Adaptive control system for small epimedium harvesting machine in hilly terrain

CN122507166APending Publication Date: 2026-08-04贵州装备制造职业学院
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
贵州装备制造职业学院
Filing Date
2026-05-08
Publication Date
2026-08-04

AI Technical Summary

Technical Problem

[0003]目前适用于丘陵地区的小型收割机控制系统依赖于机身姿态的实时反馈进行滞后补偿,导致控制系统响应延迟;割台的俯仰、滚转姿态调整与切割高度的升降控制往往相互独立;且现有系统的功率管理无法根据实时的坡度、行进速度、作物密度以及地面阻力进行动态精细化分配,影响作业流畅性与效率

Benefits of technology

[0033] 1. This invention integrates forward slope detection and inertial measurement information, and adopts a feedforward-feedback composite control algorithm to predict terrain changes in advance and compensate for system lag, so that the cutting platform can smoothly and accurately track ground undulations, thereby improving the operational stability and cutting surface flatness on continuously changing slopes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122507166A_ABST
    Figure CN122507166A_ABST
Patent Text Reader

Abstract

This invention discloses a control system for a small-scale Epimedium harvester adapted to hilly terrain, belonging to the field of agricultural machinery control technology. The system includes a terrain sensing unit, a main controller, an attitude execution unit, and a header height execution unit. The terrain sensing unit is used to collect real-time information on the forward slope, machine attitude angle, and relative height of the header relative to the ground. The main controller generates the target pitch and roll angles of the header and the target cutting height. The attitude execution unit drives the header to adjust to the target pitch and roll angles. The header height execution unit drives the header to adjust to the target cutting height. This invention can sense terrain changes in real time and adaptively adjust the header attitude and cutting height, ensuring standardized harvesting of Epimedium even in terrains with varying slopes.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of agricultural machinery control technology, specifically to a control system for a small Epimedium harvester that adapts to hilly terrain. Background Technology

[0002] Agricultural mechanization is the core driving force for the development of modern agriculture. In my country's vast hilly and mountainous areas, the widespread application of large combine harvesters faces significant challenges due to the complex and varied terrain. Small harvesters, with their flexible maneuverability, have become an important guarantee for agricultural production in these areas. However, the complex undulating terrain, rapid changes in slope, and slippery, muddy ground of hilly slopes place extremely high demands on the stability, operational adaptability, and control precision of small harvesters. Specifically, the harvesting of the medicinal plant Epimedium requires even more stringent agronomical requirements, necessitating precise cutting of the above-ground stems and leaves 5-8 centimeters above the ground to protect the underground rhizomes and ensure regeneration capacity the following year.

[0003] Currently, the control systems of small harvesters suitable for hilly areas rely on real-time feedback of the machine's posture for lag compensation, resulting in a delayed response from the control system. The pitch and roll attitude adjustment of the header and the control of the cutting height are often independent of each other. Furthermore, the power management of the existing system cannot be dynamically and finely allocated according to real-time slope, travel speed, crop density, and ground resistance, affecting the smoothness and efficiency of operation.

[0004] Therefore, developing an adaptive control system that can comprehensively sense terrain, make intelligent decisions, and coordinately control the header posture and height, while also possessing dynamic energy management functions, is of great practical value and research significance for improving the operation quality, efficiency, and reliability of small harvesters in hilly areas and promoting the mechanization of specialty cash crop production. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to address the shortcomings of the prior art by providing a control system for a small epimedium harvester that is adaptive to hilly terrain.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] The control system of a small Epimedium harvester that adapts to hilly terrain includes a terrain sensing unit, a main controller, an attitude execution unit, and a header height execution unit.

[0008] The terrain sensing unit is used to collect information on the forward slope of the terrain where the harvester is located, the machine attitude angle, and the relative height between the cutting platform and the ground in real time.

[0009] The main controller is connected to the terrain sensing unit and is used to generate the target pitch angle and target roll angle of the header based on the forward slope information and the fuselage attitude angle information through a preset terrain prediction model; and to generate the target cutting height of the header based on the relative height information and the preset agronomic height.

[0010] The attitude execution unit is connected to and controlled by the main controller, and is used to drive the cutting table to adjust to the target pitch angle and the target roll angle;

[0011] The cutting table height execution unit is connected to and controlled by the main controller, and is used to drive the cutting table to adjust to the target cutting height.

[0012] Furthermore, the terrain sensing unit includes an inertial measurement module, a forward slope detection module, and a height detection module;

[0013] The inertial measurement module is used to measure the real-time pitch and roll angles of the harvester body.

[0014] The forward slope detection module uses a lidar or ultrasonic sensor array to detect terrain slope information within a preset distance in the direction of the harvester's movement.

[0015] The height detection module includes an ultrasonic sensor installed at the bottom of the cutting platform, used to measure the real-time relative distance between the cutting platform and the ground.

[0016] Furthermore, the attitude execution unit includes a pitch servo motor, a roll servo motor, and a dual-axis servo driver;

[0017] The main controller is used to send position commands containing the target pitch angle and the target roll angle to the dual-axis servo drive;

[0018] The dual-axis servo driver is connected to the pitch servo motor and the roll servo motor respectively, and has a built-in three-loop closed-loop control structure with position loop, speed loop and current loop, which is used to drive the pitch servo motor and the roll servo motor to track the position command.

[0019] Furthermore, when generating the target pitch angle, the main controller performs the following composite operation: the target pitch angle is equal to the sum of the current fuselage pitch angle, the feedforward term proportional to the predicted slope change, and the differential feedforward term proportional to the predicted slope change rate.

[0020] Furthermore, when generating the target roll angle, the main controller adds the current fuselage roll angle to a side roll compensation angle; wherein, the calculation process of the side roll compensation angle is as follows: first, divide the difference in ground height between the left and right sides by the wheel track to obtain the ratio, then calculate the arctangent function value of the ratio, and finally multiply it by the damping coefficient that is negatively correlated with the travel speed.

[0021] Furthermore, the main controller generates the target cutting height by comprehensively considering a preset fixed agronomic height, a terrain compensation height dynamically calculated based on real-time terrain information, and a visual compensation height based on plant recognition results; wherein, the terrain compensation height is output by a fuzzy PID controller, and the proportional, integral, and derivative parameters of the fuzzy PID controller are adjusted online according to the height error and the height error change rate.

[0022] Furthermore, the system also includes a visual recognition module, used to acquire crop images in front of the cutter, identify the base position of the Epimedium stem, and output the recognition confidence signal, the base height of a single frame, and the angle between the stem and the vertical direction to the main controller for participation in the fusion calculation of the target cutting height; the main controller calculates the visual compensation height according to a weighted average formula based on multiple frames of historical data.

[0023] Furthermore, the system also includes a power management module integrated into the main controller, which is used to dynamically allocate engine power to the walking system and the cutting system according to the current working load and terrain slope.

[0024] Furthermore, the power management module performs power allocation based on the following process:

[0025] Calculate the minimum traction power required to maintain the current flight status based on the current fuselage pitch angle, speed, and rolling resistance.

[0026] Subtract the minimum traction power and auxiliary system power consumption from the total available engine power to obtain the power budget available for cutting;

[0027] By combining the stalk density information, the power budget is adjusted to determine the final real-time power allocated to the cutting system.

[0028] Furthermore, the system also includes a security protection unit integrated into the main controller, comprising:

[0029] The software limiting module is used to limit the angle adjustment range of the attitude execution unit within a preset safety mechanical range;

[0030] The overload protection module is used to trigger a power reduction or shutdown command when the motor current of the attitude execution unit or the cutting table height execution unit continues to exceed the limit;

[0031] The emergency braking module is used to control the brake attached to the pitch servo motor to close and lock in the event of a system failure or receiving an emergency stop signal.

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

[0033] 1. This invention integrates forward slope detection and inertial measurement information, and adopts a feedforward-feedback composite control algorithm to predict terrain changes in advance and compensate for system lag, so that the cutting platform can smoothly and accurately track ground undulations, thereby improving the operational stability and cutting surface flatness on continuously changing slopes.

[0034] 2. This invention combines real-time terrain compensation, visual plant recognition, and preset agronomic height, and uses fuzzy PID to dynamically adjust the cutting height, effectively overcoming the interference caused by local uneven terrain and changes in machine posture.

[0035] 3. The power management module of this invention dynamically calculates and allocates engine power to the traction system and cutting system based on real-time detected machine pitch angle, travel speed, rolling resistance and crop density, ensuring priority of traction force under heavy load conditions such as climbing slopes, thereby optimizing the overall energy consumption of the machine and improving the overall operating efficiency in complex environments.

[0036] 4. This invention integrates a safety protection unit, which uses multiple mechanisms such as software limit, overload thermal protection and emergency braking to monitor and protect the actuator's range of motion, workload and abnormal state in real time, effectively preventing mechanical overtravel, motor overheating and other faults, and improving the reliability and safety of the system for long-term operation in harsh hilly terrain. Attached Figure Description

[0037] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0038] Figure 1 This is a schematic diagram of the system structure according to an embodiment of the present invention;

[0039] Figure 2 This is a flowchart illustrating the target attitude calculation process according to an embodiment of the present invention.

[0040] Figure 3 This is a flowchart illustrating the adaptive cutting height control according to an embodiment of the present invention.

[0041] Figure 4 This is a system state transition diagram according to an embodiment of the present invention. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0043] like Figure 1 As shown, the control system of the small Epimedium harvester that adapts to hilly terrain includes a terrain sensing unit, a main controller, an attitude execution unit, and a header height execution unit.

[0044] The terrain sensing unit is used to collect information on the forward slope of the terrain where the harvester is located, the machine attitude angle, and the relative height between the cutting platform and the ground in real time.

[0045] The main controller is connected to the terrain sensing unit and is used to generate the target pitch angle and target roll angle of the header based on the forward slope information and the fuselage attitude angle information through a preset terrain prediction model; and to generate the target cutting height of the header based on the relative height information and the preset agronomic height.

[0046] The attitude execution unit is connected to and controlled by the main controller, and is used to drive the cutting table to adjust to the target pitch angle and the target roll angle;

[0047] The cutting table height execution unit is connected to and controlled by the main controller, and is used to drive the cutting table to adjust to the target cutting height.

[0048] The terrain sensing unit includes an inertial measurement module, a forward slope detection module, and a height detection module;

[0049] The inertial measurement module is used to measure the real-time pitch and roll angles of the harvester body.

[0050] The forward slope detection module uses a lidar or ultrasonic sensor array to detect terrain slope information within a preset distance in the direction of the harvester's movement.

[0051] The height detection module includes an ultrasonic sensor installed at the bottom of the cutting platform, used to measure the real-time relative distance between the cutting platform and the ground.

[0052] In the height detection module, the distance values ​​of all sensors are read in each control cycle, and the arithmetic mean of the measured values ​​is taken as the final real-time relative distance. The ground height difference is obtained by subtracting the mean of the measured values ​​of the two ultrasonic sensors on the right from the mean of the measured values ​​of the two ultrasonic sensors on the left.

[0053] The inertial measurement module uses a six-axis IMU to output a real-time pitch angle that represents the tilt angle of the fuselage in the forward and backward directions, with forward tilt being positive and backward tilt being negative; and a real-time roll angle that represents the tilt angle of the fuselage in the left and right directions, with left downward tilt being positive and right downward tilt being negative.

[0054] The forward slope detection module uses a single-line lidar or ultrasonic sensor array, installed at the front of the fuselage, to scan forward and downward, and output the predicted slope change. Based on the current scan data, it predicts the change in slope in the forward direction within a future control cycle (e.g., 1 second). A positive value indicates that it is about to go uphill, and a negative value indicates that it is about to go downhill. The predicted slope change rate, that is, the instantaneous rate of change of the predicted slope change, reflects the degree of abruptness of terrain change.

[0055] The height detection module directly measures the real-time distance between the cutting platform and the ground below. Multiple ultrasonic ranging sensors are installed on the cutting beam of the cutting platform to form a measurement array, outputting the real-time relative distance, that is, the vertical distance from the ultrasonic sensor's emitting surface to the ground. The measurement values ​​of multiple sensors can be fused to eliminate the influence of local unevenness. The ground height difference is obtained by the difference between the measurement values ​​of the grounding sensors on the left and right sides of the machine body or the left and right sides of the ultrasonic array, which is used to calculate the lateral slope.

[0056] The attitude execution unit includes a pitch servo motor, a roll servo motor, and a dual-axis servo driver;

[0057] The main controller is used to send position commands containing the target pitch angle and the target roll angle to the dual-axis servo drive;

[0058] The dual-axis servo driver is connected to the pitch servo motor and the roll servo motor respectively, and has a built-in three-loop closed-loop control structure with position loop, speed loop and current loop, which is used to drive the pitch servo motor and the roll servo motor to track the position command.

[0059] The attitude execution unit is responsible for converting the target pitch and roll angles calculated by the main controller into the actual attitude of the cutting table. The control process includes: the main controller sending position commands to the dual-axis servo driver via the communication bus. The driver internally adopts a closed-loop control structure with three cascaded loops: position loop, speed loop, and current loop. The position loop receives the target angle command, compares it with the actual angle fed back by the motor encoder, and outputs the target speed through PID calculation. The speed loop receives the target speed output by the position loop, compares it with the actual speed obtained by encoder differentiation or estimation, and outputs the target current through PID calculation. The current loop receives the target current output by the speed loop, compares it with the motor phase current sampled by the drive circuit, generates a PWM signal through a high-bandwidth PI regulator, and finally controls the motor torque to achieve high-precision position tracking.

[0060] When generating the target pitch angle, the main controller performs the following composite operation: the target pitch angle is equal to the sum of the current fuselage pitch angle, the feedforward term proportional to the predicted slope change, and the differential feedforward term proportional to the predicted slope change rate. The specific calculation formula is as follows:

[0061]

[0062] in, This indicates the target pitch angle reached by the cutter at time k+1 in the next control cycle. This represents the current fuselage pitch angle fed back by the inertial measurement module at time k. This indicates the predicted slope change provided by the forward slope detection module. This indicates the predicted slope change rate provided by the forward slope detection module. This represents the pitch angle proportional feedforward gain coefficient, calibrated based on the harvester's mass, center of gravity position, and suspension system stiffness, with a value ranging from 0.5 to 1.2. This represents the differential feedforward gain coefficient of the pitch angle, which determines the damping effect of the system on the rate of change of slope, and its value ranges from 0.1 to 0.5.

[0063] When generating the target roll angle, the main controller adds the current fuselage roll angle to a side roll compensation angle. The calculation process for the side roll compensation angle is as follows: first, the difference in ground contact height between the left and right sides is divided by the wheelbase to obtain a ratio; then, the arctangent function value of this ratio is calculated; finally, it is multiplied by a damping coefficient that is negatively correlated with the travel speed. The specific formula is as follows:

[0064]

[0065] in, This represents the target roll angle calculated by the main controller. This indicates the current fuselage roll angle reported by the inertial measurement module. This represents a function of the damping coefficient related to the speed of travel, v. These represent the grounding height on the left and the grounding height on the right, respectively. The wheelbase of the harvester is a fixed structural parameter. Represents the arctangent function;

[0066] The specific formula for the damping coefficient function is as follows:

[0067]

[0068] in, This represents the attenuation constant, typically 0.2. It is used to reduce the response intensity to lateral undulations at higher travel speeds, preventing the cutter head from shaking due to frequent adjustments and ensuring smooth travel. Indicates the speed of travel.

[0069] like Figure 2As shown, the system first reads all sensor data to obtain the current fuselage attitude angle, forward terrain prediction data, and ground height difference between the left and right sides. The system then calculates the target pitch angle and target roll angle in parallel. After the calculation is completed, a software limit check is performed to ensure that the calculation results are within the mechanical safety range. Finally, the verified target angle command is output to the dual-axis servo drive to complete the entire attitude control cycle.

[0070] The main controller generates the target cutting height by comprehensively considering a preset fixed agronomic height, a terrain compensation height dynamically calculated based on real-time terrain information, and a visual compensation height based on plant recognition results; wherein, the terrain compensation height is output by a fuzzy PID controller, and the proportional, integral, and derivative parameters of the fuzzy PID controller are adjusted online according to the height error and the height error change rate;

[0071] The specific formula for the target cutting height is as follows:

[0072]

[0073] in, This indicates the target cutting height calculated by the main controller. This indicates the preset agronomic height set according to the epimedium variety and harvesting standards. Indicates the terrain compensation height. Indicates visual compensation height;

[0074] The specific formula for the terrain compensation height is as follows:

[0075]

[0076] in, This represents the current height tracking error, which is the instantaneous deviation between the actual cutting position calculated based on the current sensor measurements and the target agronomic height. This represents the rate of change of altitude error, i.e., the first derivative of the current altitude tracking error with respect to time. Indicates the current fuselage pitch angle. This represents the slope suppression coefficient, used to control the degree to which the aircraft's pitch angle suppresses the terrain compensation height. Its value ranges from 0.1 to 0.3. Indicates fuzzy adaptive proportional gain. This represents the fuzzy adaptive integral gain. This represents the fuzzy adaptive differential gain. Indicates time;

[0077] The formula for calculating the altitude tracking error is as follows:

[0078]

[0079]

[0080] in, This indicates the mechanical reference height for the installation of the cutting table; it is a fixed mechanical parameter. This indicates the real-time ultrasonic distance measured by the height detection module. When the fuselage has a pitch angle, the slant distance measured by the ultrasonic wave needs to be converted into the actual vertical height. This represents the recognition confidence level output by the visual recognition module, with a value ranging from 0 to 1. Indicates the predicted compensation term. This represents the system's equivalent delay time, including ultrasonic response delay and actuator hysteresis, with a typical value of 0.1–0.2 seconds. This represents the estimated vertical velocity of the ground. This represents the estimated vertical acceleration of the ground, obtained in real time by using an extended Kalman filter to estimate the ultrasonic ranging sequence. This represents the pitch rate suppression coefficient, with a value ranging from 0.1 to 0.3. Indicates the fuselage pitch rate. This represents the reference angular velocity, typically 5° / s. This represents the change in slope, originating from the forward slope detection module. This represents the maximum expected rate of change in slope, typically 10%.

[0081] The confidence level is calculated using the following formula: =0.4×Edge Sharpness + 0.3×Stalk Continuity + 0.3×Segmentation Region Consistency, where edge sharpness is the normalized average gradient value of the detected edge pixels, stalk continuity is the ratio of the length of the stalk outline to the image height, and segmentation region consistency is the normalized inverse of the color variance of the segmented region.

[0082] The specific formula for the visual compensation height is as follows:

[0083]

[0084] in, This represents the visual compensation height during the k-th control cycle. This represents a window of historical frame counts, typically ranging from 5 to 10. This represents the height of the stalk base identified in the ki-th frame image in the header coordinate system, obtained from pixel coordinates through camera calibration and header pose transformation. This indicates the angle between the stem and the vertical direction within the same frame. Indicates the preset agronomic height. The weights, which combine visual confidence and time decay, are expressed by the following formula:

[0085]

[0086] in, This represents the visual recognition confidence level in the ki-th control cycle. This represents the confidence power exponent, typically ranging from 1.5 to 2.5, and is used to suppress low-confidence nonlinear amplification. This represents the timing decay constant, typically 3 to 5, which reduces the weight of earlier frames.

[0087] like Figure 3 As shown, the system first acquires various input data, including the preset agronomic height, ultrasonic ranging value, visual recognition confidence level, and current fuselage pitch angle. It then calculates the error between the current actual cutting position and the target height, and further calculates the rate of change of this error. These two parameters serve as inputs to the fuzzy PID controller. The fuzzy PID controller outputs a terrain compensation height, which is combined with the preset agronomic height and the visual compensation height to obtain the final target cutting height. Before outputting the command, the system checks whether the target height exceeds the preset safe mechanical range: if it does, software limit switching is applied to restrict the height within the safe range; if it does not exceed the limit, the height command is directly output to the actuator. After the command output is completed, the current height control cycle ends.

[0088] The specific rules for fuzzy adaptive proportional gain adjustment include:

[0089] When the altitude tracking error is large, a larger proportional gain, such as 1.5 to 2.0, is used to enable the system to respond quickly and rapidly reduce the error.

[0090] When the tracking error is moderate, a moderate proportional gain, such as 0.8~1.2, is used to balance response speed and stability.

[0091] When the tracking error is very small, a small proportional gain, such as 0.3~0.6, is used for fine adjustment to avoid overshoot and oscillation.

[0092] The specific rules for adjusting the fuzzy adaptive integral gain include:

[0093] When the altitude tracking error is large and the rate of change of altitude error is greater than 0, a smaller integral gain is used to temporarily suppress the integral action and avoid integral saturation that would cause the system to become sluggish.

[0094] When the altitude tracking error is small and the rate of change of altitude error is less than 0, a moderate integral gain, such as 0.05~0.1, is used to eliminate the residual steady-state error.

[0095] When the altitude tracking error is close to zero and the rate of change of altitude error is close to zero, a very small integral gain or zeroing is used to prevent integral accumulation from causing low-frequency oscillations in the system near the equilibrium point.

[0096] The specific rules for adjusting the fuzzy adaptive differential gain include:

[0097] When the rate of change of height error is large, a larger differential gain, such as 0.3~0.5, is used to strongly suppress the change trend and prevent overshoot.

[0098] When the rate of change of height error is moderate, a moderate differential gain, such as 0.1 to 0.2, is used to provide adequate damping.

[0099] When the rate of change of height error is very small, a very small differential gain, such as 0.02~0.05, is used to avoid being too sensitive to sensor measurement noise and causing high-frequency jitter.

[0100] The system also includes a visual recognition module, which is used to acquire crop images in front of the cutter, identify the base position of the Epimedium stem, and output the recognition confidence signal, the base height of a single frame, and the angle between the stem and the vertical direction to the main controller for participation in the fusion calculation of the target cutting height; the main controller calculates the visual compensation height according to a weighted average formula based on multiple frames of historical data.

[0101] The vision recognition module uses a global shutter industrial camera, mounted in front of the header, tilted downwards to capture images of the crop area. Specifically, it includes:

[0102] RGB images were acquired, converted to HSV color space, and threshold segmentation was performed using the specific green range of Epimedium stems and leaves to initially extract the plant area.

[0103] Morphological operations and edge detection are applied to the binarized image to identify the approximately vertical stem outline. By analyzing the distribution characteristics of the pixels at the lower end of the stem and combining them with ground line estimation, the position of the stem base is determined.

[0104] By combining edge sharpness, stem continuity, and consistency of segmented regions, the confidence level of the recognition result is calculated, and the recognized base height is transformed into the header coordinate system to obtain the base height of a single frame and the angle between the stem and the vertical direction, which is then output to the main controller.

[0105] The main controller calculates the visual compensation height according to a formula based on multiple frames of historical data.

[0106] The system also includes a power management module, which is integrated into the main controller and is used to dynamically allocate engine power to the walking system and the cutting system according to the current working load and terrain slope.

[0107] The threshold range for HSV color space segmentation is set according to the color of Epimedium stems and leaves as follows: hue 35°~85°, saturation 40%~100%, brightness 30%~100%, morphological operation adopts a closed operation of first dilation and then erosion, the structural element is a 3×3 pixel rectangle, edge detection adopts the Canny operator, and the high and low thresholds are 150 and 50 respectively.

[0108] The stalk base localization is achieved through the following steps: performing connected component analysis on the binarized image and selecting connected components with an aspect ratio greater than 3 as candidate stalks; for each candidate stalk, taking the midpoint of the bottom 5 consecutive rows of pixels as the initial base position; using a random sampling consensus algorithm to fit a straight line on the ground, and vertically projecting the initial base position onto this straight line, the projection point is the final identified stalk base position.

[0109] The power management module performs power allocation based on the following process:

[0110] Based on the current fuselage pitch angle, speed, and rolling resistance, calculate the minimum traction power required to maintain the current travel status. The specific formula is as follows:

[0111]

[0112] in, Indicates minimum traction power. Indicates the overall weight of the machine. Represents gravitational acceleration. Indicates the current speed of travel. This represents the efficiency of the traction drive system, taken as 0.85. The rolling resistance coefficient depends on ground conditions. This indicates the fixed power consumption of the auxiliary system, including controllers, sensors, hydraulic pumps, etc.

[0113] The unit for the overall mass is kg, the unit for gravitational acceleration is m / s², the unit for travel speed is m / s, the traction transmission efficiency is dimensionless, and the rolling resistance coefficient is dimensionless.

[0114] Subtracting the minimum traction power and auxiliary system power consumption from the total available engine power yields the power budget available for cutting, using the following formula:

[0115]

[0116] in, This indicates the power budget available for cutting. Indicates the total available power of the engine;

[0117] Based on the stalk density information, the power budget is adjusted to determine the final real-time power allocated to the cutting system, using the following formula:

[0118]

[0119] in, This represents the real-time power allocated to the cutting system. This indicates the maximum allowable power of the cutting system. This indicates the stalk density level estimated by the visual recognition module. This represents the density influence coefficient, used to reserve more power margin for cutting when the stalk density is high, or to ensure cutting quality by reducing the travel speed when the power budget is insufficient. The value ranges from 0.1 to 0.4.

[0120] The stalk density level is obtained by the visual recognition module by counting and mapping the number of stalks per unit area: less than 5 is level 1, 5-10 is level 2, 10-15 is level 3, 15-20 is level 4, and greater than 20 is level 5.

[0121] The system also includes a security protection unit, integrated into the main controller, comprising:

[0122] The software limiting module is used to limit the angle adjustment range of the attitude execution unit within a preset safety mechanical range;

[0123] The overload protection module is used to trigger a power reduction or shutdown command when the motor current of the attitude execution unit or the cutting table height execution unit continues to exceed the limit;

[0124] The emergency braking module is used to control the brake attached to the pitch servo motor to close and lock in the event of a system failure or receiving an emergency stop signal.

[0125] The software limit module performs real-time clamping processing on the output of the main controller to ensure that it is always within the preset mechanical limit range, specifically including: pitch angle limited to -15° to +15°, roll angle limited to -10° to +10°, and cutting height limited to 4 cm to 12 cm.

[0126] The overload protection module continuously samples the motor phase current and calculates its thermal effect integral relative to the rated current. When the thermal effect integral exceeds the safety threshold, the protection is triggered, and the output current is gradually reduced or the machine is shut down.

[0127] The emergency braking module works in concert with a hardware watchdog circuit and a software fault diagnosis thread. Once emergency braking is triggered, the program will immediately cut off the enable signals of all motors and control the electromagnetic brake integrated into the pitch servo motor to apply the brake, lock the cutter table's attitude, and prevent it from falling due to gravity.

[0128] like Figure 4As shown, after power-on, the system first performs a power-on self-test to check the status of hardware, sensors, and communication links. If the self-test passes, the system enters standby mode to await operation commands; if the self-test fails, it directly enters a fault-locked state, requiring manual intervention. Upon receiving a start command in standby mode, the system enters normal operation mode. During normal operation, the system continuously performs three parallel safety monitoring functions: monitoring whether control commands exceed mechanical limits, monitoring whether motor current is overloaded, and monitoring for serious faults or emergency stop signals. If a command exceeds limits, a software limit mechanism is triggered, clamping the command to a safe range before continuing operation; if the current is overloaded, overload protection is triggered, implementing power reduction or safe shutdown measures, returning to standby mode after a safe shutdown; if a serious fault or emergency stop signal is detected, emergency braking is immediately triggered, entering a fault-locked state. The fault-locked state requires manual reset to restart the system, ensuring operational safety.

[0129] After the system is powered on, the main controller executes the initialization and calibration process: First, the harvester is placed on a horizontal ground, and the data from the inertial measurement module is collected 100 times continuously. The average value is then calibrated as the zero point of the pitch and roll angles. The header is raised and lowered to a known mechanical reference height, and the average reading of the ultrasonic sensor array at this time is recorded and calibrated as the distance zero point. Finally, communication is established with the servo driver and engine ECU through the CAN bus, and their status information is read to complete the system readiness.

[0130] The examples described herein are merely preferred embodiments of the invention and are not intended to limit the concept and scope of the invention. Any modifications and improvements made by those skilled in the art to the technical solutions of the invention without departing from the design concept of the invention should fall within the protection scope of the invention.

Claims

1. A control system for a small epimedium harvester adapted to hilly terrain, characterized in that, It includes a terrain sensing unit, a main controller, an attitude execution unit, and a cutter height execution unit; The terrain sensing unit is used to collect information on the forward slope of the terrain where the harvester is located, the machine attitude angle, and the relative height between the cutting platform and the ground in real time. The main controller is connected to the terrain sensing unit and is used to generate the target pitch angle and target roll angle of the header based on the forward slope information and the fuselage attitude angle information through a preset terrain prediction model; and to generate the target cutting height of the header based on the relative height information and the preset agronomic height. The attitude execution unit is connected to and controlled by the main controller, and is used to drive the cutting table to adjust to the target pitch angle and the target roll angle; The cutting table height execution unit is connected to and controlled by the main controller, and is used to drive the cutting table to adjust to the target cutting height.

2. The system according to claim 1, characterized in that, The terrain sensing unit includes an inertial measurement module, a forward slope detection module, and a height detection module; The inertial measurement module is used to measure the real-time pitch and roll angles of the harvester body. The forward slope detection module uses a lidar or ultrasonic sensor array to detect terrain slope information within a preset distance in the direction of the harvester's movement. The height detection module includes an ultrasonic sensor installed at the bottom of the cutting platform, used to measure the real-time relative distance between the cutting platform and the ground.

3. The system according to claim 2, characterized in that, The attitude execution unit includes a pitch servo motor, a roll servo motor, and a dual-axis servo driver; The main controller is used to send position commands containing the target pitch angle and the target roll angle to the dual-axis servo drive; The dual-axis servo driver is connected to the pitch servo motor and the roll servo motor respectively, and has a built-in three-loop closed-loop control structure with position loop, speed loop and current loop, which is used to drive the pitch servo motor and the roll servo motor to track the position command.

4. The system according to claim 3, characterized in that, When generating the target pitch angle, the main controller performs the following composite operation: the target pitch angle is equal to the sum of the current fuselage pitch angle, the feedforward term proportional to the predicted slope change, and the differential feedforward term proportional to the predicted slope change rate.

5. The system according to claim 4, characterized in that, When generating the target roll angle, the main controller adds the current fuselage roll angle to a side roll compensation angle. The calculation process of the side roll compensation angle is as follows: first, the difference in ground height between the left and right sides is divided by the wheel track to obtain the ratio; then, the arctangent function value of the ratio is calculated; and finally, it is multiplied by the damping coefficient that is negatively correlated with the travel speed.

6. The system according to claim 5, characterized in that, The main controller generates the target cutting height by comprehensively considering a preset fixed agronomic height, a terrain compensation height dynamically calculated based on real-time terrain information, and a visual compensation height based on plant recognition results. The terrain compensation height is output by a fuzzy PID controller, whose proportional, integral, and derivative parameters are adjusted online based on the height error and the rate of change of the height error.

7. The system according to claim 6, characterized in that, The system also includes a visual recognition module, which is used to acquire crop images in front of the cutter, identify the base position of the Epimedium stem, and output the recognition confidence signal, the base height of a single frame, and the angle between the stem and the vertical direction to the main controller for participation in the fusion calculation of the target cutting height; the main controller calculates the visual compensation height according to a weighted average formula based on multiple frames of historical data.

8. The system according to claim 7, characterized in that, The system also includes a power management module, which is integrated into the main controller and is used to dynamically allocate engine power to the walking system and the cutting system according to the current working load and terrain slope.

9. The system according to claim 8, characterized in that, The power management module performs power allocation based on the following process: Calculate the minimum traction power required to maintain the current flight status based on the current fuselage pitch angle, speed, and rolling resistance. Subtract the minimum traction power and auxiliary system power consumption from the total available engine power to obtain the power budget available for cutting; By combining the stalk density information, the power budget is adjusted to determine the final real-time power allocated to the cutting system.

10. The system according to claim 9, characterized in that, The system also includes a security protection unit, integrated into the main controller, comprising: The software limiting module is used to limit the angle adjustment range of the attitude execution unit within a preset safety mechanical range; The overload protection module is used to trigger a power reduction or shutdown command when the motor current of the attitude execution unit or the cutting table height execution unit continues to exceed the limit; The emergency braking module is used to control the brake attached to the pitch servo motor to close and lock in the event of a system failure or receiving an emergency stop signal.