A leymus chinensis planting integrated operation coordination system using intelligent control

Abstract

The application discloses a kind of integrated operation coordination systems of Leymus mollis planting using intelligent control, and the application relates to the field of wisdom agriculture technology, including cloud server, the cloud server is connected with the following module communication: perception decision coordination module is used to construct soil attribute space distribution map by comprehensive acquisition soil parameter, and cloud cooperation and fruit fly algorithm optimization rotary tillage strategy is used, rotary tillage depth is determined to carry out the adaptive regulation and control of rotary tillage unit.The application constructs high-precision soil attribute space distribution map by collecting soil parameters, uses fruit fly algorithm to quickly calculate the optimal depth combination of each rotary tillage unit, and issues instructions to dynamically adjust hydraulic actuator, breaks through the limitation of traditional mechanical fixed operation mode, creates the best soil physical environment for the healthy growth of Leymus mollis root system, fundamentally solves the problem that root system is difficult to penetrate hard soil layer due to strong spatial heterogeneity of soil hardness.

Description

An integrated collaborative system for sheepgrass planting using intelligent control Technical Field

[0001] This invention relates to the field of smart agriculture technology, specifically to an integrated collaborative system for sheepgrass planting using intelligent control. Background Technology

[0002] Leymus chinensis, as a dominant species in grasslands, plays a vital role in maintaining grassland ecological balance and promoting biodiversity. However, due to factors such as climate change, soil salinization, and overgrazing, the degradation of Leymus chinensis on grasslands is becoming increasingly serious. Traditional grassland restoration methods often employ all-round tillage operations, which can easily damage grassland biodiversity. To effectively restore Leymus chinensis and maintain ecological balance, mechanized and intelligent planting techniques have become a feasible solution. The introduction of intelligent control technology can achieve precise control over multiple operational stages such as soil loosening, sowing, covering, and compaction, thereby improving planting efficiency and reducing human intervention. This integrated operation can not only improve the success rate of Leymus chinensis planting but also ensure the ecological restoration of grasslands and the maintenance of biodiversity.

[0003] In existing technologies, soil hardness and salinity in saline-alkali land exhibit strong spatial heterogeneity. Furthermore, moderately to severely degraded grassland soils have a pH of 8.5–9.5 and a salinity of 0.3–0.5%, making it difficult for traditional machinery to adaptively adjust operating parameters. This results in Leymus chinensis roots struggling to penetrate the hard soil layer, leading to inconsistent sowing depths and low germination rates. Moreover, the microenvironment at the boundary between disturbed and undisturbed areas after rotary tillage undergoes drastic changes, affecting Leymus chinensis root extension and the growth of neighboring plants. Traditional methods cannot predict the trend of root cross-growth. Additionally, in severely wind-eroded saline-alkali bare land, Leymus chinensis seeds are easily scattered after sowing, and the effectiveness of soil covering and compaction is greatly affected by wind speed and surface undulations. Traditional iron chain soil covering cannot guarantee uniformity and wind resistance. Therefore, this paper proposes an integrated collaborative system for Leymus chinensis planting using intelligent control to address the aforementioned problems. Summary of the Invention

[0004] To solve the above-mentioned technical problems, the present invention is implemented through the following technical solution: an integrated collaborative system for sheepgrass planting using intelligent control, comprising a cloud server, wherein the cloud server is communicatively connected to the following modules:

[0005] The perception and decision-making collaboration module is used to comprehensively collect soil parameters to construct a spatial distribution map of soil properties, and to optimize the rotary tillage strategy using cloud collaboration and fruit fly algorithm to determine the rotary tillage depth for adaptive control of the rotary tillage unit. The soil parameters include soil hardness, salinity and pH value.

[0006] The root growth prediction module is connected to the perception and decision-making collaboration module to construct a digital twin model of growth disturbance and predict the root interaction trend between Leymus chinensis and neighboring plants to guide the optimization of sowing density and layout.

[0007] The chain-type soil covering adaptive module combines a digital twin model of growth perturbation to analyze surface undulation data. Through reinforcement learning algorithms, it controls the swing amplitude and ground-hugging angle of the soil covering chain to adapt to uneven surfaces, ensure uniform soil covering, and reduce the risk of seed exposure and wind erosion.

[0008] The dynamic compaction and wind-resistant module is used to adjust the pressure of the compaction wheel and the spring damping in real time based on wind speed sensor data and ground undulation data, so that the compaction pressure matches the wind resistance requirements, enhances the compaction of seeds and soil, resists wind disturbance, and improves the uniformity of seedling emergence.

[0009] The task coordination and scheduling module establishes a real-time communication link between machinery and the cloud, generates collaborative operation instructions and sends them to each machinery unit, and synchronizes the rotary tillage, sowing, covering and compaction processes to achieve seamless integration of the operation process and improve operation efficiency and resource utilization efficiency.

[0010] Preferably, the sensing and decision-making collaboration module includes a soil multi-source sensing unit and an optimization decision-making unit;

[0011] The soil multi-source sensing unit, through hardness sensors and conductivity probes deployed on rotary tillage and seeding machinery, collects soil parameters including soil hardness, salinity and pH value in real time, constructs a high-precision spatial distribution map of soil properties, realizes dynamic identification of soil hardness and salinity heterogeneity, and provides a data basis for rotary tillage depth regulation.

[0012] The optimization decision-making unit is used to upload the spatial distribution map of soil properties and the soil parameters it covers to the cloud, use the fruit fly algorithm to quickly solve the optimal combination of rotary tillage pressing depth, and dynamically adjust the operating parameters of each rotary tillage unit.

[0013] Preferably, the execution steps of the perception-decision collaboration module include:

[0014] By deploying multiple hardness sensors and conductivity probes on the rotary tillage and seeding machinery, soil parameter data, including soil hardness, salinity and pH value, are collected in real time and synchronously along the travel path.

[0015] Based on the collected soil parameter data, a high-precision spatial distribution map of soil properties is constructed using a spatial interpolation algorithm to dynamically identify the spatial heterogeneity of soil hardness and salinity.

[0016] Preferably, the execution steps of the optimization decision unit include:

[0017] Upload the spatial distribution map of the soil properties and its corresponding soil parameters to the cloud server, and parse the real-time soil hardness value of each operation point as the input parameters of the fruit fly algorithm;

[0018] Using the fruit fly algorithm with the objective function of root penetration into hard soil layers, iterative optimization was performed to analyze the optimal compaction depth for each rotary tillage unit.

[0019] The optimal combination of compaction depths is sent to the local controller, which adjusts the hydraulic actuators of each rotary tillage unit to adaptively match the rotary tillage depth with soil hardness.

[0020] Preferably, the root growth prediction module includes a disturbance interface modeling unit and a root conflict prediction unit;

[0021] The disturbance interface modeling unit is used to combine rotary tillage depth and soil hardness data to construct a digital twin model of growth disturbance, simulate the microenvironmental changes at the boundary between the disturbed and undisturbed areas, reveal the root extension dynamics at the interface, and identify potential growth conflict areas.

[0022] The root conflict prediction unit predicts growth conflict areas based on a digital twin model of growth disturbance, analyzes the cross-growth trend of Leymus chinensis roots and neighboring plant roots, calculates the competition intensity index, divides different conflict risk zones, proposes differentiated sowing density adjustment strategies, guides sowing density adjustment, avoids growth inhibition caused by root competition, and improves the survival rate and community stability of Leymus chinensis.

[0023] Preferably, the execution steps of the disturbance interface modeling unit include:

[0024] The rotary tillage depth setpoint and soil hardness spatial distribution data output by the perception and decision-making collaboration module are obtained as the initial boundary conditions for the digital twin model.

[0025] Based on soil mechanics and plant root growth models, a digital twin model of growth disturbance is constructed to reflect the changes in water, nutrients and mechanical resistance at the boundary between disturbed and undisturbed areas.

[0026] Run the aforementioned growth perturbation digital twin model to simulate the extension dynamics and growth trend of Leymus chinensis roots at the interface, and visualize potential root growth conflict areas.

[0027] Preferably, the execution steps of the root conflict prediction unit include:

[0028] Based on the potential root growth conflict areas output by the growth perturbation digital twin model, the cross-growth competition trend between Leymus chinensis and neighboring plant roots is analyzed, the competition intensity index is calculated, and different conflict risk areas are divided in combination with the preset conflict threshold.

[0029] Based on the analysis of cross-growth competition trends, a sowing density adjustment strategy is generated to automatically reduce the sowing density in high-risk conflict areas and maintain or appropriately increase the density in low-risk areas.

[0030] The seeding density adjustment strategy and the spatial layout information covered by the soil property spatial distribution map are integrated into the control unit of the rotary tillage seeder to form a variable operation prescription map, which provides instructions for precision variable seeding.

[0031] Preferably, the execution steps of the chain-type soil covering adaptive module include:

[0032] Based on the digital twin model of growth perturbation and the spatial distribution map of soil properties, the surface undulation data is analyzed and used as the environmental state input for the reinforcement learning agent.

[0033] Based on reinforcement learning algorithms, the system calculates and outputs control signals for the swing amplitude and ground-hugging angle of the soil-covering chain according to the surface undulation data. The soil-covering chain mechanism is driven by a servo motor to move closely to the uneven ground surface, ensuring the uniformity of the soil covering thickness in the seed furrow and reducing seed exposure.

[0034] Preferably, the execution steps of the dynamic pressure-resistant wind-resistant module include:

[0035] By collecting ambient wind speed in real time through a wind speed sensor integrated into the rotary tillage and seeding machinery, and combining it with surface undulation data, the compaction requirements are comprehensively analyzed.

[0036] Based on the assessment results, the pressing pressure of the press wheel is adjusted in real time through the electro-hydraulic system, and the damping coefficient of the spring is dynamically adjusted through the magnetorheological damper. This ensures that the total pressing pressure of the press wheel matches the real-time wind resistance requirements, ensuring close contact between the seeds and the soil and effectively resisting seed displacement or soil erosion caused by wind.

[0037] Preferably, the execution steps of the task collaborative scheduling module include:

[0038] Establish a low-latency, highly reliable communication link between the local controller of the rotary tillage and seeding machinery and the cloud server to ensure bidirectional real-time transmission of commands and data;

[0039] Based on the real-time status and output of each module, a globally optimized collaborative operation instruction sequence is generated on the cloud server, covering the start and stop timing and parameters of rotary tillage, sowing, covering and compaction. The collaborative operation instruction sequence is then distributed to the actuators of each mechanical unit to achieve precise synchronization and seamless connection of the integrated operation process, thereby improving the overall operation efficiency of the system.

[0040] This invention provides an integrated collaborative system for sheepgrass planting using intelligent control. It offers the following advantages:

[0041] (I) This integrated collaborative system for sheepgrass planting using intelligent control constructs a high-precision spatial distribution map of soil properties by collecting soil parameters, quickly calculates the optimal combination of compaction depth for each rotary tillage unit using the fruit fly algorithm, and issues instructions to dynamically adjust the hydraulic actuator. This breaks through the limitations of traditional fixed mechanical operation mode, achieves precise adaptive matching between rotary tillage depth and real-time soil hardness in the field, creates the best soil physical environment for the healthy growth of sheepgrass roots, and fundamentally solves the problem that roots cannot penetrate hard soil layers due to the strong spatial heterogeneity of soil hardness.

[0042] (II) This intelligent control integrated operation collaborative system for Leymus chinensis planting simulates the microenvironmental changes at the boundary between disturbed and undisturbed areas and the root extension dynamics of Leymus chinensis and neighboring plants by constructing a digital twin model of growth disturbance. It calculates the probability of root spatial intersection and the competition intensity index, accurately identifies and visualizes potential growth conflict areas, and generates differentiated sowing density adjustment strategies. In high-conflict-risk areas, the density is automatically reduced to alleviate competition, while in low-risk areas, the density is maintained or increased to make full use of resources. This realizes the transformation from uniform sowing to differentiated resource allocation based on competition risk, effectively avoiding growth inhibition caused by root competition and improving the survival rate and community stability of Leymus chinensis. Attached Figure Description

[0043] Figure 1 is a schematic diagram of the workflow of a smart control integrated operation and collaborative system for sheepgrass planting according to the present invention;

[0044] Figure 2 is a block diagram illustrating the working principle of an integrated collaborative system for sheepgrass planting using intelligent control according to the present invention. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] Example 1, please refer to Figures 1 and 2. This invention provides a technical solution: an integrated collaborative system for sheepgrass planting using intelligent control, including a cloud server. The cloud server has the following communication connections:

[0047] The perception and decision-making collaboration module is used to comprehensively collect soil parameters to construct a spatial distribution map of soil properties, and to optimize the rotary tillage strategy using cloud collaboration and fruit fly algorithm to determine the rotary tillage depth for adaptive control of the rotary tillage unit. The soil parameters include soil hardness, salinity and pH value. The perception and decision-making collaboration module includes a multi-source soil perception unit and an optimization decision-making unit.

[0048] The multi-source soil sensing unit, through hardness sensors and conductivity probes deployed on rotary tillage and seeding machinery, collects soil parameters including soil hardness, salinity, and pH value in real time, constructs a high-precision spatial distribution map of soil properties, and realizes dynamic identification of soil hardness and salinity heterogeneity, providing a data foundation for rotary tillage depth regulation. By deploying multiple hardness sensors and conductivity probes on rotary tillage and seeding machinery, it collects soil parameter data including soil hardness, salinity, and pH value along the travel path in real time and synchronously. Based on the collected soil parameter data, a high-precision spatial distribution map of soil properties is constructed through spatial interpolation algorithms, dynamically identifying the spatial heterogeneity of soil hardness and salinity.

[0049] The specific work involves: using multiple sensors deployed on rotary tillage and seeding machinery, including soil hardness sensors and conductivity probes, to collect soil parameters in real time and synchronously during the movement of the rotary tillage and seeding machinery. These parameters include soil hardness (soil penetration resistance), salinity (conductivity, which indirectly reflects salinity), and pH value (hydrogen ion concentration). The data collection ensures the synchronization of data in spatial location and the consistency of data in time. Based on this, the system uses spatial interpolation algorithms to process the discrete sampling point data. Based on the measured values ​​of known points, it calculates the soil properties of unsampled areas and constructs a continuous, high-precision spatial distribution map of soil properties. This map intuitively displays the three-dimensional distribution of soil hardness and salinity in the field, enabling dynamic identification and visualization of the spatial heterogeneity of soil hardness and salinity, and accurately revealing the dramatic changes in soil conditions in space.

[0050] The optimized decision-making unit uploads the spatial distribution map of soil properties and the soil parameters it covers to the cloud. It uses the fruit fly algorithm to quickly solve the optimal combination of rotary tillage compaction depths, dynamically adjusts the operating parameters of each rotary tillage unit, breaks through the traditional fixed mechanical operation mode, and achieves adaptive matching between rotary tillage depth and soil hardness. This ensures that the roots can penetrate the hard soil layer. The spatial distribution map of soil properties and the corresponding soil parameters are uploaded to the cloud server, and the real-time soil hardness value of each operation point is parsed as the input parameter of the fruit fly algorithm. The fruit fly algorithm uses the ability of the roots to penetrate the hard soil layer as the objective function to perform iterative optimization, analyze the optimal compaction depth corresponding to each rotary tillage unit, and send the obtained optimal compaction depth combination to the local controller to adjust the hydraulic actuators of each rotary tillage unit to adaptively match the rotary tillage depth and soil hardness.

[0051] The specific work involves: uploading a high-precision spatial distribution map of soil properties and corresponding soil parameters constructed from multi-source soil sensing units to a cloud server; analyzing the map to extract real-time soil hardness values ​​at each specific work site; using these real-time soil hardness values ​​as the core input parameter for the fruit fly algorithm; and using the objective function of ensuring that the Leymus chinensis roots can successfully penetrate the hard soil layer to perform iterative optimization calculations, simulating the foraging behavior of fruit fly colonies. By continuously evaluating and adjusting possible combinations of rotary tillage depths, the algorithm quickly converges and obtains the optimal compaction depth corresponding to each independent rotary tillage unit. After completing the analysis of the optimal compaction depth, the algorithm enters the command issuance and execution stage, where the cloud server will execute the calculated optimal compaction depth. The optimal compaction depth combination is distributed to the local controller of the rotary tillage and seeding machine through a stable, low-latency communication link. This local controller acts as the field command center, receiving relevant data of the optimal compaction depth combination and parsing it into specific, executable command signals. These command signals are then precisely sent to the hydraulic actuators of the corresponding rotary tillage units. The hydraulic actuators dynamically adjust their stroke and pressure according to the command signals, accurately controlling the penetration depth of each rotary tillage unit. This achieves adaptive matching between rotary tillage operation parameters and real-time soil hardness data sensed in the field, ensuring that the rotary tillage depth is precisely controlled under any hardness heterogeneity conditions, creating the best soil physical environment for the healthy growth of Leymus chinensis roots.

[0052] The root growth prediction module, connected to the perception and decision-making collaboration module, constructs a digital twin model of growth disturbance to predict the root interaction trend between Leymus chinensis and neighboring plants, so as to guide the optimization of sowing density and layout. The root growth prediction module includes a disturbance interface modeling unit and a root conflict prediction unit.

[0053] Among them, the disturbance interface modeling unit is used to combine rotary tillage depth and soil hardness data to construct a digital twin model of growth disturbance, simulate the microenvironmental changes at the boundary between the disturbed and undisturbed areas, reveal the root extension dynamics at the interface, identify potential growth conflict areas, and obtain the rotary tillage depth setpoint and soil hardness spatial distribution data output by the perception and decision-making collaboration module as the initial boundary conditions of the digital twin model. Based on soil mechanics and plant root growth models, a growth disturbance digital twin model reflecting the changes in water, nutrients and mechanical resistance at the boundary between the disturbed and undisturbed areas is constructed. The growth disturbance digital twin model is run to simulate the extension dynamics and growth trend of Leymus chinensis roots at the interface and visualize potential root growth conflict areas.

[0054] The specific tasks are as follows: Receiving real-time data from the perception-decision collaboration module, including the rotary tillage depth settings for each location determined by the fruit fly algorithm, and a high-precision spatial distribution map of soil hardness, these are combined to form the initial boundary conditions for the digital twin model. This provides a realistic physical world mapping for the virtual space simulation. By importing the rotary tillage depth settings and the soil hardness spatial distribution map, the actual field operation scenario is accurately replicated in the digital space. This involves depicting the physical interface formed by the rotary tillage blades, between loose, disturbed soil and compacted, undisturbed soil, laying the geometric and physical foundation for the entire simulation and ensuring that root growth simulation can be conducted in a highly realistic and data-driven virtual environment. After obtaining the initial boundary conditions, the multidisciplinary knowledge of soil mechanics principles and plant root growth models is integrated to characterize the mechanical resistance gradient at the interface caused by abrupt changes in soil bulk density and porosity. The algorithm further incorporates root growth algorithms of Leymus chinensis and other typical associated plants to simulate the tropism of root growth in response to environmental stress. This allows for the construction of a digital twin model of growth disturbance that accurately reflects changes in the microenvironment of the soil disturbance boundary zone. This model reveals the driving mechanism and constraint effect of the special microenvironment at the disturbance interface on the extension behavior of plant roots. After model construction, the digital twin model of growth disturbance is run to perform dynamic simulation and prediction. It simulates the extension dynamics of Leymus chinensis roots from the sowing point to the boundary between the disturbed and undisturbed areas in three-dimensional soil space over time. The biomass distribution and extension rate of the roots are analyzed, and the spatial intersection with neighboring plant roots is predicted. Through simulation, the competitive situation and growth trend of roots at the interface are revealed, thereby identifying potential growth conflict areas, which are highlighted and visualized in the form of heat maps.

[0055] The expression for the tropism growth behavior of simulated roots in response to environmental stress is as follows:

[0056] ;

[0057] In the formula: In time step At that time, the normalized growth direction vector of the root tip unit; The gradient operator represents the rate of change and direction of a physical quantity in space; For soil mechanical resistance, gradient Pointing in the direction where mechanical resistance decreases; Nutrient concentration gradient in soil It points in the direction of increasing nutrient concentration (phosphatism); For soil water potential, gradient Pointing towards the direction of rising water potential (higher humidity) (hydrotropism); The weighting coefficients for mechanical stress, nutrients, and water factors are respectively used to reflect the sensitivity of Leymus chinensis roots to different environmental signals; at hard soil layers / boundaries, The value is large and the gradient is significant. The weight is dominant, and the root system follows The direction of growth, namely, avoiding hard areas or finding the path of least resistance, shows obvious obstacle avoidance behavior;

[0058] The biomass distribution and extension rate of the root system can be analyzed using the following expression:

[0059] ;

[0060] In the formula: In time step At that time, the actual rate of root tip extension; This represents the maximum possible rate of root tip extension under ideal conditions. In time step The biomass allocated to that root tip at that time; It is the half-saturation constant, when the allocated biomass equals At that time, the speed reached Half of; The real-time mechanical resistance of the soil in front of the root tip; The critical mechanical resistance threshold for root growth, when When the value approaches or exceeds this value, growth is severely inhibited; and Positive correlation, following a saturated growth pattern: the more biomass allocated, the faster the growth, but there is an upper limit;

[0061] The expression for predicting the spatial intersection of Leymus chinensis root systems with those of neighboring plants is as follows:

[0062] ;

[0063] In the formula: In time ,plant (Sheepgrass) and neighboring plants The root system in the soil volume The probability of spatial intersection (competition) occurring within the space; Indicates the location respectively and time ,plant and Root density; The competition coefficient represents the saturation effect of root density on competition intensity. The volume of soil under investigation; if and only if at the same location The integrand value is large only when both root densities are high. The higher the value, the more tightly the root systems of the two are intertwined in space, and the more intense the direct competition for water and fertilizer.

[0064] The potential growth conflict region is identified by the following expression:

[0065] ;

[0066] In the formula: In time The set of potential growth conflict regions defined (i.e., the highlighted areas on the heatmap). It is a point in three-dimensional space; For the entire simulated soil space; For sheep grass in Root density; Sheepgrass; For the first Planting neighboring plants Root density; A density threshold for determining the presence and ecological significance of root systems; In time Within a unit volume of soil space, sheepgrass ( ) and the An index of the intensity of spatial overlap and resource competition among the root systems of neighboring plants; A probability threshold for determining whether the intensity of competition constitutes a conflict; The logical AND operator indicates that all conditions must be met simultaneously; for a point to be considered a conflict zone, three conditions must be met simultaneously: the roots of the sheepgrass reach this point. Other plant roots also reach this area. The intensity of competition between the two exceeds a critical value. ;

[0067] The root conflict prediction unit predicts growth conflict areas based on a digital twin model of growth perturbation, analyzes the cross-growth trend of Leymus chinensis roots and neighboring plant roots, calculates the competition intensity index, divides different conflict risk zones, and proposes differentiated sowing density adjustment strategies to guide sowing density adjustment, avoid growth inhibition caused by root competition, and improve the survival rate and community stability of Leymus chinensis. Based on the potential root growth conflict areas output by the digital twin model of growth perturbation, it analyzes the cross-growth competition trend of Leymus chinensis roots and neighboring plant roots, calculates the competition intensity index, and divides different conflict risk zones in combination with preset conflict thresholds. According to the analyzed cross-growth competition trend, it generates a sowing density adjustment strategy, automatically reduces the sowing density in high conflict risk zones, and maintains or appropriately increases the density in low conflict risk zones. The sowing density adjustment strategy and the spatial layout information covered by the soil property spatial distribution map are integrated into the control unit of the rotary tillage and seeding machine to form a variable operation prescription map, providing instruction basis for precise variable sowing.

[0068] The specific work involves: extracting potential root growth conflict areas and related data from the digital twin model of growth perturbation; analyzing the cross-growth competition trend between Leymus chinensis and neighboring plant roots based on the biomass distribution and extension rate of Leymus chinensis roots; evaluating the water and nutrient uptake capacity and space occupancy efficiency of different plant roots in specific soil microenvironments (especially the boundary between perturbed and undisturbed areas); calculating the intersection probability and competition intensity index of root systems in three-dimensional space; identifying spatial points prone to high-intensity resource competition in the future timescale; generating a competition situation map; and marking different conflict risk areas by combining the competition intensity index with a preset conflict threshold. After accurately analyzing the cross-growth competition trend of roots, a concrete seeding density adjustment strategy is generated, and reverse regulation is performed based on the marked conflict risk areas: in high conflict risk areas predicted by the model, the seeding density of Leymus chinensis is automatically reduced in a command manner, thereby reducing the density of the future root network from the source by reducing the number of Leymus chinensis plants per unit area in advance. In areas with low competition and identified as low-conflict-risk zones, the seeding density is maintained or even appropriately increased to fully utilize the ecological niche space and resource potential of each area, ensuring the overall establishment success rate and yield of Leymus chinensis populations. This represents a shift from uniform seeding to differentiated resource allocation based on predictive competition risk. The abstract seeding density adjustment strategy is integrated with the spatial layout information covered by the soil property spatial distribution map. Each decision to reduce or increase density is linked to its relative position on the machine's path, forming a machine-readable, geographically referenced variable operation prescription map. This map is then sent to the control unit of the rotary tillage seeder. Based on this, the control unit automatically adjusts the speed or opening and closing frequency of the seed metering device in real time during the movement of the rotary tillage seeder, precisely executing the preset seeding density scheme in physical space, achieving precise variable seeding with a one-site-one-policy approach.

[0069] The probability of the intersection of the root systems in three-dimensional space is expressed as:

[0070] ;

[0071] In the formula: For in position and time ,plant (Sheepgrass) and plants The probability of spatial overlap between the root systems of (neighboring plants); For in position and time ,plant and Root density; A preset value for the maximum possible root density in the entire simulation system is used for normalization; when or When any value is 0, , indicating that there is no intersection here, if and only if and When all are relatively high at the same point, It only approaches 1;

[0072] The competition intensity index is expressed as follows:

[0073] ;

[0074] ;

[0075] ;

[0076] In the formula: For in position and time The intensity of competition index; These are the stress coefficients for water and nutrients, respectively. The larger the value, the scarcer the resources and the stronger the competitive potential. For position The actual available water; The optimal water level for plant growth; For position The actual available nutrients; The optimal nutrient content for plant growth; Let be the weighting coefficients of water and nutrients in the competition, respectively, and satisfy . ;when (When there is no root overlap) , indicating no competition; when It is very high, and resources are extremely scarce. , When it tends towards 1), Tendency 1) indicates extremely fierce competition for survival, with a conflict threshold of [value missing]. , As a low-conflict-risk zone, It is a high-risk conflict zone;

[0077] The chain-type soil covering adaptive module combines a digital twin model of growth perturbation to analyze surface undulation data. Through reinforcement learning algorithms, it controls the swing amplitude and ground-hugging angle of the soil covering chain to adapt to uneven surfaces, ensure uniform soil covering, and reduce the risk of seed exposure and wind erosion.

[0078] The dynamic compaction and wind-resistant module is used to adjust the pressure of the compaction wheel and the spring damping in real time based on wind speed sensor data and ground undulation data, so that the compaction pressure matches the wind resistance requirements, enhances the compaction of seeds and soil, resists wind disturbance, and improves the uniformity of seedling emergence.

[0079] The task coordination and scheduling module establishes a real-time communication link between machinery and the cloud, generates collaborative operation instructions and sends them to each machinery unit, and synchronizes the rotary tillage, sowing, covering and compaction processes to achieve seamless integration of the operation process and improve operation efficiency and resource utilization efficiency.

[0080] Example 2, as shown in Figures 1 and 2, based on Example 1, the present invention provides a technical solution: the execution steps of the chain-type soil covering adaptive module include: analyzing surface undulation data based on the growth perturbation digital twin model and the spatial distribution map of soil properties, and using it as the environmental state input of the reinforcement learning agent; calculating and outputting control signals for the swing amplitude and ground contact angle of the soil covering chain based on the surface undulation data according to the reinforcement learning algorithm; driving the soil covering chain mechanism through a servo motor to make it closely fit the uneven ground surface, ensuring the uniformity of the soil covering thickness in the seed furrow and reducing seed exposure;

[0081] The specific work involves: accessing and parsing the multidimensional data stream provided by the digital twin model of growth perturbation and the spatial distribution map of soil properties, with surface undulation data serving as key input information. Through real-time processing and feature extraction of the surface undulation data, a dynamic digital map reflecting the three-dimensional contour of the surface in front of the machinery is constructed, marking changes in absolute elevation and characterizing micro-topographic features including longitudinal slope, lateral dip angle, and local curvature. This transforms the continuous, complex, and uneven physical surface into standardized and structured environmental state parameters that can be identified and understood. After obtaining the structured environmental state input, a reinforcement learning agent is activated. Based on its offline-trained policy network, the reinforcement learning agent performs instantaneous evaluation and decision-making on the current surface undulation data. Through forward computation using a deep neural network, the topographic features, including longitudinal slope, lateral dip angle, and local curvature, are mapped to two key execution parameters for the soil-covered iron chain. The optimal control command is the swing amplitude of the entire chain and its dynamic ground contact angle relative to the ground surface. This is used to solve the optimization problem: under the premise of satisfying the uniformity constraint of soil covering, how to adjust the posture of the actuator to minimize the risk of the soil covering chain being suspended or excessively pressed down due to the undulation of the ground surface. This results in the output of continuous digital control signals, representing the theoretical action values ​​required to achieve the best adhesion effect under the current terrain conditions. The generated digital control signals are sent in real time to the servo drive system of the soil covering mechanism of the rotary tillage seeding machine. After receiving the digital control signals, the servo motor of the servo drive system converts them into precise physical control of the soil covering chain mechanism, adjusting its swing amplitude and ground contact angle. This transforms the originally rigid soil covering operation into a flexible adaptive process, ensuring that the soil raised from the seed furrow can be evenly and continuously backfilled into the seed furrow after sowing, minimizing the direct exposure of seeds and effectively ensuring the uniformity and neatness of seedling emergence.

[0082] The swing amplitude of the soil-covered iron chain is expressed as follows:

[0083] ;

[0084] In the formula: The swing amplitude of the soil-covered iron chain refers to the maximum distance the iron chain swings from the midline to one side; It is the lateral dip angle of the ground surface, that is, the slope perpendicular to the direction of the machine's movement; Local curvature of the Earth's surface, representing the degree to which the Earth's surface bends in the longitudinal or transverse direction; The weighting coefficients, obtained through reinforcement learning training, represent the weights of the influence of the lateral tilt angle and local curvature on the swing amplitude, respectively. The larger the slope, the greater the swing amplitude of the chain needs to conform to the slope and prevent one side from being suspended in the air. Increase; uneven ground ( The larger the chain, the greater the swing required to navigate around or fill the uneven areas. Increase;

[0085] The expression for the angle at which the soil-covered iron chain touches the ground is:

[0086] ;

[0087] In the formula: The dynamic ground contact angle of the soil-covered iron chain refers to the angle between the iron chain and the horizontal reference plane. The longitudinal slope of the ground surface, that is, the slope along the direction of the machine's movement; The curvature compensation coefficient, obtained through reinforcement learning, determines the sensitivity and amplitude of the chain in responding to the bending of the ground. The arctangent function is used to smoothly map the compensation amount to an angle; the base angle of the chain will follow... When going uphill, the front end of the chain lifts up ( Increase); when going downhill, the front end of the chain presses down ( (reduce)

[0088] The execution steps of the dynamic rolling and wind-resistant module include: collecting ambient wind speed in real time through a wind speed sensor integrated on the rotary tillage and seeding machine, combining it with surface undulation data, comprehensively analyzing the rolling requirements, adjusting the downward pressure of the rolling wheel in real time through an electro-hydraulic system based on the evaluation results, and dynamically adjusting the damping coefficient of the spring through a magnetorheological damper, so that the total rolling pressure of the rolling wheel matches the real-time wind resistance requirements, ensuring close contact between the seeds and the soil, and effectively resisting seed displacement or soil erosion caused by wind.

[0089] The specific work involves: continuously collecting real-time environmental wind speed and direction data of the machine's operating area using a high-precision wind speed sensor integrated on the rotary tillage and seeding machinery; simultaneously processing surface undulation data to obtain three-dimensional terrain contour information; and then fusing and analyzing the two types of data to determine the dynamic relationship between wind speed, surface roughness, and required compaction pressure. This comprehensive assessment determines the minimum and optimal compaction intensity required to ensure a sufficiently firm bond between seeds and soil, effectively resisting wind scattering or soil erosion under different wind speed disturbances. After completing the comprehensive demand assessment, a parallel dual-loop adaptive execution mechanism is activated to accurately output the target compaction pressure. The first loop targets the static downward pressure of the compaction wheel: the generated pressure setpoint is sent to the electro-hydraulic system, which precisely adjusts the output pressure of the hydraulic cylinder through a proportional valve, directly and linearly changing the static downward force of the compaction wheel on the ground. The second loop targets dynamic damping compensation: synchronously controlling the... The magnetorheological damper integrated into the suspension system instantly adjusts the viscosity of the magnetorheological fluid within the damper by changing the input current, steplessly altering its damping coefficient. This allows the press wheel assembly to provide cushioning through springs and suppress bounce caused by bumps when encountering uneven terrain, ensuring the press wheel maintains continuous and stable contact with the ground surface under any terrain conditions, preventing press failure due to suspension. Through the synergistic effect of electro-hydraulic control and magnetorheological damping technology, precise and dynamic matching of the press assembly's output force is achieved. Under strong wind conditions, the hydraulic pressure and damping coefficient are increased simultaneously, providing a powerful and stable pressing force to firmly embed the seeds into the soil and compact the surface crust, directly resisting the stripping effect of wind. When the wind weakens or the surface is flat, the output is reduced accordingly to avoid excessive soil compaction and ensure optimal contact and compaction between the seeds and the soil medium, minimizing the risk of seed displacement or surface soil erosion caused by wind.

[0090] The execution steps of the task coordination scheduling module include: establishing a low-latency, high-reliability communication link between the local controller of the rotary tillage and seeding machinery and the cloud server to ensure bidirectional real-time transmission of instructions and data; generating a globally optimized sequence of collaborative operation instructions on the cloud server based on the real-time status and output of each module, covering the start and stop timing and parameters of rotary tillage, seeding, soil covering and compaction; and then distributing the sequence of collaborative operation instructions to the actuators of each mechanical unit to achieve precise synchronization and seamless connection of the integrated operation process and improve the overall operation efficiency of the system.

[0091] The specific work involves establishing a low-latency, highly reliable communication link between the local controller of the rotary tillage and seeding machinery and a remote cloud server by deploying 5G or other communication technologies. This communication link serves as the system's information hub, responsible for full-duplex data exchange: on the one hand, it continuously uploads real-time operation data collected by the machinery's local sensors and the status information of each execution unit to the cloud server; on the other hand, it prepares to receive control commands from the cloud server, ensuring state synchronization between the physical machinery and the cloud server. After obtaining comprehensive and real-time data support, the cloud server comprehensively analyzes multiple constraints such as operation progress, terrain changes, resource competition prediction, and environmental wind speed, aiming at optimizing overall operation efficiency, and generates refined collaborative operation data. The operational instruction sequence specifies the start and stop times for the four stages of rotary tillage, sowing, covering, and compaction, precisely setting the operational parameters for each stage, including rotary tillage depth, sowing density, covering angle, and compaction intensity, forming a highly coordinated digital operation scheme in both time and space. The generated coordinated operation instruction sequence is sent to the local controller of the rotary tillage and sowing machinery through the established communication link. The local controller parses the received coordinated operation instruction sequence into specific, executable commands and promptly distributes them to the corresponding hydraulic actuators, seed metering motors, covering chain servo systems, and compaction wheel electro-hydraulic control units, ensuring precise synchronization of actions and seamless connection of processes across the four operation stages, minimizing operation waiting time and resource waste.

[0092] 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 a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0093] 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, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A collaborative system for integrated sheepgrass planting using intelligent control, comprising a cloud server, characterized in that, The cloud server communication connection includes the following modules: a perception-decision-coordination module, used to comprehensively collect soil parameters to construct a spatial distribution map of soil properties, and to optimize the rotary tillage strategy using cloud collaboration and the fruit fly algorithm to determine the rotary tillage depth for adaptive control of the rotary tillage unit. Soil parameters include soil hardness, salinity, and pH. A root growth prediction module, connected to the perception-decision-coordination module, constructs a digital twin model of growth perturbation to predict the root interaction trend between Leymus chinensis and neighboring plants, guiding the optimization of sowing density and layout. The root growth prediction module includes a perturbation interface modeling unit and a root conflict prediction unit. This unit is used to construct a digital twin model of growth disturbance by combining rotary tillage depth and soil hardness data. It simulates microenvironmental changes at the boundary between disturbed and undisturbed areas and identifies potential growth conflict zones. Specifically, it includes: acquiring the rotary tillage depth setpoint and spatial distribution data of soil hardness output by the perception-decision collaboration module as the initial boundary conditions for the digital twin model; constructing a digital twin model of growth disturbance reflecting changes in water, nutrients, and mechanical resistance at the boundary between disturbed and undisturbed areas based on soil mechanics and plant root growth models; and running the digital twin model of growth disturbance to simulate the extension dynamics and growth trend of Leymus chinensis roots at the interface, visually presenting the potential growth conflict zones. In areas of root growth conflict, the root conflict prediction unit predicts growth conflict areas based on a growth perturbation digital twin model, analyzes the cross-growth trend of Leymus chinensis roots and neighboring plant roots, calculates the competition intensity index, divides different conflict risk zones, and proposes differentiated sowing density adjustment strategies. Specifically, this includes: based on the potential root growth conflict areas output by the growth perturbation digital twin model, analyzing the cross-growth competition trend of Leymus chinensis roots and neighboring plant roots, calculating the competition intensity index, and dividing different conflict risk zones based on a preset conflict threshold; generating a sowing density adjustment strategy based on the analyzed cross-growth competition trend; and adjusting the sowing density. The spatial layout information covered by the strategy and soil property spatial distribution map is integrated into the control unit of the rotary tillage and seeding machinery to form a variable operation prescription map; the chain-type soil covering adaptive module analyzes the surface undulation data in combination with the growth disturbance digital twin model, and controls the swing amplitude and ground contact angle of the soil covering iron chain through reinforcement learning algorithm; the dynamic compaction and wind resistance module is used to adjust the pressure of the compaction wheel and the spring damping in real time according to the wind speed sensor data and the surface undulation data, so that the compaction pressure matches the wind resistance requirements; the task collaborative scheduling module establishes a real-time communication link between the machinery and the cloud, generates collaborative operation instructions and sends them to each mechanical unit to synchronize the rotary tillage, seeding, soil covering and compaction processes.

2. The integrated collaborative operation system for sheepgrass planting using intelligent control as described in claim 1, characterized in that: The perception and decision-making collaboration module includes a soil multi-source perception unit and an optimization decision-making unit. The soil multi-source perception unit collects soil parameters, including soil hardness, salinity, and pH value, in real time through hardness sensors and conductivity probes deployed on the rotary tillage and seeding machinery, and constructs a spatial distribution map of soil properties. The optimization decision-making unit uploads the spatial distribution map of soil properties and the soil parameters it covers to the cloud, uses the fruit fly algorithm to quickly solve the optimal combination of rotary tillage pressing depths, and dynamically adjusts the operating parameters of each rotary tillage unit.

3. The integrated collaborative operation system for sheepgrass planting using intelligent control as described in claim 2, characterized in that: The execution steps of the perception and decision-making collaboration module include: collecting soil parameter data, including soil hardness, salinity and pH value, on the travel path in real time and synchronously through multiple hardness sensors and conductivity probes deployed on the rotary tillage and seeding machinery; constructing a spatial distribution map of soil properties based on the collected soil parameter data through a spatial interpolation algorithm, and dynamically identifying the spatial heterogeneity of soil hardness and salinity.

4. The integrated collaborative operation system for sheepgrass planting using intelligent control as described in claim 2, characterized in that: The execution steps of the optimization decision unit include: uploading the spatial distribution map of soil properties and its corresponding soil parameters to the cloud server, and parsing the real-time soil hardness value of each operation point as the input parameter of the fruit fly algorithm; using the fruit fly algorithm with the root penetration into hard soil layer as the objective function, iteratively optimizing and analyzing the optimal compaction depth corresponding to each rotary tillage unit; sending the obtained optimal compaction depth combination to the local controller, adjusting the hydraulic actuator of each rotary tillage unit, and adaptively matching the rotary tillage depth with soil hardness.

5. The integrated collaborative operation system for sheepgrass planting using intelligent control as described in claim 1, characterized in that: The execution steps of the chain-type soil covering adaptive module include: analyzing surface undulation data based on the growth perturbation digital twin model and the spatial distribution map of soil properties, and using it as the environmental state input of the reinforcement learning agent; calculating and outputting control signals for the swing amplitude and ground contact angle of the soil covering chain based on the surface undulation data according to the reinforcement learning algorithm, and driving the soil covering chain mechanism through a servo motor to make it closely conform to the uneven ground surface.

6. The integrated collaborative operation system for sheepgrass planting using intelligent control as described in claim 1, characterized in that: The execution steps of the dynamic pressing and wind-resistant module include: collecting ambient wind speed in real time through a wind speed sensor integrated on the rotary tillage and seeding machinery, and combining it with surface undulation data to comprehensively analyze the pressing requirements; based on the evaluation results, adjusting the pressing pressure of the pressing wheel in real time through an electro-hydraulic system, and dynamically adjusting the damping coefficient of the spring through a magnetorheological damper, so that the total pressing pressure of the pressing wheel matches the real-time wind-resistant requirements.

7. The integrated collaborative operation system for sheepgrass planting using intelligent control as described in claim 1, characterized in that: The execution steps of the task coordination scheduling module include: establishing a communication link between the local controller of the rotary tillage and seeding machinery and the cloud server to ensure bidirectional real-time transmission of instructions and data; generating a globally optimized sequence of collaborative operation instructions on the cloud server based on the real-time status and output of each module, covering the start and stop timing and parameters of rotary tillage, seeding, covering and compaction, and then distributing the sequence of collaborative operation instructions to the actuators of each mechanical unit.

Images