline function interpolation, and radial basis function interpolation.

[0052] The depth reference value mapping is used to map the predicted soil hardness value to an operable reference working depth. Specifically, by using a preset mapping function between soil hardness and working depth, the depth mapping calculation is performed on the predicted soil hardness value in the continuous soil hardness space map to obtain the reference working depth corresponding to each grid cell and generate a reference working depth map.

[0053] The mapping function may specifically be a linear mapping function, a piecewise nonlinear mapping function, or an adaptive mapping function based on empirical calibration.

[0054] Preferably, the formula for calculating the reference working depth is:

[0055] ;

[0056] In the formula, D ref(x, y) represents the reference working depth at spatial coordinates (x, y), where x is the x-coordinate of the grid cell and y is the y-coordinate of the grid cell. target It is the target operation depth, k d It is the depth adjustment coefficient, which is specifically set through historical operation data calibration, soil type parameters, or empirical adjustment rules. H pred (x,y) is the predicted soil hardness value at spatial coordinates (x,y), H ref It is the standard hardness baseline, which is specifically set based on historical soil hardness statistics, regional average hardness, or preset standard hardness values.

[0057] By performing the above operations, this solution creatively adopts a spatial prediction modeling method for soil hardness based on multi-source spatiotemporal fusion, which enables early prediction of the soil hardness change trend in the target area in front of the agricultural machinery and generates a reference working depth corresponding to the soil hardness. This allows the agricultural machinery to adjust its depth in advance before entering the hard soil area, reducing mechanical impact and traction load fluctuations, improving the uniformity of working depth and the stability of agricultural machinery operation.

[0058] Example 4, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. The tillage depth control is used to predictively and dynamically adjust the tillage depth of agricultural machinery. The target operating depth generation is used to generate a target operating depth based on the soil hardness change trend. Specifically, based on the soil hardness spatial map, the soil hardness gradient between adjacent grid cells is calculated. Then, a depth adjustment increment is generated based on the soil hardness gradient. The depth adjustment increment is superimposed with the reference tillage depth to obtain the initial target operating depth. Then, the initial target operating depth is constrained by upper and lower limits to obtain the target operating depth.

[0059] The method of generating depth adjustment increment based on soil hardness gradient specifically involves mapping the soil hardness gradient proportionally by setting a depth adjustment coefficient, thereby generating a corresponding depth adjustment increment so that areas with large changes in soil hardness correspond to a larger depth adjustment range.

[0060] The formula for calculating the target working depth is:

[0061] ;

[0062] ;

[0063] In the formula, D init (x,y) is the initial target operation depth, D ref (x,y) is the reference job depth. It is a depth adjustment increment, D target(x, y) represents the target working depth, min(·) is the function to find the minimum value, and max(·) is the function to find the maximum value. min It is the minimum allowed working depth, D max This is the maximum allowed working depth;

[0064] The depth change smoothing control is used to reduce the mechanical impact and uneven tillage caused by rapid changes in working depth. Specifically, it dynamically adjusts the smoothing filter parameters by collecting the working speed and mechanical load in real time, and uses a one-dimensional smoothing filter to continuously filter the target working depth to achieve depth change smoothing control and obtain a smooth target working depth sequence.

[0065] The formula for calculating the smoothing filter parameters is as follows:

[0066] ;

[0067] In the formula, It is the smoothing filter parameter, k v It is the speed regulation coefficient, V c This is the current real-time operation speed, k. l It is the load adjustment coefficient, and the speed adjustment coefficient and the load adjustment coefficient are obtained through agricultural machinery field operation tests. L c This is the current real-time mechanical load;

[0068] The state space construction specifically involves acquiring the predicted soil hardness value and smoothed target working depth corresponding to each grid cell, and combining it with real-time collected mechanical load, working speed and slip ratio state parameters to perform multi-dimensional working state feature fusion to construct a depth control state space, providing state input for subsequent adaptive optimization of working depth;

[0069] The adaptive optimization of the working depth is used to achieve adaptive and differentiated depth control for different soil regions. Specifically, based on the depth control state space, the adjustment amount of the agricultural machinery working depth is constructed as the action space, and a reward function is constructed according to the mechanical load deviation and the working depth deviation. The depth control strategy is trained and optimized using a reinforcement learning algorithm to obtain an adaptive depth control strategy that can adapt to different soil regions, and the optimal adjustment amount of the agricultural machinery working depth is output based on the adaptive depth control strategy.

[0070] The formula for calculating the reward function is as follows:

[0071] ;

[0072] In the formula, r i It is the reward function value, L target It is the target mechanical load. This refers to the depth deviation weighting coefficient, whose initial value can be determined through empirical calibration and dynamically corrected by combining it with real-time soil hardness fluctuation indicators. c It is the current real-time job depth, D smooth It smooths the target operation depth;

[0073] The depth control execution specifically involves generating a corresponding hydraulic execution control signal based on the optimal agricultural machinery operating depth adjustment amount, and outputting the hydraulic execution control signal to the agricultural machinery hydraulic control system to drive the plow lifting mechanism to achieve real-time depth adjustment; at the same time, the operating depth and mechanical load are collected in real time during the execution process, the execution error is monitored and feedback is performed, and the hydraulic execution signal is dynamically corrected to achieve stable tracking control of the target operating depth;

[0074] Through the aforementioned depth control execution, the accuracy of agricultural machinery tillage depth control is improved, the stability of the operation process is enhanced, and the adaptive adjustment capability for different soil regions is improved, thereby achieving overall optimization and improvement of agricultural machinery tillage depth control performance.

[0075] By performing the above operations, this solution creatively adopts a tillage depth control method based on soil hardness prediction. It analyzes the trend of soil hardness changes in front of the tillage in advance and dynamically generates the target operating depth based on the trend of soil hardness changes. At the same time, it combines mechanical load, operating speed and slip ratio to smoothly adjust and adaptively optimize the depth change process, so as to realize continuous and stable depth adjustment of agricultural machinery in different soil areas. This improves the adaptability of depth control in complex tillage environments and the continuity of tillage operations, and improves the crop root growth environment.

[0076] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0077] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

[0078] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A big data-based agricultural machinery tillage depth control and improvement system, characterized in that: It includes a farmland data sensing module, a soil hardness spatial prediction modeling module, and a farmland depth control module; The cultivated land data sensing module is used to sense cultivated land data, obtain cultivated land operation data, and send the cultivated land operation data to the soil hardness spatial modeling module and the cultivated land depth control module. The soil hardness spatial modeling module is used for soil hardness spatial modeling. Based on farmland operation data, it adopts a soil hardness spatial prediction modeling method based on multi-source spatiotemporal fusion to construct farmland spatiotemporal association, perform multi-source sensor fusion modeling, spatial interpolation prediction and depth reference value mapping, and obtain a soil hardness spatial map and depth reference value mapping results covering the target area in front of the agricultural machinery. The soil hardness spatial map and depth reference value mapping results are then sent to the farmland depth control module. The multi-source sensor fusion modeling is used to combine historical operation data and real-time sensor feedback to predict soil hardness. Specifically, based on gridded spatiotemporal correlation data, the historical baseline of soil hardness is obtained by calculating the relationship between historical mechanical load and operation depth of each grid cell. The historical baseline of soil hardness is then dynamically corrected by combining real-time collected mechanical load and acoustic vibration feedback signals to obtain the predicted value of soil hardness for the current grid cell. The spatial interpolation prediction specifically involves obtaining the predicted soil hardness values ​​of the sampled grid cells and, in conjunction with the tractor's current direction of travel, extracting the target area ahead. Then, a spatial interpolation algorithm is used to continuously estimate the soil hardness of the unsampled grid cells within the target area, thereby reconstructing the continuous hardness field of the target area and obtaining a continuous soil hardness spatial map covering the target area in front of the agricultural machinery. The tillage depth control module is used for tillage depth control. Based on tillage operation data, soil hardness spatial map and depth reference value mapping results, it adopts a tillage depth control method based on soil hardness prediction to generate target operation depth, control depth change smoothing, generate corresponding hydraulic execution control signals based on the optimal agricultural machinery operation depth adjustment, adaptive optimization of operation depth and execution of depth control, so as to obtain adaptive continuous tillage depth control results for different soil regions.

2. The agricultural machinery tillage depth control and improvement system based on big data according to claim 1, characterized in that: The construction of the spatiotemporal correlation of cultivated land is specifically carried out by performing time alignment, outlier processing and normalization on cultivated land operation data to construct standardized time series data. Then, the cultivated land area is divided into grids according to GPS spatial location, and the standardized time series data is mapped to the corresponding grid cells and spatiotemporal correlation encoding is performed to obtain gridded spatiotemporal correlation data.

3. The agricultural machinery tillage depth control and improvement system based on big data according to claim 2, characterized in that: The depth reference value mapping is used to map the predicted soil hardness value to an operable reference working depth. Specifically, by using a preset mapping function between soil hardness and working depth, the depth mapping calculation is performed on the predicted soil hardness value in the continuous soil hardness space map to obtain the reference working depth corresponding to each grid cell and generate a reference working depth map.

4. The agricultural machinery tillage depth control and improvement system based on big data according to claim 3, characterized in that: The target operation depth generation is used to generate the target operation depth based on the soil hardness change trend. Specifically, based on the soil hardness spatial map, the soil hardness gradient between adjacent grid cells is calculated. Then, a depth adjustment increment is generated based on the soil hardness gradient. The depth adjustment increment is superimposed with the reference tillage depth to obtain the initial target operation depth. Finally, upper and lower limits are constrained on the initial target operation depth to obtain the target operation depth.

5. The agricultural machinery tillage depth control and improvement system based on big data according to claim 4, characterized in that: The depth change smoothing control specifically involves dynamically adjusting the smoothing filter parameters based on real-time collected operating speed and mechanical load, and continuously filtering the target operating depth using a one-dimensional smoothing filter to achieve depth change smoothing control and obtain a smooth target operating depth sequence.

6. The agricultural machinery tillage depth control and improvement system based on big data according to claim 5, characterized in that: The process of generating corresponding hydraulic execution control signals based on the optimal agricultural machinery operation depth adjustment is specifically achieved by acquiring the predicted soil hardness value and smoothed target operation depth corresponding to each grid unit, and combining it with real-time collected mechanical load, operation speed and slip ratio state parameters to perform multi-dimensional operation state feature fusion, thereby constructing a depth control system that generates corresponding hydraulic execution control signals based on the optimal agricultural machinery operation depth adjustment, providing state input for subsequent adaptive optimization of operation depth. The adaptive optimization of the operating depth is used to achieve adaptive and differentiated depth control for different soil regions. Specifically, it generates corresponding hydraulic execution control signals based on the optimal agricultural machinery operating depth adjustment amount according to the depth control, constructs the agricultural machinery operating depth adjustment amount as an action space, constructs a reward function based on the mechanical load deviation and operating depth deviation, and uses a reinforcement learning algorithm to train and optimize the depth control strategy to obtain an adaptive depth control strategy that can adapt to different soil regions, and outputs the optimal agricultural machinery operating depth adjustment amount based on the adaptive depth control strategy.

7. The agricultural machinery tillage depth control and improvement system based on big data according to claim 6, characterized in that: The depth control execution specifically involves generating a corresponding hydraulic execution control signal based on the optimal agricultural machinery operating depth adjustment amount, and outputting the hydraulic execution control signal to the agricultural machinery hydraulic control system to drive the plow lifting mechanism to achieve real-time depth adjustment; at the same time, the operating depth and mechanical load are collected in real time during the execution process, the execution error is monitored and feedback is performed, and the hydraulic execution signal is dynamically corrected to achieve stable tracking control of the target operating depth.

8. The agricultural machinery tillage depth control and improvement system based on big data according to claim 7, characterized in that: The aforementioned farmland data sensing specifically involves sampling the location information and mechanical operation parameters of agricultural machinery during the tillage process, and then combining this data with basic soil information to form farmland operation data.

Images