Corn harvester energy-saving regulation and control method and system based on reinforcement learning

Abstract

The invention belongs to the technical field of corn harvester management and control, and particularly relates to a corn harvester energy-saving regulation and control method and system based on reinforcement learning, and the system comprises a multi-source heterogeneous data collection module, a reinforcement learning decision module, an action execution regulation and control module, a safety constraint module and an operation control supervision terminal. The multi-source heterogeneous data acquisition module is used for acquiring related data in the operation process of the corn harvester, the reinforcement learning decision module is used for designing a continuous action set based on received data information, and the action execution regulation and control module is used for converting an action instruction into a specific control signal and carrying out automatic adaptive regulation and control on the operation of the corn harvester. And the abnormal operation risk of the corn harvester is reduced by introducing a safety constraint mechanism, the whole-process intelligent control of the corn harvesting operation from environment perception to parameter optimization is realized, the safe, stable and efficient energy-saving operation of the corn harvester is ensured, the intelligent and automatic level is high, and the operation management and control difficulty is small.

Description

Technical Field

[0001] The present invention relates to the technical field of corn harvester control, and specifically to an energy-saving regulation method and system for corn harvesters based on reinforcement learning. Background Art

[0002] A corn harvester is an agricultural machinery and equipment specifically used for corn harvesting operations. Its core functions are to efficiently and accurately complete the harvesting, threshing, and cleaning of corn plants, so as to collect corn kernels from the field and prepare for subsequent storage, processing, or sales.

[0003] Traditional corn harvesters usually adopt a fixed-parameter control mode, which cannot adaptively adjust the operating state according to real-time changes such as field conditions and crop density, resulting in serious energy waste. Moreover, they cannot reasonably analyze and timely warn about the operation performance and risk of the constraint mechanism of the corn harvester, which is not conducive to ensuring the safe, stable, efficient, and energy-saving operation of the corn harvester, and the operation supervision is difficult and the intelligent level is low.

[0004] In view of the above technical deficiencies, a solution is proposed now. Summary of the Invention

[0005] The purpose of the present invention is to provide an energy-saving regulation method and system for corn harvesters based on reinforcement learning, which solves the problems that the prior art cannot adaptively adjust the operating state according to real-time changes such as field conditions and crop density, and cannot reasonably analyze and timely warn about the operation performance and risk of the constraint mechanism of the corn harvester, which is not conducive to ensuring the safe, stable, efficient, and energy-saving operation of the corn harvester, and the operation supervision is difficult and the intelligent level is low.

[0006] To achieve the above purpose, the present invention provides the following technical solutions:

[0007] An energy-saving regulation system for corn harvesters based on reinforcement learning includes a multi-source heterogeneous data acquisition module, a reinforcement learning decision-making module, an action execution regulation module, a safety constraint module, and an operation control and supervision terminal; the multi-source heterogeneous data acquisition module real-time collects the operating environment and state parameters of the corn harvester, performs spatio-temporal alignment and noise filtering on the original data, and generates a three-dimensional state vector including a crop density distribution map, a terrain slope, and the current fuel consumption rate.

[0008] The reinforcement learning decision-making module maps the three-dimensional state vector output by the multi-source heterogeneous data acquisition module into a discretized state encoding, designs a continuous action set including engine speed regulation, cutter bar height adjustment, and threshing cylinder speed optimization, and the action execution regulation module converts the action instruction output by the reinforcement learning decision-making module into a specific control signal to automatically adaptively regulate the operation of the corn harvester.

[0009] The safety constraint module pre-sets several groups of risk matters to be monitored in advance. If corresponding risk matters are detected during the operation of the corn harvester, corresponding safety constraint operations are taken, and constraint alarm information is generated and sent to the operation control and supervision terminal. The operation control and supervision terminal displays and warns about the constraint alarm information.

[0010] Furthermore, the multi-source heterogeneous data acquisition module scans the crop row spacing and density distribution through lidar. The lidar obtains the crop canopy point cloud data in a 360° rotation scanning mode, eliminates weed interference points through the voxel grid filtering algorithm, and identifies the crop row spacing and local density distribution based on the density clustering algorithm to generate a density heat map with a spatial resolution of 0.1m×0.1m.

[0011] The body attitude of the corn harvester is monitored through an inertial measurement device. The pitch angle, roll angle, and acceleration changes of the body are monitored at a sampling frequency of 500Hz, and the multi-sensor data is fused using the Kalman filter algorithm to construct an attitude feature vector including the terrain slope and the body vibration amplitude.

[0012] The load change is obtained through the engine torque sensor, the instantaneous torque of the crankshaft is collected in real time, and the load fluctuation coefficient is calculated in combination with the ring gear speed signal; and the operation trajectory is recorded through the GPS positioning module.

[0013] Furthermore, the action execution and regulation module sends the engine speed adjustment instruction to the ECU via the CAN bus, and uses the feedforward-feedback composite control to reach the target speed within 200ms. The cutting table height adjustment instruction drives the hydraulic proportional valve through the PWM signal, and cooperates with the displacement sensor to achieve closed-loop control with a positioning accuracy of ±2mm. And the threshing cylinder torque limit instruction realizes the torque soft limit by changing the current of the clutch solenoid valve.

[0014] Furthermore, the operation control and supervision terminal is communicatively connected to the harvesting continuity detection module. The harvesting continuity detection module analyzes the harvesting continuity execution status of the corn harvester per unit time, generates a continuity qualified signal or a continuity abnormal signal through the analysis, and sends the continuity qualified signal or the continuity abnormal signal to the operation control and supervision terminal. When the operation control and supervision terminal receives the continuity abnormal signal, it issues a corresponding warning.

[0015] Furthermore, the specific analysis process of the harvesting continuity detection module is as follows:

[0016] The engine speed, header height, and threshing cylinder speed of the corn harvester are obtained. The difference between the engine speed and the currently set standard engine speed value is calculated and the absolute value is taken to obtain the engine detection value. Similarly, the header detection value and the threshing cylinder detection value are obtained. The engine detection value, header detection value, and threshing cylinder detection value are respectively compared with the preset engine detection threshold, preset header detection threshold, and preset threshing cylinder detection threshold. If the engine detection value, header detection value, or threshing cylinder detection value exceeds the corresponding preset threshold, it is determined that the corn harvester is in a non-optimal harvesting state;

[0017] The total duration of the corn harvester being in a non-optimal harvesting state within a unit time is obtained and marked as the non-optimal harvesting characteristic value. The non-optimal harvesting characteristic value is compared with the preset non-optimal harvesting characteristic threshold. If the non-optimal harvesting characteristic value exceeds the preset non-optimal harvesting characteristic threshold, a continuous abnormal signal is generated.

[0018] Furthermore, if the non-optimal harvesting characteristic value does not exceed the preset non-optimal harvesting characteristic threshold, the engine detection value, header detection value, and threshing cylinder detection value are weighted and summed to obtain the harvester operation measurement value. The average value of all harvester operation measurement values within a unit time is calculated to obtain the harvester operation analysis value, and the maximum harvester operation measurement value within a unit time is marked as the harvester abnormal value;

[0019] The continuous evaluation value is obtained by weighted summing the non-optimal harvesting characteristic value, harvester operation analysis value, and harvester abnormal value. The continuous evaluation value is compared with the preset continuous evaluation threshold. If the continuous evaluation value exceeds the preset continuous evaluation threshold, a continuous abnormal signal is generated; if the continuous evaluation value does not exceed the preset continuous evaluation threshold, a continuous qualified signal is generated.

[0020] Furthermore, the harvesting continuity detection module is communicatively connected to the constraint effectiveness evaluation module. The harvesting continuity detection module sends the continuous qualified signal to the constraint effectiveness evaluation module. When the constraint effectiveness evaluation module receives the continuous qualified signal, it evaluates and analyzes the safety constraint effectiveness of the corn harvester within a unit time, generates a constraint qualified signal or a constraint potential hazard signal through the analysis, and sends the constraint qualified signal or the constraint potential hazard signal to the operation control supervision terminal. When the operation control supervision terminal receives the constraint potential hazard signal, it issues a corresponding warning.

[0021] Furthermore, the specific analysis process of the constraint effectiveness evaluation module includes:

[0022] All risk events that occur during the operation of a corn harvester within a unit time are obtained. The occurrence moment of the corresponding risk event is marked as the first moment, the moment when the corresponding constraint operation is made is marked as the second moment, and the time difference between the second moment and the first moment is calculated to obtain a constraint delay value. The constraint delay value is numerically compared with the corresponding preset constraint delay threshold. If the constraint delay value exceeds the corresponding preset constraint delay threshold, a constraint invalid symbol TW-1 is assigned to the corresponding risk event;

[0023] The number of times the constraint invalid symbol TW-1 is assigned within a unit time is obtained and the ratio is calculated with the total number of occurrences of the risk event to obtain a constraint invalid detection value. The constraint invalid detection value is numerically compared with the preset constraint invalid detection threshold. If the constraint invalid detection value exceeds the preset constraint invalid detection threshold, a constraint hidden danger signal is generated;

[0024] If the constraint invalid detection value does not exceed the preset constraint invalid detection threshold, the ratio of the constraint delay value to the corresponding preset constraint delay threshold is calculated to obtain a constraint delay measurement value, and the average value of all constraint delay measurement values within a unit time is calculated to obtain a constraint delay analysis value, and the maximum constraint delay measurement value within a unit time is marked as a constraint delay risk value;

[0025] By calculating the weighted sum of the constraint invalid detection value, the constraint delay analysis value, and the constraint delay risk value to obtain a constraint effectiveness evaluation value, the constraint effectiveness evaluation value is numerically compared with the preset constraint effectiveness evaluation threshold. If the constraint effectiveness evaluation value exceeds the preset constraint effectiveness evaluation threshold, a constraint hidden danger signal is generated; if the constraint effectiveness evaluation value does not exceed the preset constraint effectiveness evaluation threshold, a constraint qualified signal is generated.

[0026] Furthermore, the present invention also proposes an energy-saving regulation method for a corn harvester based on reinforcement learning, including the following steps:

[0027] Step 1: Collect the operating environment and state parameters of the corn harvester and perform data processing;

[0028] Step 2: Design a continuous action set for the corn harvester;

[0029] Step 3: Generate specific control signals based on action instructions to automatically and adaptively regulate the operation of the corn harvester;

[0030] Step 4: Monitor risk events of the corn harvester. If a corresponding risk event is detected during the operation of the corn harvester, go to Step 5;

[0031] Step 5: Make corresponding safety constraint operations and generate constraint alarm information.

[0032] Compared with the prior art, the beneficial effects of the present invention are:

[0033] 1. In the present invention, by collecting relevant data during the operation of the corn harvester, reasonably designing a set of continuous actions, converting action instructions into specific control signals, automatically adaptively regulating the operation of the corn harvester, and introducing a safety constraint mechanism to reduce the risk of abnormal operations of the corn harvester, the full-process intelligent control from environmental perception to parameter optimization of the corn harvesting operation is achieved, which is beneficial to ensuring the safe, stable, efficient and energy-saving operation of the corn harvester.

[0034] 2. In the present invention, through the harvesting continuity detection module, the harvesting execution performance of the corn harvester is reasonably analyzed and comprehensively evaluated. When generating a continuity qualified signal, the constraint effectiveness evaluation module accurately evaluates the timeliness of the response of the constraint mechanism of the corn harvester, which is beneficial for management personnel to promptly conduct cause investigation and analysis and make targeted improvement measures, further ensuring the high-efficiency, stable and energy-saving operation of the corn harvester, significantly reducing the operation supervision difficulty of the corn harvester, and having a high level of intelligence and automation. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] For the convenience of those skilled in the art to understand, the present invention will be further described below in conjunction with the accompanying drawings.

[0036] Figure 1 It is the system block diagram of Embodiment 1 in the present invention.

[0037] Figure 2 It is the system block diagram of Embodiments 2 and 3 in the present invention.

[0038] Figure 3 It is the method flowchart of Embodiment 4 in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0039] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0040] Embodiment 1: As Figure 1 shown, a corn harvester energy-saving regulation system based on reinforcement learning proposed by the present invention includes a multi-source heterogeneous data acquisition module, a reinforcement learning decision module, an action execution regulation module, a safety constraint module and an operation control and supervision terminal.

[0041] The multi-source heterogeneous data acquisition module collects the operating environment and status parameters of the corn harvester in real time (as the system perception layer, three groups of threads for lidar scanning, inertial navigation monitoring, and engine condition acquisition are started synchronously), performs spatio-temporal alignment and noise filtering on the raw data, and generates a three-dimensional state vector including the crop density distribution map, terrain slope, and current fuel consumption rate. That is, after spatio-temporal alignment of the three groups of raw data, a three-dimensional state representation including crop density encoding, terrain complexity score, and current fuel consumption rate is generated through a feature splicing network.

[0042] Specifically, the multi-source heterogeneous data acquisition module scans the crop row spacing and density distribution through lidar. The lidar obtains the crop canopy point cloud data in a 360° rotation scanning mode, eliminates weed interference points through the voxel grid filtering algorithm, and identifies the crop row spacing and local density distribution based on the density clustering algorithm to generate a density heat map with a spatial resolution of 0.1m×0.1m.

[0043] The inertial measurement unit (IMU) is used to monitor the body attitude of the corn harvester. The six-axis IMU monitors the body pitch angle, roll angle, and acceleration changes at a sampling frequency of 500Hz, and uses the Kalman filter algorithm to fuse multi-sensor data to construct an attitude feature vector including terrain slope (accuracy ±0.5°) and body vibration amplitude (RMS value). The load change is obtained through the engine torque sensor, the instantaneous torque of the crankshaft is collected in real time, and the load fluctuation coefficient is calculated in combination with the ring gear speed signal. And the operation trajectory is recorded through the GPS positioning module.

[0044] The reinforcement learning decision module maps the three-dimensional state vector output by the multi-source heterogeneous data acquisition module into a discretized state encoding, including dimensions such as crop density level (low / medium / high), terrain slope angle (-15° to 15°), and current fuel consumption rate interval (0-5L / min). A continuous action set including engine speed regulation, cutter bar height adjustment, and threshing cylinder speed optimization is designed. The action execution and regulation module converts the action instructions output by the reinforcement learning decision module into specific control signals to automatically and adaptively regulate the operation of the corn harvester.

[0045] Among them, the action execution and regulation module sends the engine speed regulation instruction to the ECU via the CAN bus, and uses a feedforward-feedback composite control to reach the target speed within 200ms. The cutter bar height adjustment instruction drives the hydraulic proportional valve through a PWM signal, and cooperates with the displacement sensor to achieve closed-loop control with a positioning accuracy of ±2mm. And the threshing cylinder torque limit instruction realizes torque soft limit by changing the current of the clutch solenoid valve.

[0046] The safety constraint module pre-sets several groups of risk items to be monitored (i.e., items that need to be monitored and constrained, such as the constraint of operating speed, the constraint of cutter bar height, etc. For the constraint of cutter bar height, it is necessary to ensure that the height is not less than 80% of the crop plant height to avoid missed harvesting). If it is detected that the corn harvester has corresponding risk items during operation, corresponding safety constraint operations are taken, and a constraint alarm message is generated and sent to the operation control and supervision terminal. The operation control and supervision terminal displays and warns of the constraint alarm message, and the safety constraint mechanism reduces the abnormal operation risk of the corn harvester.

[0047] Embodiment 2: As Figure 2 shown, the difference between this embodiment and Embodiment 1 is that the operation control and supervision terminal is communicatively connected to the harvesting continuity detection module. The harvesting continuity detection module analyzes the harvesting continuity execution status of the corn harvester within a unit time, generates a continuity qualified signal or a continuity abnormal signal through the analysis, and sends the continuity qualified signal or the continuity abnormal signal to the operation control and supervision terminal;

[0048] When the operation control and supervision terminal receives the continuity abnormal signal, it issues a corresponding warning, which can reasonably analyze and comprehensively evaluate the harvesting execution performance of the corn harvester, facilitating the management personnel to timely conduct cause investigation and analysis and make targeted improvement measures to ensure the efficient, stable and energy-saving operation of the corn harvester; The specific analysis process of the harvesting continuity detection module is as follows:

[0049] Obtain the engine speed, cutter bar height and threshing cylinder speed of the corn harvester, calculate the difference between the engine speed and the currently set standard engine speed value and take the absolute value to obtain the engine detection value, calculate the difference between the cutter bar height and the currently set standard cutter bar height value and take the absolute value to obtain the cutter bar detection value, and calculate the difference between the threshing cylinder speed and the currently set standard threshing cylinder speed value and take the absolute value to obtain the threshing cylinder detection value;

[0050] Compare the engine detection value, cutter bar detection value and threshing cylinder detection value with the preset engine detection threshold, preset cutter bar detection threshold and preset threshing cylinder detection threshold respectively. If the engine detection value, cutter bar detection value or threshing cylinder detection value exceeds the corresponding preset threshold, it is determined that the corn harvester is in a non-optimal harvesting state;

[0051] Obtain the total duration of the corn harvester being in a non-optimal harvesting state within a unit time and mark it as the non-optimal harvesting characteristic value. Compare the non-optimal harvesting characteristic value with the preset non-optimal harvesting characteristic threshold. If the non-optimal harvesting characteristic value exceeds the preset non-optimal harvesting characteristic threshold, it indicates that the harvesting execution status of the corn harvester within a unit time is poor, which is not conducive to ensuring the efficient, stable and energy-saving operation of the harvester, and then a continuity abnormal signal is generated.

[0052] Furthermore, if the number of harvested non-ideal eigenvalues does not exceed the preset threshold for harvesting non-ideal eigenvalues, the engine detection value, the header detection value, and the threshing cylinder detection value are weighted and summed to obtain the harvester operation measurement value. That is, corresponding preset weight coefficients are assigned to the engine detection value, the header detection value, and the threshing cylinder detection value respectively. The engine detection value, the header detection value, and the threshing cylinder detection value are multiplied by the corresponding preset weight coefficients respectively, and the sum of the three product results is marked as the harvester operation measurement value;

[0053] It should be noted that the larger the value of the harvester operation measurement value, the worse the overall real-time operation performance of the corn harvester; the mean value of all harvester operation measurement values within a unit time is calculated to obtain the harvester operation analysis value, and the harvester operation measurement value with the largest value within a unit time is marked as the harvester abnormal value;

[0054] The continuous evaluation value is obtained by weighted summation calculation of the harvested non-ideal eigenvalue, the harvester operation analysis value, and the harvester abnormal value; that is, corresponding preset weight coefficients are assigned to the harvested non-ideal eigenvalue, the harvester operation analysis value, and the harvester abnormal value respectively. The harvested non-ideal eigenvalue, the harvester operation analysis value, and the harvester abnormal value are multiplied by the corresponding preset weight coefficients respectively, and the sum of the three product results is marked as the continuous evaluation value;

[0055] It should be noted that the larger the value of the continuous evaluation value, the worse the overall harvesting execution status of the corn harvester within a unit time, and it is less conducive to ensuring the efficient, stable, and energy-saving operation of the harvester; the continuous evaluation value is compared with the preset continuous evaluation threshold. If the continuous evaluation value exceeds the preset continuous evaluation threshold, it indicates that the overall harvesting execution status of the corn harvester within a unit time is poor and it is not conducive to ensuring the efficient, stable, and energy-saving operation of the harvester, then a continuous abnormality signal is generated; if the continuous evaluation value does not exceed the preset continuous evaluation threshold, it indicates that the overall harvesting execution status of the corn harvester within a unit time is good, then a continuous qualified signal is generated.

[0056] Embodiment 3: As Figure 2 shown, the difference between this embodiment and Embodiment 1 and Embodiment 2 is that the harvesting continuity detection module is communicatively connected to the constraint effectiveness evaluation module. The harvesting continuity detection module sends the continuous qualified signal to the constraint effectiveness evaluation module. When the constraint effectiveness evaluation module receives the continuous qualified signal, it evaluates and analyzes the safety constraint effectiveness of the corn harvester within a unit time, and generates a constraint qualified signal or a constraint hidden danger signal through the analysis;

[0057] And send the constraint qualified signal or constraint hidden danger signal to the operation control and supervision terminal. When the operation control and supervision terminal receives the constraint hidden danger signal, it issues a corresponding warning, which can reasonably analyze and comprehensively evaluate the timeliness of the response of the constraint mechanism of the corn harvester, facilitating the management personnel to conduct cause investigation and analysis in a timely manner and make targeted improvement measures, further ensuring the efficient, stable and energy-saving operation of the corn harvester, significantly reducing the operation supervision difficulty of the corn harvester, and having a high level of intelligence and automation; The specific analysis process of the constraint effectiveness evaluation module includes:

[0058] Obtain all risk events that occur during the operation of the corn harvester per unit time. Mark the occurrence moment of the corresponding risk event as the first moment, mark the moment when the corresponding constraint operation is made as the second moment, and calculate the time difference between the second moment and the first moment to obtain the constraint delay value; Among them, the larger the value of the constraint delay value, the less timely the constraint for the corresponding risk event.

[0059] Compare the constraint delay value with the corresponding preset constraint delay threshold. If the constraint delay value exceeds the corresponding preset constraint delay threshold, indicating that the constraint for the corresponding risk event is not timely, then assign the constraint invalid symbol TW-1 to the corresponding risk event.

[0060] Obtain the number of times the constraint invalid symbol TW-1 is assigned per unit time and calculate the ratio with the total number of occurrences of risk events to obtain the constraint invalid detection value. Compare the constraint invalid detection value with the preset constraint invalid detection threshold. If the constraint invalid detection value exceeds the preset constraint invalid detection threshold, indicating that the constraint risk for the safety constraint of the corn harvester is relatively high and is not conducive to ensuring the safe and stable operation of the corn harvester, then generate a constraint hidden danger signal.

[0061] If the constraint invalid detection value does not exceed the preset constraint invalid detection threshold, then calculate the ratio of the constraint delay value to the corresponding preset constraint delay threshold to obtain the constraint delay measurement value, calculate the average value of all constraint delay measurement values per unit time to obtain the constraint delay analysis value, and mark the maximum constraint delay measurement value per unit time as the constraint delay risk value.

[0062] Calculate the constraint effectiveness evaluation value by weighted summing the constraint invalid detection value, the constraint delay analysis value and the constraint delay risk value; that is, assign corresponding preset weight coefficients to the constraint invalid detection value, the constraint delay analysis value and the constraint delay risk value, multiply the constraint invalid detection value, the constraint delay analysis value and the constraint delay risk value by the corresponding preset weight coefficients respectively, and mark the sum value of the three product results as the constraint effectiveness evaluation value.

[0063] It should be noted that the larger the value of the constraint effectiveness evaluation value, the higher the comprehensive constraint risk of the safety constraint module for the corn harvester; comparing the constraint effectiveness evaluation value with the preset constraint effectiveness evaluation threshold numerically, if the constraint effectiveness evaluation value exceeds the preset constraint effectiveness evaluation threshold, it indicates that the comprehensive constraint risk for the corn harvester is relatively high, and then a constraint hidden danger signal is generated; if the constraint effectiveness evaluation value does not exceed the preset constraint effectiveness evaluation threshold, it indicates that the comprehensive constraint risk for the corn harvester is relatively low, and then a constraint qualified signal is generated.

[0064] Embodiment 4: As Figure 3 shown, the difference between this embodiment and Embodiment 1, Embodiment 2, and Embodiment 3 is that a corn harvester energy-saving regulation method based on reinforcement learning proposed by the present invention includes the following steps:

[0065] Step 1, collect the operating environment and state parameters of the corn harvester and perform data processing;

[0066] Step 2, design a continuous action set for the corn harvester;

[0067] Step 3, generate specific control signals based on the action instructions to automatically and adaptively regulate the operation of the corn harvester;

[0068] Step 4, monitor the risk matters of the corn harvester. If it is detected that the corn harvester has corresponding risk matters during the operation process, then go to Step 5;

[0069] Step 5, make corresponding safety constraint operations and generate constraint alarm information.

[0070] The working principle of the present invention: When in use, the relevant data during the operation of the corn harvester is collected through the multi-source heterogeneous data collection module. The reinforcement learning decision-making module designs a continuous action set based on the received data information. The action execution and regulation module converts the action instructions into specific control signals to automatically and adaptively regulate the operation of the corn harvester. And by introducing a safety constraint mechanism to reduce the abnormal operation risk of the corn harvester, the full-process intelligent control from environmental perception to parameter optimization of the corn harvesting operation is realized, which is beneficial to ensuring the safe, stable, efficient and energy-saving operation of the corn harvester, significantly reducing the operation management and control difficulty of the corn harvester and improving its intelligence and automation level.

[0071] In the technical solution of the present invention, the setting of the threshold value, preset value, preset range, etc. is for the result comparison and analysis to determine whether it is good or bad. Regarding the value selection of these, it is set and stored by combining the large model analysis of sample data and manual experience, and can also be appropriately adjusted according to seasonal or regular influencing conditions. Moreover, the setting of the preset weight coefficient, influencing factor, etc. is based on the influence degree of each parameter on the result to allocate specific values to ultimately reflect the influence situation on the result. It is also set and stored by combining the large model analysis of sample data and manual experience, and can also be appropriately adjusted according to seasonal or regular influencing conditions.

[0072] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the present invention to only the specific implementation manners. Obviously, according to the content of this specification, many modifications and changes can be made. These embodiments are selected and specifically described in this specification to better explain the principle and practical application of the present invention, so that those skilled in the relevant technical field can well understand and utilize the present invention. The present invention is only limited by the claims and their full scope and equivalents.

Claims

1. An energy-saving regulation system for a corn harvester based on reinforcement learning, characterized in that, It includes a multi-source heterogeneous data acquisition module, a reinforcement learning decision-making module, an action execution control module, a safety constraint module, and an operation control and supervision terminal; the multi-source heterogeneous data acquisition module collects the operating environment and status parameters of the corn harvester in real time, performs spatio-temporal alignment and noise filtering on the original data, and generates a three-dimensional state vector; The reinforcement learning decision-making module maps the three-dimensional state vector output by the multi-source heterogeneous data acquisition module into a discretized state code, designs a continuous action set, and the action execution control module converts the action instruction output by the reinforcement learning decision-making module into a specific control signal to automatically and adaptively control the operation of the corn harvester; The safety constraint module pre-sets several groups of risk items to be monitored. If corresponding risk items are detected during the operation of the corn harvester, corresponding safety constraint operations are performed, and a constraint alarm message is generated and sent to the operation control and supervision terminal, and the operation control and supervision terminal displays and warns of the constraint alarm message.

2. The energy-saving regulation system for a corn harvester based on reinforcement learning according to claim 1, characterized in that, The multi-source heterogeneous data acquisition module scans the crop row spacing and density distribution through a lidar. The lidar obtains the crop canopy point cloud data in a 360° rotation scanning mode, eliminates weed interference points through a voxel grid filtering algorithm, and identifies the crop row spacing and local density distribution based on a density clustering algorithm to generate a density heat map with a spatial resolution of 0.1m×0.1m; The body attitude of the corn harvester is monitored through an inertial measurement device, the body pitch angle, roll angle, and acceleration changes are monitored at a sampling frequency of 500Hz, and the multi-sensor data is fused using a Kalman filter algorithm to construct an attitude feature vector including the terrain slope and body vibration amplitude; The load change is obtained through an engine torque sensor, the instantaneous torque of the crankshaft is collected in real time, and the load fluctuation coefficient is calculated in combination with the ring gear speed signal; and the operation trajectory is recorded through a GPS positioning module.

3. The energy-saving regulation system for a corn harvester based on reinforcement learning according to claim 1, wherein, The action execution control module sends the engine speed adjustment instruction to the ECU via the CAN bus, and uses a feedforward-feedback composite control to reach the target speed within 200ms. The cutting table height adjustment instruction drives the hydraulic proportional valve through a PWM signal, and cooperates with a displacement sensor to achieve closed-loop control with a positioning accuracy of ±2mm. And the threshing cylinder torque limit instruction realizes torque soft limit by changing the current of the clutch solenoid valve.

4. A corn harvester energy-saving regulation system based on reinforcement learning according to claim 1, characterized in that, The operation control and supervision terminal is communicatively connected to the harvesting continuity detection module. The harvesting continuity detection module analyzes the harvesting continuity execution status of the corn harvester within a unit time, and sends a continuity qualified signal or a continuity abnormal signal to the operation control and supervision terminal.

5. The energy-saving regulation system of a corn harvester based on reinforcement learning according to claim 4, characterized in that, The specific analysis process of the harvesting continuity detection module is as follows: the total duration of the corn harvester in a non-optimal harvesting state within a unit time is obtained and marked as the non-optimal harvesting characteristic value. If the non-optimal harvesting characteristic value exceeds the preset non-optimal harvesting characteristic threshold, a continuity abnormal signal is generated.

6. The energy-saving regulation system of a corn harvester based on reinforcement learning according to claim 5, wherein If the harvested non-ideal eigenvalue does not exceed the preset harvested non-ideal eigenvalue threshold, the continuity evaluation value is calculated by weighted summation of the harvested non-ideal eigenvalue, the harvester operation analysis value, and the harvester anomaly value. If the continuity evaluation value exceeds the preset continuity evaluation threshold, a continuity anomaly signal is generated; if the continuity evaluation value does not exceed the preset continuity evaluation threshold, a continuity qualified signal is generated.

7. The energy-saving regulation system for a corn harvester based on reinforcement learning according to claim 4, wherein The harvesting continuity detection module is communicatively connected to the constraint effectiveness evaluation module. When the constraint effectiveness evaluation module receives the continuity qualified signal, it evaluates and analyzes the safety constraint effectiveness of the corn harvester within a unit time, generates a constraint qualified signal or a constraint hidden danger signal through the analysis, and sends the constraint qualified signal or the constraint hidden danger signal to the operation control supervision terminal.

8. An energy-saving regulation system for a corn harvester based on reinforcement learning according to claim 7, characterized in that, The specific analysis process of the constraint effectiveness evaluation module includes: Obtain the number of times the constraint non-effective symbol TW-1 is assigned within a unit time and calculate the ratio with the total number of occurrences of risk events to obtain the constraint non-effective detection value. If the constraint non-effective detection value exceeds the preset constraint non-effective detection threshold, a constraint hidden danger signal is generated; if the constraint non-effective detection value does not exceed the preset constraint non-effective detection threshold, the constraint effectiveness evaluation value is calculated by weighted summation of the constraint non-effective detection value, the constraint extension analysis value, and the constraint extension risk value. If the constraint effectiveness evaluation value exceeds the preset constraint effectiveness evaluation threshold, a constraint hidden danger signal is generated; otherwise, a constraint qualified signal is generated.

9. An energy-saving regulation method for a corn harvester based on reinforcement learning, characterized in that, It includes the following steps: Step 1: Collect the operating environment and state parameters of the corn harvester and perform data processing; Step 2: Design the continuous action set of the corn harvester; Step 3: Generate specific control signals based on the action instructions; Step 4: Monitor the risk events of the corn harvester. If a corresponding risk event is detected during the operation of the corn harvester, go to Step 5; Step 5: Make corresponding safety constraint operations and generate constraint alarm information.

Images