Multi-parameter adaptive control system and method for combine harvester

Abstract

The present application belongs to the technical field of harvesting equipment control, and provides a multi-parameter adaptive control system and method for a combine harvester. The system is provided on the frame of the combine harvester, and the harvester is provided with a driving mechanism capable of adjusting the walking speed. The system mainly comprises a threshing cylinder, a concave plate, an upper and lower screen assembly, a fan, a detection assembly and a control unit. The threshing cylinder and the concave plate can be adjusted in speed and gap by corresponding driving mechanisms, and the upper and lower screen assembly and the fan can be adjusted in screen opening and air outlet opening. The detection assembly integrates a camera, a load detection unit, a collision sensor, a moisture sensor and a speed sensor, and can collect data on the harvesting operation state in all directions. The control unit can calculate the impurity content and the crushing rate of the harvested material according to the detection data, and adaptively generate control signals for each mechanism. The system can realize multi-parameter collaborative adaptive regulation and control of the harvesting operation, reduce the impurity content and the crushing rate of the harvested material, adapt to different operation conditions, and has strong practicality.

Description

Technical Field

[0001] This invention relates to the field of harvesting equipment control technology, specifically to a multi-parameter adaptive control system and method for a combine harvester. Background Technology

[0002] When combine harvesters operate in the field, their operational quality is typically evaluated using a comprehensive approach, including grain loss rate, grain impurity content in the grain bin, grain breakage rate, and productivity. The loss rate is primarily caused by the threshing and cleaning systems. The impurity content is mainly related to the cleaning airflow, screen opening, and material distribution. The breakage rate is closely related to the threshing drum speed, concave plate gap, and feed rate. Due to the continuous changes in crop variety, maturity, moisture content, and field weed conditions, the threshing and cleaning system exhibits significant nonlinearity, time-varying characteristics, and large time delays. Traditional methods, relying on the driver's experience to manually adjust parameters such as fan speed, upper / lower screen opening, drum speed, concave plate gap, and travel speed, are insufficient to track changes in operating conditions in a timely manner, often resulting in large fluctuations in impurity content and breakage rate, leading to unstable operational quality.

[0003] In the existing technology, there are threshing adjustment schemes for combine harvesters, but a comprehensive adaptive control system covering multiple actuators such as fans, screens, threshing drums, concave plates, and travel speed has not yet been built. Summary of the Invention

[0004] The purpose of this invention is to solve the above-mentioned technical problems and provide a multi-parameter adaptive control system and method for a combine harvester.

[0005] To achieve the above objectives, some embodiments of the present invention provide the following technical solutions: A multi-parameter adaptive control system for a combine harvester, wherein the combine harvester is connected to a harvester drive mechanism, the harvester drive mechanism being capable of controlling the travel speed of the harvester; the combine harvester includes a frame, and the frame is equipped with: granary; Threshing drum: includes a main body and a grid plate disposed on the main body. The threshing drum main body is connected to a threshing drum drive mechanism, which can adjust the rotational speed of the threshing drum. Concave plate: It is spaced apart from the threshing drum and forms a gap with the grid plate of the threshing drum. It is connected to the concave plate drive mechanism, which can adjust the gap between the concave plate and the threshing drum. Upper sieve assembly; located inside the grain bin, facing the discharge port of the threshing drum; connected to an upper sieve assembly adjustment mechanism, the upper sieve assembly adjustment mechanism being able to adjust the opening degree of the upper sieve assembly screen plates; Lower sieve assembly: Located inside the grain bin, it is set below the upper sieve assembly and connected to the lower sieve assembly adjustment mechanism, which can adjust the opening degree of the lower sieve assembly screen. Fan: Its air outlet faces the lower screen assembly; connected to a fan drive mechanism, which is used to adjust the opening of the fan outlet; Detection components: The detection components include a camera installed inside the grain silo, a load detection unit, a collision sensor installed at the straw throwing point, a moisture sensor installed at the grain silo outlet, and a speed sensor; Control unit: Connected to the detection component, the harvester drive mechanism, the threshing drum drive mechanism, the upper screen assembly adjustment mechanism, the lower screen assembly adjustment mechanism, and the blower drive mechanism respectively, the control unit is configured as follows: Based on the detection data of the harvested material in the grain warehouse detected by the detection component, the impurity content and breakage rate of the harvested material are calculated. Based on the impurity content and the breakage rate, control signals are generated for the harvester drive mechanism, the threshing drum drive mechanism, the concave plate drive mechanism, the upper screen assembly adjustment mechanism, the lower screen assembly adjustment mechanism, and the fan drive mechanism.

[0006] In some embodiments of this application, the control unit includes: Data acquisition module: Used to collect data from the detection components; Data preprocessing module: Used to preprocess the data collected by the data acquisition module; Calculation module: Used to calculate the impurity content of the harvested material based on the data processed by the data preprocessing module. and breakage rate ; and according to the impurity content and the breakage rate The system calculates and generates control signals for the harvester drive mechanism, the threshing drum drive mechanism, the concave plate drive mechanism, the upper screen assembly adjustment mechanism, the lower screen assembly adjustment mechanism, and the fan drive mechanism.

[0007] In some embodiments of this application, the detection component further includes: a stroke sensor installed in the upper screen assembly adjustment mechanism, a stroke sensor installed in the lower screen assembly adjustment mechanism, and a stroke sensor installed in the concave plate drive mechanism; The calculation module is configured to: calibrate the stroke function of the upper screen assembly based on the measurement data of the stroke sensor of the upper screen assembly adjustment mechanism and the relationship between the opening degree of the upper screen assembly. Based on the measurement data of the stroke sensor of the lower screen assembly adjustment mechanism and the relationship between the opening degree of the lower screen assembly, the stroke function of the lower screen assembly is calibrated. The positional relationship between the stroke sensor of the concave plate drive mechanism and the gap between the concave plate and the threshing drum is used to calibrate the stroke function of the concave plate drive mechanism. .

[0008] This application also proposes a multi-parameter adaptive control method for a combine harvester, which is based on the multi-parameter adaptive control system for a combine harvester described above, and includes the following steps: S1: Data calibration steps: Perform a self-calibration process on the upper screen assembly adjustment mechanism, the lower screen assembly adjustment mechanism and the concave plate drive mechanism to calibrate the relationship between the stroke of the upper screen assembly adjustment mechanism and the opening of the upper screen assembly, the relationship between the stroke of the lower screen assembly adjustment mechanism and the opening of the lower screen assembly, and the relationship between the stroke of the concave plate drive mechanism and the gap between the concave plate and the threshing drum. S2: Data Acquisition Steps: After the combine harvester starts, the detection component collects the combine harvester's operating data in real time and transmits it to the appropriate department. The operating data of the first control cycle is converted into the first... Multi-source measurement vector for each control cycle Based on the first The image data collected in the control cycle is used to calculate the... Visual credibility of each control cycle ; S3: Based on the first Multi-source measurement vector for each control cycle , No. Quality state vector within each control cycle: and the Visual reliability of each control cycle Calculate the first Quality state vector within each control cycle And calculate the first The first control cycle and the first The difference of the quality state vectors over each control cycle , This indicates that the first [item] was not used. Before periodically collecting data, the predicted first Estimated periodic impurity level; This indicates that the first [item] was not used. Before periodically collecting data, the predicted first... Estimated cycle breakage rate; In order to use the first After periodic data collection, the first Estimated impurity rate for each control cycle In order to use the first After periodic data collection, the first Estimated breakage rate per control cycle; S4: Define the first Control vector for each control cycle ,Will Control vector for each control cycle With the Quality state vector within each control cycle Pair up to get the first Time-delay aligned samples for each control cycle ; S5: Based on the first Quality state vector within each control cycle Aligned samples with time delay Construct the first Sensitivity matrix for each control cycle The sensitivity matrix is ​​used to characterize the correspondence between the control vector and the quality state vector. S6: Based on the first The first control cycle and the first The difference of the quality state vectors over each control cycle , No. Sensitivity matrix for each control cycle Calculate the first Control increment per control cycle ; S7: Based on the first Control increment per control cycle The system generates control commands, which include control signals for the harvester drive mechanism, the threshing drum drive mechanism, the concave plate drive mechanism, the upper screen assembly adjustment mechanism, the lower screen assembly adjustment mechanism, and the fan drive mechanism.

[0009] In some embodiments of this application, the method further includes multi-source measurement vectors. Normalization steps:

[0010] The normalized version Measurement vector for each control cycle, For synthesis operators; In step S2, the normalized measurement vector is used to calculate the first... Visual reliability of each control cycle ; In step S3, the normalized measurement vector and the first... Quality state vector within each control cycle: and the Visual reliability of each control cycle Calculate the first Periodic quality state vector within each control cycle .

[0011] In some embodiments of this application, step S3 includes: Adopting the first The control vector of the control cycle is updated. Quality state estimation vector for each control cycle: , obtained the Quality state vector for each control cycle: ; Based on the Visual reliability of each control cycle Calculate the first Visual clutter noise variance per control cycle and visual damage rate noise variance :

[0012]

[0013] in, To calibrate the baseline impurity content noise variance, The noise variance for calibrating the baseline breakage rate; It is a positive number; Visual impurity noise variance and visual damage rate noise variance Construct the covariance matrix, and use the covariance matrix to calculate the first... Impurity estimate within each control cycle and estimated breakage rate .

[0014] In some embodiments of this application, based on the first Quality state vector within each control cycle Aligned samples with time delay Construct the first Sensitivity matrix for each control cycle ,include: Calculate the first The control cycle is relative to the first Quality status change over each control cycle :

[0015] Calculate the first The control cycle is relative to the first The change in the control vector over each control cycle :

[0016] Constructing a local linear model:

[0017] in, For the first Modeling error term for each control cycle; Obtaining the first based on the local linear model Sensitivity matrix for each control cycle .

[0018] In some embodiments of this application, the method further includes: constructing an updated gain vector. Update the local linear model to obtain the first Sensitivity matrix for each control cycle ,include: . In some embodiments of this application, step S6 includes: Construct a quality error prediction model:

[0019] in, For the first The quality error vector of each control cycle; For the first Sensitivity matrix for each control cycle; For the first The increment vector of the control quantity to be determined for each control cycle; In order to apply The next cycle quality error vector predicted after the prediction; Construct the comprehensive cost objective function :

[0020] With minimization as the optimization objective, For the first Quality error weight matrix for each control cycle For the first The control increment penalty matrix for each control cycle; Constraints are set for optimizing the overall cost objective function, and the overall cost objective function is then optimized based on these constraints. To obtain the first one that satisfies the constraints The optimal control increment solution for each control cycle .

[0021] Compared with the prior art, the embodiments of this application have at least the following technical effects: 1. This invention does not simply rely on a single visual signal from the grain silo or a tail loss signal. Instead, it establishes a quality state space model using multi-source signals such as grain silo visual inspection, tail loss, load, and grain moisture. It also introduces a Kalman filter fusion algorithm that weights image quality assessment with visual confidence. Even under complex lighting, dust, and varying accumulation morphology conditions, it can still obtain smooth and reliable estimates of the "true impurity rate" and "true breakage rate." This reduces random fluctuations and the probability of misjudgment in the quality signal, improving control robustness.

[0022] 2. This invention, based on an online sensitivity matrix, constructs a quadratic cost function centered on the prediction errors of impurity content and breakage rate. It introduces a multi-objective weight matrix and a control increment penalty matrix, unifying the adjustments of the fan, upper / lower screen, drum, concave plate, and travel speed into a multi-objective, multi-constraint optimization problem. By rationally setting weights and constraints, an adjustable dynamic balance between impurity content and breakage rate can be achieved, and new problems caused by over-adjustment of a single parameter can be avoided. Compared with existing methods that adjust parameters one by one through manual experience or simple rules, this invention provides a new control scheme with quantitative trade-offs and overall optimization, maintaining synchronous and stable control of impurity content and breakage rate even when field conditions vary significantly.

[0023] 3. This invention incorporates travel speed as a control variable alongside that of the fan, screen surface, drum, and concave plate into the optimization framework. Under normal operating conditions, it maintains a predetermined efficiency. When the impurity content and breakage rate are detected to be difficult to control within allowable ranges, and other structural parameters are approaching physical limits, the invention guides the calculation results to prioritize reducing travel speed by adjusting the weights of quality error and control increment. This automatically reduces the feed rate and system load, preventing severe blockages and overloads while ensuring that the work quality does not deteriorate. Compared to traditional "manual deceleration after load alarm" or simple engine protection logic, this invention achieves proactive coordination among work quality, work efficiency, and load status, which helps reduce the number of blockage shutdowns and improves effective working time and output efficiency per unit area. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a schematic diagram of the combine harvester structure according to an embodiment of this application; Figure 2 This is a schematic diagram of the main controller structure according to an embodiment of this application; Figure 3 This is a flowchart of the harvester control method according to an embodiment of this application; in: 1. Rack; 2. Threshing drum; 3. Concave plate; 4. Upper sieve assembly; 5. Lower sieve assembly; 6. Fan, 601 air outlet; 7. Deflector plate; 8. Fan drive motor; 9. Threshing drum drive motor; 10. Diagonal plate gap adjustment mechanism; 11. Main controller; 1101. Data acquisition module; 1102. Data preprocessing module; 1103. Calculation module; 12. Fan opening adjustment mechanism. Detailed Implementation

[0026] To make the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the present invention.

[0027] The prefixes such as "first" and "second" used in this application embodiment are merely for distinguishing different descriptive objects and do not limit the position, order, priority, quantity, or content of the described objects. The use of ordinal numbers and other prefixes used to distinguish descriptive objects in this application embodiment does not constitute a limitation on the described objects. The description of the described objects is given in the claims or the context of the embodiments, and should not constitute unnecessary restrictions due to the use of such prefixes. Furthermore, in the description of this embodiment, unless otherwise stated, "multiple" means two or more.

[0028] The technical solutions of the embodiments of this application will be described below with reference to the accompanying drawings. In the description of the embodiments of this application, unless otherwise stated, " / " means "or," for example, A / B can mean A or B; the term "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone.

[0029] In the embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0030] To improve the quality of grain harvesting operations, it is necessary to control the operation of the harvester. Existing control schemes include the following technical means.

[0031] 1. By monitoring signals such as threshing drum load and grain loss, key parameters such as drum speed and concave plate gap are automatically adjusted to improve threshing performance. However, the main objective is to focus on drum load or loss rate, without fully considering comprehensive quality indicators such as impurity content and breakage rate in the grain bin.

[0032] 2. A grain entrainment loss monitoring system and a grain impurity rate and breakage rate monitoring device are set on the threshing and separating device. The control system adjusts parameters such as the rotation speed of the second threshing drum and the angle of the top cover guide strip to enable the combine harvester to operate with a lower entrainment loss rate and breakage rate. However, its adjustment objects are mainly concentrated on the local threshing and separating device, the number of control parameters is relatively limited, and the control strategy mostly adopts preset rules or range adjustment, which makes it difficult to coordinate and optimize multiple actuators.

[0033] 3. A closed-loop control of drum speed based on grain damage observers has been constructed to balance separation loss and grain damage. However, most studies focus on the control of a single actuator (such as the drum), and there is relatively little research on the fusion of multi-source measurement data and the unified trade-off of multiple objectives (loss, impurities, and breakage).

[0034] 4. Some combine harvesters also collect crop information, feed rate, or overall machine load to automatically adjust the threshing drum, guide vanes, straw shredder, and travel speed to keep the overall machine load within a reasonable range. However, this type of solution focuses primarily on load balance and pays insufficient attention to operational quality indicators such as grain impurity rate and grain breakage rate, resulting in situations where "load control is good but operational quality is unstable."

[0035] 5. Regarding comprehensive quality adaptive control, by collecting historical harvesting data, an optimal working range is established between the loss rate or breakage rate and the drum speed and concave plate gap. During operation, the system matches the real-time collected production parameters (production efficiency, loss rate, breakage rate, etc.) with preset correlations. When production parameters are abnormal, the drum speed and concave plate gap are adjusted in the corresponding "low loss rate mode" or "low breakage rate mode". This method primarily relies on static optimal ranges formed by offline historical data for mode switching. The control variables are concentrated on the drum and concave plate parameters, lacking coordinated optimization of multiple parameters such as the cleaning fan, screen opening, and travel speed. Furthermore, a unified dynamic control model is not established for complex time-delay systems.

[0036] In summary, the existing technologies for grain harvester control mainly have the following problems.

[0037] 1. Existing combine harvester operation quality detection devices mainly focus on loss rate or single quality indicators. Although online measurement of grain impurity rate and breakage rate in grain bins has been applied to some extent, it is mostly used for information display or simple alarm, and has not formed a control closed loop that is deeply coupled with multi-parameter automatic adjustment. 2. Existing adaptive adjustment schemes often only adjust the threshing drum, concave plate gap or individual structural parameters. The control algorithms mostly rely on preset rules, static optimal ranges or empirical mappings, making it difficult to take into account the coupling relationship between multiple actuators and the rapid changes in working conditions. 3. There is a significant time lag in the transport of grain quality signals from the threshing and cleaning section to the grain bin. The quality signals fluctuate greatly and their reliability is affected by operating conditions and image quality. Existing solutions generally lack a unified quality state estimation and parameter identification mechanism for time lag and multi-source measurement uncertainties, making it difficult to achieve fine control of impurity content and breakage rate while ensuring stability. 4. For the two inherently contradictory quality indicators of impurity content and breakage rate, existing technologies mostly adopt separate switching control of "low loss rate mode" and "low breakage rate mode", lacking the ability to perform multi-objective collaborative optimization of multiple parameters such as fan, screen surface, drum, concave plate and travel speed under a unified framework.

[0038] To address the above issues, this application proposes a multi-parameter adaptive control system and method for a combine harvester. Using grain impurity content and breakage rate as core feedback indicators, it comprehensively utilizes multi-source information such as grain bin visual inspection, tail loss, load, and moisture content to overcome the time lag and uncertainty in quality feedback. This enables coordinated and adaptive adjustment of multiple operating parameters, including the cleaning fan, screen opening, threshing drum speed, concave plate gap, and travel speed. Consequently, it stably controls impurity content and breakage rate under complex field conditions, improving the overall operational quality and adaptability of the combine harvester.

[0039] The combine harvester described in this application is mainly used for harvesting various granular crops. This equipment integrates harvesting, conveying, threshing, cleaning, and dehulling, enabling simultaneous crop harvesting and grain sorting, as well as shell removal in the field. It has comprehensive operational functions and is suitable for harvesting various granular crops.

[0040] refer to Figure 1 First, the structure and working principle of the combine harvester will be explained.

[0041] The combine harvester includes a frame 1, which serves as the mounting base for all working components of the machine, providing load-bearing, fixing, and positioning for each functional structure. The frame is equipped with core operating mechanisms such as a grain bin (not shown in the attached diagram), a threshing drum 2, a concave plate 3, an upper screen assembly 4, a lower screen assembly 5, a blower 6, and a control unit. These components work together to complete harvesting operations such as threshing, screening, and cleaning of crops. The specific structure and connection relationships are as follows.

[0042] Grain storage: The core cavity structure for the temporary storage and screening of crop grains.

[0043] Threshing drum: includes a main body and a grid plate disposed on the main body. The threshing drum main body is connected to a threshing drum drive mechanism (in this embodiment, a threshing drum drive motor 9). The threshing drum drive mechanism can adjust the rotation speed of the threshing drum. The threshing drum drive mechanism can respond to control commands to realize stepless or multi-level adjustment of the operating speed of the threshing drum, which can adapt to the threshing needs of different crop types, moisture content and harvesting conditions, ensuring the threshing effect while reducing the grain breakage rate.

[0044] Concave plate 3: Spaced apart from the threshing drum, forming a gap with the drum's grid, and connected to a concave plate drive mechanism. This mechanism adjusts the gap between the concave plate and the threshing drum. The concave plate 3 works in conjunction with the threshing drum 2 to complete the threshing process. The gap between the concave plate 3 and the threshing drum 2 can be precisely adjusted via the concave plate drive mechanism to accommodate crops of different particle sizes and straw hardness, effectively improving threshing cleanliness and avoiding problems such as clogging and incomplete threshing.

[0045] For example, in this embodiment of the application, the concave plate is driven to move by the concave plate gap adjustment mechanism 10, thereby realizing the adjustment of the gap between the concave plate 3 and the threshing drum 2.

[0046] For example, the threshing drum 2 is connected to an engine or hydraulic motor via gears, providing the main threshing and frictional action for the combine harvester, and its rotational speed... The adjustment can be achieved through a gearbox or frequency converter / variable control; the concave plate 3 is arranged around the lower side of the drum and adopts a grid structure to separate the grains from the straw. It can be designed as an integral concave plate 3 or a front and rear segmented concave plate 3 as needed; the concave plate drive mechanism includes a swing arm connecting the concave plate 3, a connecting rod, and a set of electric push rods for adjusting the gap. The change in the stroke of the push rods causes the concave plate 3 to swing eccentrically relative to the threshing drum 2 to achieve the adjustment of the concave plate gap. It is adjustable online. In some embodiments, a gap position detection device can also be set, which can be a push rod stroke sensor, to measure the current position and convert it into the corresponding concave plate gap value. This unit is used to adjust the threshing intensity and the degree of compression, and is the main execution object for controlling the crushing rate. It also affects the threshing rate and the entrainment situation.

[0047] Furthermore, the combine harvester is also equipped with a harvester drive mechanism to control the combine harvester's movement in the field and the harvesting of crops.

[0048] The harvester's drive mechanism includes a walking drive system, a travel speed adjustment mechanism, and a feeding and conveying device. The walking drive system consists of a hydraulic walking motor or a mechanical gearbox, and controls retraction and travel speed via a throttle, a travel handle, or an electronically controlled proportional valve. The system includes: a speed control mechanism; a travel speed setting mechanism (in electronically controlled models), where the main controller sends the travel speed setting via the CAN bus interface to automatically decelerate or restore the combine harvester's travel speed; and a feeding and conveying device, including a header conveyor auger, a bridge chain rake, or a conveyor scraper, used to continuously and evenly convey crops to the threshing unit. For models with speed control capabilities, this system can also output the feed chain rake speed setting to coordinate with the combine harvester's travel speed and achieve overall control over the feed amount. This unit primarily acts as a "yield execution object" for overall load and workload, ensuring that quality indicators do not spiral out of control when cleaning and threshing parameters are close to saturation by automatically decelerating or reducing the feed amount.

[0049] The upper screen assembly 4 is located inside the grain bin, facing the discharge port of the threshing drum. It is connected to an adjustment mechanism, which allows for adjustment of the opening degree of the screen plates. The upper screen assembly 4 performs the first round of screening of the threshed mixture. The upper screen assembly 4 is connected to the adjustment mechanism, which allows for flexible adjustment of the screen plate opening degree, changing the material throughput and screening accuracy in the first round of screening to adapt to different material mixing ratios. To facilitate the transport of crops between the threshing drum and the upper screen assembly 4, a guide plate 7 can be installed between them. Crops flowing out of the threshing drum are transported to the upper screen assembly 4 via the guide plate 7.

[0050] Lower Screen Component 5: Located inside the grain bin, below the upper screen component 4, and connected to the lower screen component 5 adjustment mechanism. This mechanism adjusts the opening of the screen plates in the lower screen component 5. The lower screen component 5 performs a secondary, finer screening of the material after screening by the upper screen component 4, further separating impurities from qualified grains. The lower screen component 5 is connected to the lower screen component 5 adjustment mechanism, which allows adjustment of the screen plate opening to match the operating parameters of the upper screen component 4, forming a graded screening system and improving the overall screening and impurity removal effect of the machine.

[0051] For example, the upper screen assembly 4 and the lower screen assembly 5 can adopt a structure similar to louvers. The opening degree of the louvers can be adjusted by a corresponding adjustment mechanism, thereby controlling the volume of the filterable material of the corresponding screen plate.

[0052] For example, the cleaning screen opening control unit mainly includes an upper screen assembly 4, a lower screen assembly 5, an electric push rod for adjusting the opening of the upper / lower screen assembly 5, a stroke sensor, and related mounting brackets. The upper screen assembly 4 is installed in the upper layer of the cleaning chamber and is used for coarse separation of the mixture containing a large amount of straw fragments and ear cob residue. The screen plates are mostly of louvered or fish-scale structure. The lower screen assembly 5 is installed in the lower layer of the cleaning chamber and mainly performs fine screening of the mixture falling from the upper screen assembly 4, obtaining clean grain and a small amount of impurities. The electric push rod for opening adjustment is connected to the opening adjustment mechanisms of the upper screen assembly 4 and the lower screen assembly 5 respectively. By extending and retracting, it changes the opening size of the screen plates, thus controlling the opening of the upper screen assembly 4. 5-degree opening of the lower sieve assembly Continuously adjustable; the stroke sensor is integrated inside or externally connected to the electric push rod to detect the current position of the push rod, converting the stroke signal into an opening signal to achieve a one-to-one mapping between "stroke" and "opening". Under the control of the control system, this unit changes the screen surface throughput and grading characteristics by adjusting the screen opening, thereby affecting the distribution of impurities, clean grains and tailings between the upper and lower screen components 5. It is an important means of controlling the impurity content and loss rate.

[0053] Fan 6: Its air outlet 601 faces the lower screen assembly 5; it is connected to the fan drive motor 8 and the fan opening adjustment mechanism 12, which is used to adjust the opening of the fan outlet 601; the fan 6 can remove light impurities such as straw fragments and dust remaining during the screening process through airflow purging. The fan is connected to the fan drive mechanism, which can precisely adjust the opening and closing degree of the fan outlet 601, thereby changing the air volume and wind speed to adapt to cleaning conditions with different impurity contents and different material moisture.

[0054] In some embodiments of this application, the fan drive mechanism includes a fan drive motor 8, a speed control driver, and an airflow regulation mechanism. The fan is installed at the air inlet or side air outlet of the cleaning chamber, forming a stable airflow field with the air duct and the screen box consisting of the upper screen assembly 4 and the lower screen assembly 5. The fan drive motor 8 is connected to the power output shaft of the combine harvester, and its speed is adjustable via a frequency converter. The speed control driver provides controllable speed commands to the motor, adjusting the fan speed. The air volume is continuously adjusted within the allowable range according to the output of the control system. Depending on the model, the air volume adjustment mechanism can be equipped with dampers, guide vanes, etc., and the damper opening can be changed via a small electric actuator to achieve fine adjustment of localized zones of air volume. The function of this unit is to adjust the fan speed and air volume distribution under the command of the control system to influence the cleaning airflow intensity, thereby controlling the impurity content index of the grain silo.

[0055] Detection components: The detection components include a camera installed inside the grain silo, a load detection unit, a collision sensor installed at the straw throwing point, a moisture sensor, a speed sensor, and a position sensor installed at the grain silo outlet.

[0056] In some embodiments of this application, the camera in the detection component includes an industrial camera component for acquiring images of harvested crops inside a grain silo.

[0057] Examples include industrial cameras, lens assemblies, light sources, mounting brackets, and dustproof protective structures.

[0058] Industrial camera and lens: Installed on the rear wall of the grain silo, the camera's field of view covers the falling grain stream or accumulated grain layer, acquiring images to obtain information on the clean grain appearance within a unit of time. Light source assembly, typically a ring-shaped LED light source, is installed near the camera to provide stable and uniform supplementary lighting to the observation area, reducing the impact of external light variations and shadows. Mounting bracket and protective cover: Used to firmly fix the camera and light source in appropriate positions, employing transparent protective covers, dustproof air curtains, or scraper structures to reduce dust adhesion and impact. This unit, through an embedded vision processing module or image processing program in the main controller, outputs estimated values ​​for the impurity content and breakage rate of the clean grain in the silo, as well as image quality confidence scores, serving as the primary information source for subsequent quality data fusion.

[0059] The collision sensor is used to detect harvest loss; in this embodiment, loss refers to grain loss during cleaning and separation. The collision sensor is installed at the straw throwing point to detect impact signals or reflect the grain loss rate through light, electro-induction, or other methods. The collision sensor is also connected to a signal acquisition circuit, a signal amplification circuit, and a signal conditioning circuit to process the collision signal. The signal conditioning circuit amplifies, filters, and shapes the electrical signal output from the collision sensor, converting it into a digital quantity proportional to the actual loss rate. This unit provides the controller with auxiliary criteria for determining whether the cleaning level is too low or too high, assisting in correcting impurity estimates or adjusting control target weights.

[0060] The load detection unit's data can help provide feedback on the harvester's current operating load level. It can be detected using a torque sensor or a current sensor, and multiple sets can be configured. In this embodiment, the load detection unit is configured as follows.

[0061] The engine torque sensor is used for engine load / torque detection. It reads parameters such as torque, fuel injection quantity, and throttle opening from the engine ECU to reflect the overall load. The threshing drum torque (or a current sensor can be used instead) is installed on the drum drive shaft or motor power circuit to monitor the instantaneous load of the threshing process. The cleaning fan current is detected to indirectly reflect changes in airflow and resistance. The feed conveyor motor current is detected to determine whether the feed is blocked or whether the load is abnormal. These signals combined constitute the load index (Pk), which is used to determine whether the two cassette harvesters are under high load, overload, or light load conditions, providing a basis for weight adjustment and yielding strategies in optimized control.

[0062] Moisture sensors are installed at the cleaning outlet of the grain silo to measure the moisture content of the cleaned grain online.

[0063] Speed ​​sensors, including wheel speed sensors, are used to calculate the working area, working unit length, etc.

[0064] In some embodiments, an environmental monitoring unit is also included for detecting ambient light and dust, using photosensitive elements to reflect the image acquisition environment to assist in assessing visual confidence.

[0065] The output of the detection component, along with the visual quality signal, is input into the quality data fusion and control strategy module, forming the basis of multi-source information.

[0066] Control unit: connected to the detection component, harvester drive mechanism, threshing drum drive mechanism, upper screen assembly 4 adjustment mechanism, and lower screen assembly 5 adjustment mechanism respectively.

[0067] The control unit is configured as follows: Based on the detection data of the harvested grain in the grain warehouse detected by the detection component, the true impurity content of the harvested grain is calculated. and true breakage rate ; Based on the true impurity content and the actual breakage rate The system generates control signals for the harvester drive mechanism, threshing drum drive mechanism, concave plate drive mechanism, upper screen assembly 4 adjustment mechanism, lower screen assembly 5 adjustment mechanism, and fan drive mechanism.

[0068] It should be noted that, in the embodiments of this application, the subscripts of the parameters are... Indicates the first The current control cycle. The current control cycle indicates the [number]th control cycle. The previous control cycle represents the first control cycle. One control cycle.

[0069] Specifically, the control unit includes the main controller. The main controller adopts an automotive-grade embedded control unit, including a multi-core processor, memory, analog / digital input / output ports, and a CAN communication interface. The processor has sufficient computing power to run quality data fusion, online identification, and constraint optimization algorithms, while ensuring that the control cycle meets real-time requirements; the memory includes program memory and data memory, used to store the control program, configuration parameters, and historical operation data; the communication interface exchanges data with the engine ECU, the travel control unit, various sensors, and actuator drive units via the CAN bus.

[0070] The main controller includes a data acquisition module 1101, a data preprocessing module 1102, and a calculation module 1103.

[0071] The data acquisition module 1101 is used to collect data from the detection components. The data acquisition module 1101 consists of multiple data acquisition channels. The analog input channel is used to acquire analog signals such as loss, current, voltage, and moisture; the digital input channel is used to acquire combine harvester travel speed pulses, limit switch status, etc.; and the communication channel is used to receive bus data such as engine load and fan speed feedback.

[0072] The data preprocessing module 1102 is used to preprocess the data acquired by the data acquisition module 1101. Specifically, the data preprocessing module 1102 performs filtering, calibration, and normalization on the data, converting all raw signals into engineering quantity form to provide standardized input for subsequent algorithms.

[0073] The calculation module 1103 executes the calculation of the control instructions of each drive unit, and is used to calculate the impurity content of the harvest based on the data processed by the data preprocessing module 1102. and breakage rate ; and based on the impurity content and breakage rate The system calculates and generates control signals for the harvester drive mechanism, threshing drum drive mechanism, concave plate drive mechanism, upper screen assembly 4 adjustment mechanism, lower screen assembly 5 adjustment mechanism, and fan drive mechanism.

[0074] The computing module 1103 is equipped with a multi-sensor fusion algorithm. The input is multi-source signals such as visual impurity rate, visual breakage rate, loss, load, and moisture, combined with sensor confidence information. Internally, a state-space model is used to estimate the state of impurity rate and breakage rate and suppress noise. The output is the fused impurity rate estimate, breakage rate estimate and corresponding estimate confidence.

[0075] In some embodiments of this application, the main processor further includes a time-delay compensation module and an online identification module. These two modules work together to perform time-correlation analysis on quality feedback and historical control quantities based on the material conveying path structure and average conveying time, and to complete online sensitivity identification. The time-delay compensation module uses preset or online estimated time-delay parameters to establish a mapping between the current quality estimate and the control quantities from several control cycles ago. The online identification module uses recursive least squares or other adaptive identification methods to estimate the local sensitivity matrices of impurity rate and breakage rate on each control quantity in real time. This unit enables the control system to dynamically learn the "parameter-quality" relationship based on the actual operation process, rather than relying on fixed empirical rules.

[0076] In some embodiments of this application, an actuator drive and closed-loop control module is also included, which is responsible for converting the control signals obtained by the calculation module's optimization decision into physical execution actions and monitoring the execution effect in real time. It sets the output speed for motor-type actuators (fan 6, threshing drum 2, feeding conveyor, etc.) and forms a closed-loop control in conjunction with an encoder or speed sensor; it outputs target stroke or angle commands to electric push rod actuators (upper screen plate assembly, lower screen plate assembly, and concave plate adjustment), and achieves position closed-loop control through stroke feedback; it detects the actuator status in real time, including faults such as overload, jamming, and overtravel, and feeds the fault information back to the main controller, triggering protection strategies when necessary. This unit ensures that control commands are executed accurately and provides real execution feedback for online identification and subsequent control.

[0077] Furthermore, the calculation function configuration of the data calculation module is as follows.

[0078] The calculation module is capable of performing calibration functions.

[0079] In some embodiments of this application, the detection component further includes a stroke sensor installed in the adjustment mechanism of the upper sieve assembly 4, a stroke sensor installed in the adjustment mechanism of the lower sieve assembly 5, and a stroke sensor installed in the concave plate drive mechanism.

[0080] The calculation module is configured to: perform position calibration of the combine harvester; and calibrate the stroke function of the upper screen assembly 4 based on the measurement data of the stroke sensor of the upper screen assembly 4 adjustment mechanism and the opening degree relationship of the upper screen assembly 4. Based on the measurement data of the stroke sensor of the lower screen assembly 5 adjustment mechanism and the relationship between the opening degree of the lower screen assembly 5, the stroke function of the lower screen assembly 5 is calibrated. The positional relationship between the stroke sensor of the concave plate drive mechanism and the gap between the concave plate and the threshing drum is used to calibrate the stroke function of the concave plate drive mechanism. .

[0081] The relationship between the position of the upper screen assembly 4 adjustment mechanism and the opening degree of the upper screen assembly 4, the relationship between the position of the lower screen assembly 5 adjustment mechanism and the opening degree of the lower screen assembly 5, and the relationship between the position of the concave plate drive mechanism and the key gap between the concave plate and the threshing drum are obtained.

[0082] Specifically, S1: The data calibration step is executed as follows: the calculation module is configured to perform calibration in the following way.

[0083] Before powering on or during the reset of the combine harvester, the position of the combine harvester should be calibrated.

[0084] Specifically, the main controller performs a self-calibration process on the upper screen assembly 4 adjustment mechanism, the lower screen assembly 5 adjustment mechanism, and the concave plate drive mechanism.

[0085]

[0086]

[0087]

[0088] in, The opening degree of the upper screen component 4. The opening degree of the lower screen component 5. The gap between the concave plate and the threshing drum: In some embodiments, if an electric push rod is used as the corresponding adjustment or drive mechanism, then... , , The stroke position of the corresponding electric actuator is measured by the internal or external stroke sensor of the actuator; the correspondence function between the stroke and the opening / clearance is obtained by least squares fitting or table lookup.

[0089] The calibration points are for the adjustment mechanism of the upper screen assembly 4 to move to the mechanical limit position and several intermediate positions; The calibration points are for adjusting the movement of the lower screen assembly 5 to its mechanical limit position and several intermediate positions; These are the calibration points for the concave plate drive mechanism to move to its mechanical limit position and several intermediate positions.

[0090] During self-calibration, the controller drives the push rod to the mechanical limit, records the minimum and maximum values ​​output by the stroke sensor, and generates calibration points based on the corresponding opening / clearance values ​​given by the mechanical design. These calibration points are then used to fit... and its inverse function And stored in the main controller.

[0091] After calibration, the combine harvester is started. During the harvester's crop collection process, the detection components collect the harvester's operating data in real time and transmit it to the main controller.

[0092] The computing unit is configured to: in each control cycle Construct multi-source measurement vectors based on collected data ,and.

[0093] Specifically, in step S2, the data acquisition step is executed as follows: the computing unit is configured to process the acquired data into multi-source measurement vectors in the following manner.

[0094] After the combine harvester is started, the detection component collects the combine harvester's operating data in real time and converts the operating data into multi-source measurement vectors. ;

[0095] in: For the first Impurity rate estimate for each control cycle, For the first The estimated breakage rate for each control cycle. For the first The collision sensor detection values ​​for each control cycle. For the first Load / torque signal for each control cycle For the first Moisture signals for each control cycle. This is an estimate of the impurity content. The breakage rate estimate can be obtained based on real-time collected data.

[0096] In some embodiments of this application, visual confidence is also included. Calculation steps: Based on the image data acquired by the detection component, calculate the image quality score and normalize it into visual confidence level. In this embodiment of the application, the image data mainly refers to the image data collected by the camera inside the grain warehouse.

[0097]

[0098] The comprehensive score is obtained by weighted summation or other empirical functions based on indicators such as image contrast, number of edges, uniformity of brightness histogram, and proportion of occlusion area. A pre-defined mapping function, such as linear compression or a sigmoid function, is used in the design phase to compress the score to... interval; For the first The reliability of periodic visual signals.

[0099] As an optional step, step S2' is also included: a normalization process for the multi-source measurement vectors. Normalization is performed to transform the raw measurements of the detection components into smoother inputs that can be used for state estimation.

[0100] The computing unit is configured to perform multi-source measurement vector calculations as follows: Normalization processing.

[0101]

[0102] in, The normalized version The measurement vectors of each control cycle are used as inputs to the subsequent fusion algorithm. The normalized measurement vectors are then used to calculate the impurity rate and breakage rate, as well as the control signals of each drive mechanism and adjustment mechanism. The comprehensive operators mainly include median filtering (filtering out high-frequency noise and isolated outliers), linear normalization (mapping each channel to a uniform numerical range, making it easier to set the covariance matrix), and saturation limiting (preventing interference from accidental extreme values ​​of the sensor).

[0103] S3: Based on the first Multi-source measurement vector for each control cycle , No. Quality state vector within each control cycle: and visual credibility Calculate the first Quality state vector within each control cycle and the The first control cycle and the first The difference of the quality state vectors over each control cycle , This indicates that the first [item] was not used. Before periodically collecting data, the predicted first... Estimated periodic impurity level; This indicates that the first [item] was not used. Before periodically collecting data, the predicted first... Estimated cycle breakage rate In order to use the first After periodic data collection, the first Estimated impurity rate for each control cycle In order to use the first After periodic data collection, the first Estimated breakage rate for each control cycle.

[0104] It should be understood that if step S2' is included, then the normalized measurement vector is used to calculate the first... Impurity estimate within each control cycle and estimated breakage rate .

[0105] Specifically, the execution steps of step S3 are as follows.

[0106] Definition of the first Control vector for each control cycle: ,in, For the first The target speed of the fan in each control cycle. For the first The target opening degree of the upper screening component 4 in each control cycle. For the first The target opening degree of the lower screen component 5 in each control cycle. For the first The target rotational speed of the threshing drum for each control cycle For the first The target gap of the concave plate in each control cycle For the first The target speed of the combine harvester for each control cycle. The initial state of the control vector can be obtained through the initialization state in step S1.

[0107] Define the mass state vector: ;in, For the first Impurity rate corresponding to each control cycle For the first Breakage rate corresponding to each control cycle The initial states of impurity rate and damage rate can be obtained through the initialization state in step S1.

[0108] Specifically, the quality status is not a raw quantity directly measured by a single sensor, but an estimated value obtained by fusing multiple sources of data, such as grain bin visual inspection values, tail loss signals, load signals, and moisture signals. It is used to represent the comprehensive quality status of the grain output by the combine harvester under the current operating conditions.

[0109] Adopting the first The control vector of the control cycle is updated. Quality state estimation vector for each control cycle: , obtained the Quality state vector for each control cycle: ,in, This indicates that the first [item] was not used. Before periodically collecting data, the predicted first... Estimated periodic impurity level; This indicates that the first [item] was not used. Before periodically collecting data, the predicted first... Estimated periodic breakage rate.

[0110] Based on the Visual reliability of each control cycle Calculate the visual impurity noise variance for the current control cycle. and visual damage rate noise variance :

[0111]

[0112] in, To calibrate the baseline impurity content noise variance, The noise variance for calibrating the baseline breakage rate; Visual impurity noise variance and visual damage rate noise variance Construct the covariance matrix, and use the covariance matrix to calculate the first... Impurity estimate within each control cycle and estimated breakage rate .

[0113] Specifically, the execution process of step S3 is as follows.

[0114] 1. Define the state equations.

[0115]

[0116] in, For the current period (the 1st cycle) Quality status (each control cycle); For the previous cycle (the 1st) Quality status (each control cycle); The process noise is assumed to be zero-mean Gaussian white noise.

[0117] It should be understood that the quality status here refers to the quality status of the grain output by the combine harvester under the current operating conditions, specifically including the impurity content and breakage rate of the grain in the grain bin after fusion estimation. This quality status differs from the multi-source measurement vectors directly collected by sensors. The sensor measurement vectors mainly include the impurity content estimate, breakage rate estimate, tail loss signal, operating load signal, and grain moisture signal obtained from visual detection of the grain bin. Based on the above multi-source measurement values, the controller obtains a more stable and closer-to-real-operational quality of impurity content and breakage rate through a quality status fusion estimation algorithm.

[0118] 2. Define the observation equation.

[0119]

[0120] in, The observation matrix has dimensions of . Each row represents the linear sensitivity of the corresponding measurement to impurity content and breakage rate, which can be set through offline calibration tests (e.g., changing quality indicators and observing changes in the output of each sensor) or engineering experience; For the first The noise measured during each control cycle is assumed to be zero-mean Gaussian white noise. For the first The noise covariance matrix is ​​measured over each control cycle. initial values ​​of diagonal elements Based on static measurement noise statistical estimation, and then according to Dynamic adjustment.

[0121] 3. Prediction steps.

[0122]

[0123] in, For the current period (the 1st cycle) Prior state estimation (for each control cycle) ; For the previous cycle (the 1st) The posterior state estimate (per cycle). If no more complex dynamic prediction model is added between two adjacent control cycles, the impurity content and breakage rate of the current cycle can be temporarily assumed to be the same as those of the previous cycle (per cycle). Since the values ​​of the previous cycle are similar (the impurity rate and breakage rate after fusion of the previous cycle), the predicted values ​​of the current cycle are directly used as the predicted values ​​of the current cycle.

[0124] Then, once the measurement data for the grain silo, such as visual parameters, losses, load, and moisture content, are received for the current cycle, the controller makes corrections to obtain a "posterior state estimate". :

[0125]

[0126] This is the covariance matrix; a larger value indicates that the state can change more rapidly, and it is generally adjusted through simulation and experimentation. The prior estimate is the error covariance; For the previous cycle (the 1st) (per cycle) posterior estimation error covariance.

[0127] 4. Update steps.

[0128]

[0129] in, For the current period (the 1st cycle) The Kalman gain (per control cycle) represents the weighting of prediction and measurement, i.e., whether to place more trust in the "model prediction" or the "measurement in this cycle". The larger the value, the greater the weight of the current cycle measurement information in the estimation of impurity and breakage rates, and the more the system relies on the current measurement results of the grain warehouse vision and multi-source sensors. The smaller the value, the greater the weight of the prior prediction information, and the more the system relies on the state prediction results of the previous cycle. To measure the noise covariance, according to Dynamic adjustment.

[0130]

[0131] in, For posterior state estimation after fusion; For measurement innovation.

[0132]

[0133] in, For the posterior estimation of the error covariance matrix, according to and To construct. It is an identity matrix.

[0134] For example, , All are baseline noise variances obtained through offline statistics; The impurity content was obtained through offline calibration experiments. Specifically, before the combine harvester operation, several groups of standard wheat grain samples with known impurity and breakage rates were selected, or the true impurity content of the samples was obtained through manual sorting, weighing, and labeling. and true breakage rate Then, under the same camera, light source, installation distance, and shooting angle conditions as the grain warehouse visual inspection module, images were repeatedly acquired for each group of samples, and the corresponding impurity content estimate was output using the visual inspection algorithm. and breakage rate estimate .

[0135] For the For the second calibration sample, the visual detection error is defined as:

[0136] in: For the first Visual detection error of impurity content in secondary calibration samples; This is the estimated impurity content output by the grain warehouse visual inspection module; The true impurity content given for manual sorting, weighing, or standard samples.

[0137]

[0138] in: For the first Visual detection error of breakage rate of secondary calibration samples; This is an estimate of the breakage rate; The actual breakage rate is obtained through manual calibration.

[0139] If it is carried out together After the first effective calibration, the reference noise variance can be calculated using the following formula:

[0140] The baseline noise variance for visual detection of impurity content; To effectively determine the number of calibrated samples; For the first Error in secondary impurity content detection; This represents the average error of all impurity detections.

[0141]

[0142] in: The baseline noise variance for visual inspection of breakage rate; For the first Error in secondary breakage rate detection; This represents the average error of the total breakage rate detection.

[0143] S4: Define the first Control vector for each control cycle ,Will Control vector for each control cycle With the Quality state vector within each control cycle Pairing to obtain time-delay aligned samples .

[0144] Because of the time lag in the transport of grain from the threshing / cleaning device, the quality currently observed reflects the control quantity from several cycles ago. This invention employs an average time lag model:

[0145] in, The delay step is measured in control cycles and is a non-negative integer. The average conveying time of material from the threshing / cleaning unit to the grain bin measurement location can be obtained in the following ways: offline test, where marked particles are fed into the feed inlet and the time to reach the grain bin is measured; or model estimation, which is estimated based on the conveyor belt / chain rake speed, conveying distance, and material filling rate. The control cycle is set during the system design phase.

[0146] In the Period, considered as quality state Mainly corresponds to the first Control quantity implemented periodically That is, forming sample pairs:

[0147] Here It is not a simple general mathematical expression, but a corresponding sample pair of "historical operation parameters - grain warehouse quality results" formed after compensation for the time delay of transportation.

[0148] Specifically Indicates the first A set of combine harvester operation parameters actually issued and executed by the main controller in each control cycle:

[0149] in: : No. Periodic cleaning fan speed; : No. The periodic upper screen component has an opening of 4 degrees; : No. The lower screen assembly is opened to a depth of 5 degrees during the cycle. : No. Cyclic threshing drum speed; : No. The gap between the periodic concave plate and the threshing drum; : No. The travel speed of the periodic combine harvester.

[0150] Because crops need to undergo cleaning, hygienic transport, and lifting processes from the threshing and cleaning area into the grain warehouse, the visual inspection of the grain warehouse is crucial. The impurity rate and breakage rate seen during the cycle are not the current parameters. The immediate result, but mainly the first The results generated by the control parameters in each control cycle. Therefore, this scheme pairs the two:

[0151] The meaning of this sample pair is: in the first... The cycle, the combine harvester based on the fan speed 4-degree opening of the upper screening component 5-degree opening of the lower sieve assembly Drum speed The gap between the concave plate and the threshing drum Driving speed Perform the operation; after the material conveying time lag, the controller is in the... The cycle obtains its corresponding operational quality result, namely the impurity rate, by fusing visual data from the grain depot with multi-source data. and breakage rate .

[0152] Based on this, this scheme constructs a data sample to establish the correspondence between "multi-parameter setpoints of the combine harvester and grain silo quality feedback values". This sample is subsequently used for online sensitivity identification, thereby obtaining the influence relationship of each control parameter on impurity content and breakage rate.

[0153] By introducing time-delay estimation and offset length alignment methods, the current grain silo quality estimate is paired with the control settings from several cycles ago. Based on this, online identification methods such as recursive least squares are used to estimate the local sensitivity matrix of multiple control variables, including impurity content and breakage rate, to fan speed, upper / lower screen opening, drum speed, concave plate gap, and travel speed, in real time. Compared with existing practices that rely on offline experiments to divide "empirical optimal intervals" or fixed rule tables, this invention no longer relies on pre-fixed empirical mappings, but continuously "learns" the real relationship between parameters and quality under the current operating conditions during operation, significantly enhancing the system's adaptive ability to changes in crop variety, moisture content, and feed rate.

[0154] S5: Based on the first Quality state vector within each control cycle Aligned samples with time delay Construct the sensitivity matrix .

[0155] Among them, the sensitivity matrix This matrix is ​​used to characterize the correspondence between the control vector and the quality state vector. Specifically, it describes the impact of changes in control parameters such as the cleaning fan speed, the opening degree of the upper screen assembly 4, the opening degree of the lower screen assembly 5, the threshing drum speed, the gap between the concave plates, and the travel speed on the impurity content and breakage rate of the grain in the grain bin. The first row of this matrix corresponds to the sensitivity of the impurity content to each control parameter, and the second row corresponds to the sensitivity of the breakage rate to each control parameter. This matrix is ​​not a pre-fixed empirical parameter, but is obtained online through time-delay aligned changes in historical control quantities and changes in the grain bin's quality state. It is used to predict the impact of parameter adjustments on operational quality in subsequent multi-objective constrained optimization control.

[0156] Sensitivity matrix This can be specifically expressed as follows:

[0157] The first line indicates the effect of changes in various control parameters on the impurity content. The degree of influence; the second line indicates the effect of changes in each control parameter on the breakage rate. The degree of influence; for example, This indicates the degree of influence of changes in the cleaning fan speed on changes in the impurity content; This indicates the degree of influence of changes in the threshing drum speed on the breakage rate. This indicates the degree of influence of the change in the gap between the concave plates on the change in the breakage rate; , This indicates the degree of influence of changes in the combine harvester's travel speed on the impurity content and breakage rate.

[0158] 1. Calculate the first... The control cycle is relative to the first Quality status change over each control cycle :

[0159]

[0160] Among them, For two adjacent cycles (the first) The first control cycle and the first The change in impurity content estimate (per control cycle) is obtained by subtracting the state estimate from the opening output of the upper sieve component 4 of S4. For two adjacent cycles (the first) The first control cycle and the first The change in breakage rate estimated for each control cycle.

[0161] 2. Calculate the first... The control cycle is relative to the first The change in the control vector over each control cycle :

[0162] In practice, this refers to the six control parameters mentioned above at two adjacent delay alignment periods (the...). The first control cycle and the first The amount of change between control cycles, i.e.:

[0163] This indicates the change in the rotational speed of the cleaning fan; This indicates the change in the opening degree of the lower screen assembly 5; This indicates the change in the opening degree of the lower screen assembly 5; This indicates the change in the rotational speed of the threshing drum; This indicates the change in the gap between the concave plate and the threshing drum; This indicates the change in the speed of the vending machine.

[0164] In the embodiments of this application, Instead This is because the quality results detected by the visual inspection of the grain silo have a transport time lag. (Currently...) Quality changes over each control cycle It mainly corresponds to the previous part. The changes in control parameters around the cycle, not the changes in control parameters that were just issued.

[0165] In this embodiment, the controller reads two sets of operating parameters after delay alignment from the historical records, calculates the changes in control parameters such as fan speed, screen opening, drum speed, concave plate gap, and travel speed, and uses these parameters to analyze the impact of changes in these parameters on the current impurity content and breakage rate.

[0166] 3. Construct a locally linear model:

[0167] in, This is the modeling error term, assumed to have a mean of zero, caused by modeling noise, such as sudden changes in crop density, changes in moisture content, sudden increases in field weeds, and sensor noise. This is the local sensitivity matrix (2×6) under the current operating conditions.

[0168] The physical meaning of the local linear model is that the currently detected changes in impurity content and breakage rate are mainly caused by changes in control parameters such as fan speed, screen opening, drum speed, concave plate gap, and travel speed after delayed alignment.

[0169] Furthermore, based on the local linear model, the local sensitivity matrix can be derived. .

[0170] In some embodiments of this application, in order to obtain a more accurate local sensitivity matrix It also includes: constructing and updating the gain vector Update the local linear model to obtain the sensitivity matrix. .

[0171]

[0172] For the first Periodic parameter updates the gain vector; used to determine the sensitivity matrix of newly added sample pairs in the current period. The correction magnitude is essentially the weighted calculation within the controller: when the current sample provides strong new information about the "control parameter-quality change" relationship, the gain is larger, and the sensitivity matrix is ​​updated more significantly; when historical samples are sufficient or the current sample has low reliability, the gain is smaller, and the sensitivity matrix is ​​only slightly corrected. This gain vector is used to distribute the current quality change prediction residual to the sensitivity coefficients corresponding to each control parameter, thereby enabling online learning of the influence relationships of parameters such as fan speed, screen opening, drum speed, concave plate gap, and travel speed. The covariance matrix of the parameters for the previous period is estimated, with the initial value being a diagonal large number matrix, indicating uncertainty about the initial sensitivity; The forgetting factor (0-1) is set by the designer based on the rate of change of operating conditions; the smaller the value, the more sensitive it is to new data. This represents the transpose of the control increment vector after delay alignment, i.e., converting the column vector into a row vector, to satisfy the matrix multiplication dimension requirement in the recursive least squares update formula. Specifically, this vector includes the cleaning fan speed increment, upper screen opening increment, lower screen opening increment, threshing drum speed increment, concave plate gap increment, and travel speed increment.

[0173] In actual control, the threshing and cleaning process of a combine harvester is non-linear, and the above-mentioned influencing factors will change under different wheat moisture contents, feed rates, and crop densities. Therefore, this invention does not... Instead of being fixed parameters, they are updated online using a recursive least squares method, so that they can reflect the actual impact of each control parameter on the impurity content and breakage rate under the current field conditions.

[0174]

[0175] This can be understood as: Current sensitivity matrix = previous period sensitivity matrix + correction amount obtained from the prediction error of mass change in this period.

[0176] If, during a particular operation, the fan speed, screen opening, drum speed, concave plate clearance, or combine harvester travel speed changes, and subsequently the impurity content and breakage rate detected in the grain silo also change, the system uses this set of "parameter change - quality change" data to update... Gradually learn "which parameter to adjust and how the quality will change under the current working conditions".

[0177] For the previous cycle (the 1st) Sensitivity matrix (-1 control cycle); This represents the fitting residual for this sample.

[0178]

[0179] in, The covariance matrix of the current period parameter estimate is used to measure the uncertainty of the estimate.

[0180] The covariance matrix for parameter estimation in the recursive least squares algorithm is used to characterize the current sensitivity matrix. The estimation uncertainty. This matrix is ​​not used directly as actuator control command, but rather the parameter update gain is calculated in the next control cycle. Time is used as input, thus affecting the sensitivity matrix in the next cycle. The update frequency. Through continuous updates. The controller can adaptively adjust the learning intensity of new samples based on the number and validity of existing samples, so that the influence of control variables such as fan speed, screen opening, drum speed, concave plate gap and travel speed on impurity content and breakage rate gradually converges and adapts to changes in working conditions.

[0181] Initial sensitivity matrix The direction of influence of the fan, screen surface, drum, concave plate, and travel speed on the impurity content and breakage rate can be set according to experience, or it can be set as a zero matrix; the initial parameter estimation covariance matrix Set as ,in For larger positive numbers, The identity matrix represents the significant uncertainty in the sensitivity relationships of various control parameters during the system startup phase, enabling the online identification algorithm to quickly correct the sensitivity matrix based on subsequently collected operational data; the forgetting factor... The adjustment is typically determined based on the rate of change of operating conditions through simulation or field trials, and is usually taken as follows: .

[0182] The "diagonal large number matrix" here is an initialization method used in the recursive least squares algorithm. It represents the state at the beginning of system operation, when the controller's sensitivity to the relationships between various control parameters and quality indicators is still uncertain. It is not a fixed mechanical parameter, nor a sensor measurement value, but rather a numerical matrix used internally by the controller to initialize the online identification algorithm.

[0183] To avoid the term "diagonal large number matrix" being too abstract, it will be written in a more specific way:

[0184] in: : Initial parameter estimation covariance matrix of the recursive least squares algorithm; The initial uncertainty coefficient is set by the system designer based on experimental experience, and is usually taken as a large positive number, for example... ; The identity matrix has the same dimensions as the number of control variables. In this scheme, the control variables include fan speed, upper screen assembly 4 opening, lower screen assembly 5 opening, drum speed, concave plate gap, and travel speed. Therefore, we can take... Identity matrix.

[0185] When the system first started operating, there weren't enough historical samples to determine "how much influence adjusting the speed of the fan, screen, drum, concave plate, and combine harvester would have on the impurity content and breakage rate, respectively." Therefore, the controller initially set the sensitivity uncertainty of each control parameter to be relatively high, so that the newly collected "control parameter change - quality change" samples could quickly correct the sensitivity matrix. .

[0186] In other words, The larger the value, the less the system "distrusts" the initial sensitivity matrix. The larger the update magnitude after a new sample enters; The smaller the value, the more reliable the system believes the initial sensitivity matrix is, and the smaller the correction required by new samples.

[0187] S6: Based on the first The first control cycle and the first The difference of the quality state vectors over each control cycle Sensitivity matrix Calculate the first Control increment per control cycle .

[0188] 1. Construct a quality error prediction model

[0189] in, For the first The quality error vector of each control cycle is calculated from the output state of step S3 and the preset target. For the first The sensitivity matrix for each control cycle is output by S5; For the first For each control cycle, we need to find the control increment vector. In order to apply The predicted first Each control cycle quality error vector.

[0190] 2. Construct the comprehensive cost objective function

[0191] in, Let the overall cost be the objective function, and minimization be the optimization objective. For the first A quality error weight matrix for each control cycle, typically shown below. It can adaptively adjust according to the current working conditions (such as feed rate and moisture content); To control the incremental penalty matrix, it is a diagonal matrix. The values ​​are determined by engineering experience and overall machine dynamic requirements; , To enable setting the limit to increase when impurities exceed the limit Increased when breakage exceeds limits The more you want to control, the more stable the situation becomes. For the first The weighting coefficient of the impurity error in each control cycle is used to represent the degree of importance the controller attaches to the impurity exceedance problem. For the first The weighting coefficient of the breakage rate error in each control cycle is used to indicate the degree of importance the controller attaches to the problem of breakage rate exceeding the limit.

[0192] Constraints are set for the optimization of the comprehensive cost objective function, including magnitude constraints and rate constraints.

[0193] Amplitude constraints:

[0194] in, Indicates the first The first control vector in the control cycle One control component Indicates the first The first control cycle The change in each control component Indicates the first The minimum value of each control component. Indicates the first The maximum value of each control component is determined by the mechanical structure limits, rated speed, and safety regulations (given during the design phase). This is the target control quantity after applying the increment.

[0195] Rate constraints:

[0196] in, For the first The absolute value of the rate of change of each control component; The maximum rate variation in a single cycle is determined by a combination of the actuator's speed capability and the requirement to avoid frequent large-scale parameter adjustments. Furthermore, it is achieved by adjusting... and And, if necessary, introduce additional inequalities to implement the following priority strategy: When the impurity content is severely exceeded and the breakage rate is normal, the penalty for the increase related to the fan, upper screen assembly 4 and lower drying assembly should be appropriately reduced, and the adjustment of the threshing drum and concave plate should be restricted; when the breakage rate is severely exceeded and the impurity content is normal, the weight should be adjusted in the opposite direction; when all structural parameters are close to the limit and still cannot meet the quality requirements, the constraint on the direction of speed reduction should be relaxed, and deceleration should be used as a concession.

[0197] 4. Solve for the overall cost objective function.

[0198]

[0199] in, The optimal control increment solution that satisfies the constraints is calculated by an embedded quadratic programming (QP) solver; This represents the parameter values ​​that minimize the objective function. , , All parameters are determined by the overall machine design and field tests, and recorded in the controller configuration; the solution algorithm calls library functions and is implemented by software engineering.

[0200] S7: Based on the first Control increment vector per control cycle The system generates control commands, including control signals for the harvester drive mechanism, threshing drum drive mechanism, concave plate drive mechanism, upper screen assembly 4 adjustment mechanism, lower screen assembly 5 adjustment mechanism, and blower drive mechanism.

[0201] The control quantity is updated based on the optimal increment, and the target setpoints for each actuator are generated using the S1 calibration function:

[0202] in, For the next cycle (the (each control cycle) target control quantity vector; This is a function that saturates and clips a vector within amplitude and rate constraints to achieve final amplitude limiting. For the current period (the 1st cycle) (each control cycle) target control vector; For this cycle (the 1st) The incremental vector of control quantity obtained by optimization (one control cycle).

[0203] Target stroke of screen surface and concave plate:

[0204]

[0205]

[0206] in, Depend on Take it out from the middle; , Obtained by inverting the S1 calibration function; This is the target stroke sent to the electric linear actuator. The target values ​​for the fan, rollers, and travel system are directly determined by... The corresponding components are generated and converted into inverter frequency, hydraulic valve opening, etc.

[0207] S8: Actuator closed-loop control and status monitoring.

[0208] Each actuator drive unit forms a closed loop based on the target value and feedback signal: Construct a speed closed loop for motor-driven actuators (fans, drums):

[0209] in, The target rotational speed of the wind turbine is generated by S7. The actual rotational speed of the fan is measured by a speed sensor. The error signal is used to adjust the motor drive based on the internal PID controller.

[0210] Constructing a position closed loop for push rod actuators:

[0211] in, The target stroke for the upper (or lower) screen assembly; This is the feedback value from the stroke sensor; This is an error signal used for push rod drive control.

[0212] Simultaneously monitor motor current, push rod speed, stroke limits, etc., and report the fault status to the main controller when jamming or overload is detected. If necessary, suspend the adjustment of parameters of this channel or trigger deceleration / stop protection.

[0213] S9: Data recording and next cycle loop (determine whether to continue to the next cycle based on job status (whether it has ended), fault status, etc. If to continue, return to S2 and execute again.)

[0214] At the end of each control cycle, the controller writes the following data to a circular buffer or external storage: the fusion quality status of the current cycle. Actual control quantity Sensitivity matrix Quality error Some raw sensor data is provided for offline analysis and optimization.

[0215] In existing technologies, key parameters such as fan speed, screen opening, drum speed, concave plate gap, and travel speed rely heavily on repeated trial and error adjustments based on the driver's experience, resulting in significant differences in operational quality between different drivers and different plots of land. This application's embodiment, through quality status fusion, online identification, and multi-parameter optimization control, enables the combine harvester to automatically provide parameter adjustment suggestions based on grain bin quality feedback and directly drive the actuators in closed-loop tracking, significantly reducing the workload of drivers frequently adjusting parameters in the field. Even drivers with limited experience can achieve operational quality levels approaching those of "skilled machine operators" in a shorter time, thereby significantly improving the consistency and replicability of overall machine operational quality. In some embodiments of this application, to facilitate system integration and operation monitoring, a communication and human-machine interaction unit is provided, configured on the display terminal of the combine harvester, to display the current impurity content, breakage rate, loss estimate, set and actual values ​​of various execution quantities, and alarm information, etc.; the operation input device includes a touch screen, buttons, or knobs, allowing the operator to select the control mode, adjust the target quality level, or clear the alarm; the communication bus adopts the CAN bus standard to realize data interaction between this system and the engine ECU, body controller, and remote monitoring platform. This unit enables the system of this invention to have good human-machine interaction and overall machine integration capabilities.

[0216] The combine harvester provided in this application embodiment has a control unit that can receive working condition data collected by the detection component in real time. After data analysis and processing, it controls the actions of each drive and adjustment mechanism accordingly to achieve adaptive and precise adjustment of threshing speed, screening degree, cleaning air volume and working interval, ensuring stable and efficient operation of the combine harvester throughout the process.

[0217] This invention adopts a modular system structure and a general control framework: the quality state fusion module, time delay alignment and online identification module, multi-objective constraint optimization module, and multi-actuator closed-loop control module all have good versatility at the algorithm level. Only some parameters need to be recalibrated according to different combine harvester models and different small-grain crops (such as wheat, rice, barley, etc.) to reuse the overall control approach. Compared with empirical rule schemes built only for specific models and operating conditions, this invention is more suitable as a platform technology for the intelligent upgrade of combine harvesters. It provides a unified integration interface for the subsequent introduction of more sensors and more actuators (such as guide vane angles, secondary cleaning units, etc.), which is beneficial for machine manufacturers to achieve intelligent upgrades and differentiated competition of their products without significantly increasing hardware costs.

[0218] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

[0219] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. It should be noted that any modifications, equivalent substitutions, and improvements made by those skilled in the art within the spirit and principles of the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of this patent application should be determined by the scope of the appended claims.

Claims

1. A multi-parameter adaptive control system for a combine harvester, characterized in that, The combine harvester is connected to a harvester drive mechanism, which controls the harvester's travel speed; the combine harvester includes a frame, on which are mounted: granary; Threshing drum: includes a main body and a grid plate disposed on the main body. The threshing drum main body is connected to a threshing drum drive mechanism, which can adjust the rotational speed of the threshing drum. Concave plate: It is spaced apart from the threshing drum and forms a gap with the grid plate of the threshing drum. It is connected to a concave plate drive mechanism, which can adjust the gap between the concave plate and the threshing drum. Upper sieve assembly; located inside the grain bin, facing the discharge port of the threshing drum; connected to an upper sieve assembly adjustment mechanism, the upper sieve assembly adjustment mechanism being able to adjust the opening degree of the upper sieve assembly screen plates; Lower sieve assembly: Located inside the grain bin, it is set below the upper sieve assembly and connected to the lower sieve assembly adjustment mechanism, which can adjust the opening degree of the lower sieve assembly screen. Fan: Its air outlet faces the lower screen assembly; connected to a fan drive mechanism, which is used to adjust the opening of the fan outlet; Detection components: The detection components include a camera installed inside the grain silo, a load detection unit, a collision sensor installed at the straw throwing point, a moisture sensor installed at the grain silo outlet, and a speed sensor; Control unit: Connected to the detection component, the harvester drive mechanism, the threshing drum drive mechanism, the upper screen assembly adjustment mechanism, the lower screen assembly adjustment mechanism, and the blower drive mechanism respectively, the control unit is configured as follows: Based on the detection data of the harvested material in the grain warehouse detected by the detection component, the impurity content and breakage rate of the harvested material are calculated; Based on the impurity content and the breakage rate, control signals are generated for the harvester drive mechanism, the threshing drum drive mechanism, the concave plate drive mechanism, the upper screen assembly adjustment mechanism, the lower screen assembly adjustment mechanism, and the fan drive mechanism.

2. The multi-parameter adaptive control system for a combine harvester according to claim 1, characterized in that, The control unit includes: Data acquisition module: Used to collect data from the detection components; Data preprocessing module: Used to preprocess the data collected by the data acquisition module; Calculation module: Used to calculate the impurity content of the harvested material based on the data processed by the data preprocessing module. and breakage rate ; and according to the impurity content and the breakage rate The system calculates and generates control signals for the harvester drive mechanism, the threshing drum drive mechanism, the concave plate drive mechanism, the upper screen assembly adjustment mechanism, the lower screen assembly adjustment mechanism, and the fan drive mechanism.

3. The multi-parameter adaptive control system for a combine harvester according to claim 1, characterized in that, The detection assembly further includes: a stroke sensor installed in the upper screen assembly adjustment mechanism, a stroke sensor installed in the lower screen assembly adjustment mechanism, and a stroke sensor installed in the concave plate drive mechanism. The calculation module is configured to: calibrate the stroke function of the upper screen assembly based on the measurement data of the stroke sensor of the upper screen assembly adjustment mechanism and the relationship between the opening degree of the upper screen assembly. Based on the measurement data of the stroke sensor of the lower screen assembly adjustment mechanism and the relationship between the opening degree of the lower screen assembly, the stroke function of the lower screen assembly is calibrated. The positional relationship between the stroke sensor of the concave plate drive mechanism and the gap between the concave plate and the threshing drum is used to calibrate the stroke function of the concave plate drive mechanism. .

4. A multi-parameter adaptive control method for a combine harvester, implemented based on the multi-parameter adaptive control system for a combine harvester as described in any one of claims 1 to 3, characterized in that, Includes the following steps: S1: Data calibration steps: Perform a self-calibration process on the upper screen assembly adjustment mechanism, the lower screen assembly adjustment mechanism and the concave plate drive mechanism to calibrate the relationship between the stroke of the upper screen assembly adjustment mechanism and the opening of the upper screen assembly, the relationship between the stroke of the lower screen assembly adjustment mechanism and the opening of the lower screen assembly, and the relationship between the stroke of the concave plate drive mechanism and the gap between the concave plate and the threshing drum. S2: Data Acquisition Steps: After the combine harvester starts, the detection component collects the combine harvester's operating data in real time and transmits it to the appropriate department. The operating data of the first control cycle is converted into the first... Multi-source measurement vector for each control cycle Based on the first The image data collected in the control cycle is used to calculate the... Visual reliability of each control cycle ; S3: Based on the first Multi-source measurement vector for each control cycle , No. Quality state vector within each control cycle: and the Visual reliability of each control cycle Calculate the first Quality state vector within each control cycle And calculate the first The first control cycle and the first The difference of the quality state vectors over each control cycle , This indicates that the first [item] was not used. Before periodically collecting data, the predicted first... Estimated periodic impurity level; This indicates that the first [item] was not used. Before periodically collecting data, the predicted first Estimated cycle breakage rate; In order to use the first After periodic data collection, the first Estimated impurity rate for each control cycle In order to use the first After periodic data collection, the first Estimated breakage rate for each control cycle; S4: Define the first Control vector for each control cycle , will the Control vector for each control cycle With the Quality state vector within each control cycle Pair up and get the first Time-delay aligned samples for each control cycle ; S5: Based on the first Quality state vector within each control cycle Aligned samples with time delay Construct the first Sensitivity matrix for each control cycle The sensitivity matrix is ​​used to characterize the correspondence between the control vector and the quality state vector. S6: Based on the first The first control cycle and the first The difference of the quality state vectors over each control cycle , No. Sensitivity matrix for each control cycle Calculate the first Control increment per control cycle ; S7: Based on the first Control increment per control cycle The system generates control commands, which include control signals for the harvester drive mechanism, the threshing drum drive mechanism, the concave plate drive mechanism, the upper screen assembly adjustment mechanism, the lower screen assembly adjustment mechanism, and the fan drive mechanism.

5. The multi-parameter adaptive control method for a combine harvester according to claim 4, characterized in that, It also includes multi-source measurement vectors. Normalization steps: The normalized version Measurement vector for each control cycle, For synthesis operators; In step S2, the normalized measurement vector is used to calculate the first... Visual reliability of each control cycle ; In step S3, the normalized measurement vector and the first... Quality state vector within each control cycle: and the Visual reliability of each control cycle Calculate the first Periodic quality state vector within each control cycle .

6. The multi-parameter adaptive control method for a combine harvester according to claim 4 or 5, characterized in that, Step S3 includes: Adopting the first The control vector of the control cycle is updated. Quality state estimation vector for each control cycle: , obtained the Quality state vector for each control cycle: ; Based on the Visual reliability of each control cycle Calculate the first Visual clutter noise variance per control cycle and visual damage rate noise variance : in, To calibrate the baseline impurity content noise variance, The noise variance for calibrating the baseline breakage rate; It is a positive number; Visual impurity noise variance and visual damage rate noise variance Construct the covariance matrix, and use the covariance matrix to calculate the first... Impurity estimate within each control cycle and estimated breakage rate .

7. The multi-parameter adaptive control method for a combine harvester according to claim 4, characterized in that, Based on the Quality state vector within each control cycle Aligned samples with time delay Construct the first Sensitivity matrix for each control cycle ,include: Calculate the first The control cycle is relative to the first Quality status change over each control cycle : Calculate the first The control cycle is relative to the first The change in the control vector over each control cycle : Constructing a local linear model: in, For the first Modeling error term for each control cycle; Obtaining the first based on the local linear model Sensitivity matrix for each control cycle .

8. The multi-parameter adaptive control method for a combine harvester according to claim 7, characterized in that, Also includes: Construct updated gain vector Update the local linear model to obtain the first Sensitivity matrix for each control cycle ,include: 。 9. The multi-parameter adaptive control method for a combine harvester according to claim 4, characterized in that, Step S6 includes: Construct a quality error prediction model: in, For the first The quality error vector of each control cycle; For the first Sensitivity matrix for each control cycle; For the first The increment vector of the control quantity to be determined for each control cycle; In order to apply The next cycle quality error vector predicted after the prediction; Construct the comprehensive cost objective function : With minimization as the optimization objective, For the first Quality error weight matrix for each control cycle For the first The control increment penalty matrix for each control cycle; To optimize the overall cost objective function, constraints are set, and the overall cost objective function is then optimized based on these constraints. To obtain the first one that satisfies the constraints The optimal control increment solution for each control cycle .

Images