Self-adaptive heat dissipation method of intelligent weeding robot for orchard

Abstract

The invention discloses a self-adaptive heat dissipation method for an orchard intelligent weeding robot, and relates to the technical field of weeding robots, and the method comprises the following steps: S1, sensor installation and initialization, S2, data acquisition, S3, sensor fusion and heat dissipation demand judgment, S4, heat dissipation control strategy adjustment, and S5, continuous monitoring and feedback. Through the multi-sensor fusion technology, the internal and external environment parameters of the robot are comprehensively monitored, the heat dissipation requirement is accurately judged through a sensor fusion algorithm, the heat dissipation efficiency is improved, equipment damage caused by environmental factors is avoided, the air cooling strength is intelligently adjusted according to the environment humidity, and damage to equipment caused by water vapor condensation is prevented; normal operation of equipment is guaranteed, the acquisition time interval of a sensor is dynamically adjusted, system resource consumption is reduced, timely response can be achieved when the heat dissipation requirement changes, all heat dissipation control links are accurately regulated and controlled, the heat dissipation system can better adapt to different working states and environment changes of the robot, and the heat dissipation efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of weeding robots, and specifically to an adaptive heat dissipation method for an intelligent orchard weeding robot. Background Art

[0002] With the continuous expansion of orchard planting scale and the advancement of agricultural automation, intelligent orchard weeding robots have emerged. They play an important role in improving weeding efficiency and reducing labor costs. The weeding robot can automatically identify weeds and perform weeding operations, greatly reducing the labor intensity of farmers. The use of weeding robots makes orchard management more efficient, and farmers can focus more on other important agricultural activities. By improving weeding efficiency and reducing the use of chemical drugs, weeding robots contribute to enhancing the economic benefits of orchards. The application of weeding robots in orchards is of great significance and role. It can not only improve weeding efficiency, ensure orchard quality, but also reduce the use of chemical drugs, bringing more economic benefits to fruit farmers. However, the orchard environment is complex and changeable, and the weeding robot faces many challenges during operation, especially the heat dissipation problem seriously restricts the improvement of its performance and stable operation.

[0003] In the orchard environment, factors such as temperature, humidity, and wind speed vary significantly in different seasons, different time periods, and different regions. Traditional heat dissipation methods often lack consideration of the adaptability to this complex environment. Traditional heat dissipation methods often adopt a single heat dissipation method, such as a fan with a fixed rotation speed for heat dissipation or a simple liquid cooling system, and cannot effectively adjust heat dissipation according to the complex orchard environment where the robot is located and the changes in its own working state. For example, in a high-temperature and high-humidity environment, if it still operates according to the conventional heat dissipation strategy, it may cause water vapor to condense inside the device, damaging electronic components. Moreover, the data accuracy monitored by a single sensor is limited, making it difficult to comprehensively and accurately judge the heat dissipation requirements, thus affecting the working stability and service life of the robot. Therefore, it is necessary to propose an adaptive heat dissipation method for an intelligent orchard weeding robot to solve the problems in the prior art. Summary of the Invention

[0004] The purpose of the present invention is to make up for the deficiencies of the prior art and provide an adaptive heat dissipation method for an intelligent orchard weeding robot. It can adopt multi-sensor fusion technology to comprehensively monitor the internal and external environment parameters of the robot, and based on these parameters, achieve intelligent heat dissipation control. By integrating multiple sensors such as temperature sensors, humidity sensors, and wind speed sensors, and using sensor fusion algorithms to comprehensively analyze the collected data, it can more accurately judge the heat dissipation status and requirements, effectively improve the heat dissipation efficiency, and avoid equipment damage caused by environmental factors.

[0005] To solve the above technical problems, the present invention provides the following technical solution: an adaptive heat dissipation method for an orchard intelligent weeding robot, the method comprising the following steps:

[0006] S1. Sensor installation and initialization: Install temperature sensors, humidity sensors and wind speed sensors on the surface and external positions of the heat-generating components of the orchard intelligent weeding robot respectively, and perform initialization calibration to establish a data transmission connection with the robot control system;

[0007] S2. Data acquisition: During the operation of the robot, each sensor collects temperature data of the internal heat-generating components of the robot, humidity data of the surrounding environment and wind speed data at a predetermined time interval, and transmits them to the robot control system;

[0008] S3. Sensor fusion and heat dissipation requirement judgment: The robot control system processes the received data using a sensor fusion algorithm, and the sensor fusion algorithm calculates the heat dissipation requirement index Q based on the following formula: where T is the average temperature value of the heat-generating component, in °C, H is the environmental humidity value, in %, S is the environmental wind speed value, in m / s, k1, k2, k3 are weighting coefficients preset according to the heat dissipation characteristics of the robot, and k1 + k2 + k3 = 1. The Q value calculated is used to determine the current heat dissipation requirement level of the robot according to the preset heat dissipation requirement level division standard;

[0009] S4. Heat dissipation control strategy adjustment: According to the heat dissipation requirement level, the robot control system automatically adjusts the heat dissipation control strategy, including adjusting the fan speed, coolant flow rate and starting the auxiliary heat dissipation device. When the humidity exceeds the set threshold, the air-cooled heat dissipation intensity is reduced;

[0010] S5. Continuous monitoring and feedback: After the heat dissipation control strategy is adjusted, the heat dissipation status of the robot is continuously monitored through sensors, and the real-time monitoring data is compared and analyzed with the preset target heat dissipation parameters, and the heat dissipation control strategy is adjusted again according to the deviation situation to form a closed-loop feedback control.

[0011] Further, in step S1, sensor installation and initialization, the installation position of the temperature sensor is determined according to the heat distribution model of the heat-generating component, and the heat distribution model is constructed by the following method:

[0012] When the robot is in different working modes including low-speed driving for weeding, high-speed turning, and climbing operation, use an infrared thermal imager to perform thermal imaging scanning on key heat-generating components to obtain heat distribution data points;

[0013] Use a clustering algorithm based on neural network to analyze and process these data points, classify the data points with similar thermal characteristics into one category, and determine the typical heat distribution areas of the heat-generating components in different working modes;

[0014] An improved self-organizing mapping neural network SOM is adopted according to the clustering algorithm of these typical thermal distribution networks, and the neuron connection weight update formula is as follows: Among them, w ij (t) represents the connection weight between the i-th input data and the j-th neuron at time t, α(t) and β(t) are respectively learning rate functions that decrease with time, h ij (t) is the neighborhood function, representing the influence degree of the j-th neuron on the i-th input data, and g ik (t) is the global influence function, representing the global influence of the k-th neuron on the i-th input data, x i is the i-th input data, and K is the total number of neurons.

[0015] Furthermore, in S2, the data acquisition step, the data acquisition time interval of each sensor adopts a dynamic adjustment strategy, and the dynamic adjustment of the data acquisition time interval is determined according to the following formula: Among them, I is the actual data acquisition time interval, in milliseconds, I0 is the initially set data acquisition time interval, in milliseconds, Q is the currently calculated heat dissipation demand index, Q0 is the preset standard heat dissipation demand index. When the robot is in a stable working state and the heat dissipation demand level is low, the data acquisition time interval is once every 20 milliseconds. When the robot is in a high-load working state and the heat dissipation demand level is high, the data acquisition time interval is once every 5 milliseconds.

[0016] Furthermore, in S3, the sensor fusion and heat dissipation demand judgment step, the determination process of the weighting coefficients k1, k2, and k3 is as follows:

[0017] Construct a multi-variable test matrix, including different combinations of temperature, humidity, and wind speed, covering various working conditions that may occur in the orchard environment;

[0018] For each combination, use the finite element analysis method to simulate the heat dissipation process of the robot under this working condition to obtain the corresponding theoretical heat dissipation demand value;

[0019] Compare the heat dissipation effect data of the robot obtained from actual tests under the same working condition with the theoretical heat dissipation demand value, and adjust the weighting coefficients k1, k2, and k3 through the least squares optimization algorithm to minimize the error between the calculated heat dissipation demand index Q and the actual heat dissipation demand. This finite element analysis method is based on an improved heat conduction equation solving algorithm, and a correction term for the influence of environmental humidity on the heat conduction coefficient is introduced into the traditional heat conduction equation. The corrected heat conduction equation is: Among them, ρ is the material density, c is the specific heat capacity, T is the temperature, t is the time, k(T, H) is the thermal conductivity considering the influence of humidity, and its expression is k(T, H) = k0×(1 - γH)×(1 + δT), where k0 is the initial thermal conductivity, and γ and δ are correction factors determined according to the material and environmental characteristics, and q is the internal heat source term.

[0020] Furthermore, in S4, the adjustment of the fan speed in the heat dissipation control strategy adjustment step is based on the following formula: N = N0×(1 + λ×(Q - Q0)), where N is the adjusted fan speed in revolutions per minute, N0 is the initial fan speed in revolutions per minute, λ is the speed adjustment coefficient determined according to the fan characteristics and the matching of the robot heat dissipation system, Q is the current heat dissipation demand index, Q0 is the preset standard heat dissipation demand index. A speed adjustment threshold ΔN is set. When the calculated speed adjustment amount exceeds ΔN, the fan speed is actually adjusted. When adjusting the fan speed, an optimization algorithm based on the acoustic model is used. This acoustic model is constructed based on the aerodynamic acoustics principle of the fan blades. By analyzing the relationship between the shape, number, speed of the fan blades and the air flow speed factors and the generation of noise, the adjustment of the fan speed is constrained and optimized.

[0021] Furthermore, in S4, the heat dissipation control strategy adjustment step, the adjustment of the coolant flow rate adopts a method based on fuzzy logic control. The input variables are defined as the heat dissipation demand index Q and the current coolant temperature T c , and the output variable is the coolant flow rate adjustment ratio F. A fuzzy control rule table is established. When the heat dissipation demand index Q is high and the current coolant temperature T c is also high, the coolant flow rate adjustment ratio F is large. When the heat dissipation demand index Q is low and the current coolant temperature T c is low, the coolant flow rate adjustment ratio F is small. The fuzzification process uses a triangular membership function to fuzzify the input variables. The inference process uses the Mamdani fuzzy inference method. The defuzzification uses the centroid method to obtain the accurate coolant flow rate adjustment ratio F. The actual coolant flow rate adjustment amount ΔV is calculated according to the coolant flow rate adjustment ratio F: ΔV = F×V0, where V0 is the initial coolant flow rate in liters per minute.

[0022] Further, in step S4, the heat dissipation control strategy adjustment step, when the auxiliary heat dissipation device is started and the heat dissipation fin deployment mechanism is adopted, the control of the deployment angle θ is based on the following formula: θ = arctan(μ×(Q - Q0)), where μ is the angle adjustment coefficient determined according to the structural characteristics of the heat dissipation fins and the heat dissipation requirements of the robot, Q is the current heat dissipation requirement index, and Q0 is the preset standard heat dissipation requirement index. The servo control algorithm based on position feedback is used. By installing an angle sensor on the heat dissipation fins, the deployment angle of the heat dissipation fins is monitored in real time and compared with the target deployment angle. According to the deviation, the output torque of the servo motor is adjusted so that the heat dissipation fins can reach the predetermined deployment angle. The servo control algorithm uses an improved PID controller, and its control law is: where u(t) is the output torque of the servo motor, K p 、K i 、K d are the proportional, integral, and differential coefficients respectively, e(t) is the angle deviation, and ω(t) is the compensation term used to overcome the influence of the friction and inertia of the system.

[0023] Further, in step S5, the continuous monitoring and feedback step, a predictive heat dissipation control strategy based on data trend analysis is adopted. Time series analysis is performed on the continuously collected heat dissipation-related data including temperature, humidity, wind speed, and heat dissipation control parameters to construct a data trend model. The autoregressive integrated moving average ARIMA model is used to model the temperature data: where T t is the temperature value at time t, φ i and θ j are the autoregressive coefficient and the moving average coefficient respectively, L is the lag operator, ∈ t is the white noise sequence, and p and q are the model orders. The values of φ i 、θ j 、p, and q are determined by fitting the historical data. According to the constructed data trend model, the change trend of the heat dissipation condition in the future period is predicted. When it is predicted that the temperature will rise rapidly in the future, even if the current heat dissipation requirement level has not reached a high level, the fan speed and the coolant flow rate are appropriately increased in advance.

[0024] Furthermore, a fault diagnosis and early warning system is established during the operation of the entire heat dissipation method. This system comprehensively analyzes sensor data, heat dissipation control parameters, and robot working state data to determine whether there are potential faults in the heat dissipation system. When the temperature data collected by the temperature sensor shows abnormal fluctuations and does not match other sensor data and heat dissipation control strategies, and when the heat dissipation control parameters such as the fan speed and coolant flow rate cannot be adjusted according to the predetermined strategy, the fault diagnosis and early warning system determines that there may be problems with sensor faults, heat dissipation component faults, and control system faults. At this time, the system issues a fault warning signal to notify the operator to conduct inspections and repairs. At the same time, it automatically adjusts the heat dissipation control strategy and takes emergency heat dissipation measures, including starting a standby cooling fan and increasing the pressure in the coolant storage tank. The fault diagnosis and early warning system uses an intelligent diagnosis algorithm based on multi-source data fusion. This algorithm maps sensor data, control parameter data, and robot working state data to a high-dimensional feature space, and uses a support vector machine (SVM) classifier to identify and classify different fault modes. Its classification decision function is: where x is the data sample to be diagnosed, x i is the training sample, y i is the class label of the training sample, α i is the Lagrange multiplier, K(x i , x) is the kernel function, b is the bias term, and n is the number of training samples.

[0025] Compared with the prior art, the adaptive heat dissipation method of this orchard intelligent weeding robot has the following beneficial effects:

[0026] Through multi-sensor fusion technology, the present invention comprehensively monitors the internal and external environmental parameters of the robot, accurately judges the heat dissipation requirements using the sensor fusion algorithm, effectively improves the heat dissipation efficiency, and avoids equipment damage caused by environmental factors. In terms of the heat dissipation control strategy, the air-cooling intensity is intelligently adjusted according to the environmental humidity, which can prevent damage to the equipment caused by water vapor condensation and ensure the normal operation of the equipment in a high-humidity environment. The sensor acquisition time interval is dynamically adjusted, which not only reduces system resource consumption but also can respond in a timely manner when the heat dissipation requirements change. The precise regulation of each heat dissipation control link, such as the fan speed, coolant flow rate, and auxiliary heat dissipation device, enables the heat dissipation system to better adapt to different working states and environmental changes of the robot, improving the heat dissipation efficiency.

[0027] Other advantages, objectives, and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on the study of the following text, or can be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0029] Figure 1 It is a flowchart of an adaptive heat dissipation method for an intelligent orchard weeding robot. Specific embodiments

[0030] The following will clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0031] Embodiment 1

[0032] During the design and manufacture of an intelligent orchard weeding robot, temperature sensors are installed and arranged for its key heat-generating components such as motors and controllers. First, when the robot is in different working modes such as low-speed driving for weeding, high-speed turning, and climbing operations, an infrared thermal imager is used to conduct a comprehensive thermal imaging scan of parts such as the motor housing and the controller heat sink, collecting a large number of heat distribution data points.

[0033] Subsequently, an improved self-organizing map neural network (SOM) clustering algorithm is used to analyze these data points. Among them, the neuron connection weight update formula is Through this algorithm, data points with similar thermal characteristics are classified to determine the typical thermal distribution areas of the heat-generating components in different working modes. According to the positions and ranges of these areas, the optimal installation sites of temperature sensors are determined on the surfaces of the key heat-generating components, and initialization calibration is completed. At the same time, humidity sensors and wind speed sensors are installed at appropriate positions on the top and side of the robot respectively, establishing a stable data transmission connection with the robot control system.

[0034] When the robot conducts weeding operations in the orchard, each sensor collects data according to the dynamic adjustment strategy. For example, when the robot is in a stable straight-line weeding working state and the initial heat dissipation requirement level is low, the data collection time interval is set to once every 20 milliseconds. At this time, the temperature sensor collects the motor temperature of 45°C, the humidity sensor measures the ambient humidity of 60%, and the wind speed sensor detects the wind speed of 3 m / s. As the robot starts climbing operations and the working load increases, the heat dissipation requirement level rises, and the data collection time interval is based on the formula Shorten it to once every 5 milliseconds (where I0 is 20 milliseconds) to ensure accurate environmental and device status information can be obtained in a timely manner.

[0035] After the robot control system receives the sensor data, it uses a sensor fusion algorithm to calculate the heat dissipation demand index Q. The algorithm formula is Suppose that through previous tests and analyses of the heat dissipation characteristics of the robot, the weighting coefficients k1, k2, and k3 are determined (these weighting coefficients are obtained by constructing a multi-variable test matrix covering various combinations of temperature, humidity, and wind speed conditions, using the finite element analysis method to simulate the heat dissipation process and combining actual test data, and using the least squares optimization algorithm). Taking the currently collected data as an example, the Q value can be calculated. Then, according to the preset heat dissipation demand level classification standard, it is determined that the current heat dissipation demand level of the robot is moderately high.

[0036] According to the above-determined heat dissipation demand level, the control system automatically adjusts the heat dissipation strategy. For the adjustment of the fan speed, according to the formula N = N0×(1 + λ×(Q - Q0)) (assuming the initial fan speed N0 is 1000 revolutions per minute, the speed adjustment coefficient λ is 0.2, and the standard heat dissipation demand index Q0 corresponds to the initial state), the adjusted fan speed is calculated and adjusted. At the same time, considering the fan noise, an acoustic model constructed based on the aerodynamic acoustics principle of the fan blades is used for constraint optimization to ensure that the fan speed adjustment meets the heat dissipation requirements while the noise is within an acceptable range.

[0037] In terms of the adjustment of the coolant flow rate, a method based on fuzzy logic control is adopted. Define the heat dissipation demand index Q and the current coolant temperature T c as input variables, and the coolant flow rate adjustment ratio F as the output variable. Establish a fuzzy control rule table. When Q is high and T c is also high, F is large; when Q is low and T c is low, F is small. The fuzzification uses a triangular membership function, the inference uses the Mamdani fuzzy inference method, and the defuzzification uses the centroid method to obtain the accurate F value. Then, according to the formula ΔV = F×V0 (assuming the initial coolant flow rate V0 is 5 liters per minute), the actual coolant flow rate adjustment amount is calculated and the adjustment is executed.

[0038] If the heat dissipation demand further increases, start the heat dissipation fin deployment mechanism. Its deployment angle θ is calculated according to the formula θ = arctan(μ×(Q - Q0)) (assuming the angle adjustment coefficient μ is 0.3). By installing an angle sensor on the heat dissipation fins and using a servo control algorithm based on position feedback, an improved proportional-integral-derivative (PID) controller (the control law is Adjust the output torque of the servo motor according to the angular deviation to accurately reach the predetermined deployment angle of the heat dissipation fins. And when the ambient humidity exceeds the set threshold (such as 80%), appropriately reduce the air-cooling heat dissipation intensity to avoid water vapor condensation.

[0039] After the heat dissipation strategy is adjusted, the sensor continuously monitors the heat dissipation status of the robot. For example, an autoregressive integrated moving average (ARIMA) model is used to model the temperature data: Determine φ i , θ j , p and q values by fitting historical data, construct a data trend model to predict the change trend of the heat dissipation status. If it is predicted that the temperature will rise rapidly in the future, even if the current heat dissipation demand level has not reached a high level, appropriately increase the fan speed or coolant flow rate in advance. At the same time, compare the real-time monitored data with the preset target heat dissipation parameters. If a deviation is found, adjust the heat dissipation control strategy again according to the deviation situation to form a closed-loop feedback control.

[0040] In addition, during the entire operation process, the fault diagnosis and early warning system runs continuously. Based on the intelligent diagnosis algorithm of multi-source data fusion, this system maps the sensor data, control parameter data, and robot working state data to a high-dimensional feature space, and identifies and classifies fault modes through a support vector machine (SVM) classifier. Once it is found that the temperature sensor data fluctuates abnormally (such as the temperature suddenly rises or falls significantly and does not match the other sensor data and heat dissipation control strategy), or the control parameters such as the fan speed and coolant flow rate cannot be adjusted according to the predetermined strategy, it is immediately determined that there may be a potential fault. At this time, on the one hand, the system issues a fault early warning signal to notify the operator or maintenance system to check and repair; on the other hand, it automatically starts emergency heat dissipation measures, such as starting a standby cooling fan, increasing the pressure of the coolant storage tank to increase the coolant flow rate, etc., to maintain the normal operation of the robot.

[0041] The effects brought by this embodiment are as follows:

[0042] Through the application of this embodiment, the orchard intelligent weeding robot can achieve precise and efficient heat dissipation control in the complex and changeable orchard environment. The multi-sensor fusion technology ensures the accuracy of the judgment of heat dissipation requirements, avoiding the problems of untimely or excessive heat dissipation caused by the limitations of single-sensor information, effectively protecting the electronic components inside the robot and extending its service life. The adaptive heat dissipation control strategy can flexibly change whether it is the fan speed, the coolant flow rate or the adjustment of the auxiliary heat dissipation device according to the actual working state and environmental changes, greatly improving the heat dissipation efficiency and reducing energy consumption. In a high-humidity environment, the intelligent regulation of the air-cooling heat dissipation intensity effectively prevents the damage of water vapor condensation to the equipment, ensuring the stable operation of the robot. The continuous monitoring and feedback mechanism and the fault diagnosis and early warning system further enhance the reliability and maintainability of the robot, reducing the downtime caused by heat dissipation failures and improving the operation continuity.

[0043] Embodiment 2

[0044] In a large hilly orchard, the newly deployed orchard intelligent weeding robot needs to cope with complex terrains and changeable climate conditions. In the robot assembly workshop, technicians first install high-precision temperature sensors near the stator and rotor of the motor and at key heat-generating parts such as the power module of the controller. These parts are determined based on the thermal simulation analysis of the robot in the simulated orchard environment. The thermal simulation uses a complex model considering the thermal characteristics of various materials and the heat conduction path.

[0045] At the same time, a humidity sensor is installed on the top of the robot's shell, and its position is optimized through wind tunnel experiments to ensure accurate measurement of the surrounding environmental humidity and not be interfered by the robot's own air flow. A wind speed sensor is installed on the side of the robot in the forward direction, and its probe can automatically adjust the angle to adapt to different wind directions. After the installation is completed, all sensors are strictly calibrated. The calibration process uses a standard temperature and humidity source and a high-precision anemometer to ensure the accuracy of the sensor data, and a high-speed data transmission link with the robot's built-in control system is established, with a transmission rate of up to 100 Mbps to ensure real-time data transmission.

[0046] On a late spring and early summer morning, the temperature in the orchard gradually rises, the humidity is relatively high and there is a gentle breeze from time to time. The robot starts to work in the orchard, and the data acquisition system is activated. The sensors collect data at a predetermined dynamic time interval. Initially, since the robot is performing low-load flat-ground weeding operations and the environmental temperature is relatively mild, the data acquisition interval is set to once every 15 milliseconds. At this time, the temperature sensor collects the temperature of the key parts of the motor as 40 °C, the humidity sensor measures the environmental humidity as 70%, and the wind speed sensor detects the wind speed as 2.5 m / s.

[0047] As the robot enters a region with a relatively large slope for weeding operations, the workload increases sharply. Based on the preliminary calculation and changing trend of the heat dissipation demand index, the control system shortens the data acquisition interval to once every 3 milliseconds to more timely grasp heat dissipation-related information. For example, during high-load operations, the motor temperature rapidly rises to 55 °C, the humidity slightly increases to 75% due to the evaporation of vegetation moisture, and the wind speed decreases to 1.5 m / s due to terrain obstruction.

[0048] After the robot control system receives the sensor data, it calculates the heat dissipation demand index Q using a sensor fusion algorithm optimized for the orchard environment. In this algorithm, the weighting coefficients k1, k2, and k3 are obtained through a large number of on-site tests in different seasons and time periods in the orchard. For example, during the high-temperature and high-humidity period in summer, the influence weight k2 of humidity on heat dissipation will relatively increase, while in spring and autumn when it is relatively dry and the temperature is moderate, the temperature weight k1 will be more prominent. Assuming that in the current environment, k1 = 0.45, k2 = 0.35, k3 = 0.2, according to the formula the Q value is calculated, and based on the robot heat dissipation demand level classification table, it is determined that the current heat dissipation demand level is relatively high. Among them, the heat dissipation demand level classification table is obtained through performance tests and component life analysis of the robot under different heat dissipation demands, and is divided into four levels: low, medium, relatively high, and high. Each level corresponds to different heat dissipation control strategies.

[0049] According to the determined relatively high heat dissipation demand level, the control system starts to adjust the heat dissipation strategy. For the fan heat dissipation part, the adjustment of the fan speed is based on the formula N = N0×(1 + λ×(Q - Q0)), where the initial fan speed N0 = 800 revolutions per minute, the speed adjustment coefficient λ = 0.25, and the standard heat dissipation demand index Q0 corresponds to the Q value under normal working conditions. After calculating the new fan speed, the control system will also consider the relatively high current environmental humidity to avoid water vapor condensation and limit the increase amplitude of the fan speed. The limit coefficient is determined according to the humidity value and the pre-set humidity-speed limit curve.

[0050] In terms of the coolant circulation system, the adjustment of the coolant flow rate adopts a method based on fuzzy logic control. In addition to the heat dissipation demand index Q and the current temperature T of the coolant c as input variables, the environmental humidity H is also added as a correction variable. When the humidity is relatively high, even if the heat dissipation demand index is relatively high, the increase amplitude of the coolant flow rate will be appropriately reduced to prevent condensate from forming on the surface of the coolant pipeline. The fuzzy control rule table is formulated based on a large number of experimental data and expert experience. For example, when Q is relatively high, T c is relatively high and H is relatively high, the adjustment ratio F of the coolant flow rate is a medium increase; when Q is very high, T cWhen it is very high and H is moderate, F increases significantly. The fuzzification process uses a Gaussian membership function, the inference uses the Takagi-Sugeno-Kang (TSK) fuzzy inference method, and the defuzzification uses the weighted average method to obtain an accurate F value. Then, the actual coolant flow rate adjustment amount is calculated according to the formula ΔV = F × V0 (the initial coolant flow rate V0 = 4 liters per minute) and the adjustment is executed.

[0051] If the heat dissipation requirement further intensifies, the heat dissipation fin deployment mechanism is activated. Its deployment angle θ is calculated according to the formula θ = arctan(μ × (Q - Q0)). The angle adjustment coefficient μ = 0.35. During the deployment process, the servo control algorithm based on position feedback uses a high-precision optoelectronic encoder as the angle sensor, whose resolution can reach 0.1°. It real-time monitors the deployment angle of the heat dissipation fins and compares it with the target deployment angle. The servo motor uses a brushless DC motor with high torque density and fast response characteristics, and its control law is By adjusting the proportional coefficient K p 、the integral coefficient K i and the derivative coefficient K d , the heat dissipation fins can quickly and accurately reach the predetermined deployment angle and remain stable during the deployment process, avoiding damage to the heat dissipation fins caused by mechanical shock.

[0052] After the heat dissipation control strategy is adjusted, the continuous monitoring system closely monitors the heat dissipation status of the robot. A time series prediction model based on deep learning is used to predict the temperature data. This model uses a long short-term memory network (LSTM) structure. By learning a large amount of historical temperature data, it can accurately predict the temperature change trend in the next few minutes. For example, when it is predicted that the future temperature will continue to rise and may exceed the component safety temperature threshold, even if the current heat dissipation requirement level has not reached the highest level, the control system will increase the fan speed, increase the coolant flow rate in advance, and further deploy the heat dissipation fins to achieve preventive heat dissipation control.

[0053] Meanwhile, the fault diagnosis and early warning system analyzes the sensor data, heat dissipation control parameters, and the working state data of the robot in real time. When it is found that the data of the temperature sensor shows abnormal fluctuations, such as multiple jumps in temperature within a short period of time and does not match the data of other sensors and the heat dissipation control strategy, or when the heat dissipation control parameters such as the fan speed and the coolant flow rate cannot be adjusted according to the predetermined strategy, for example, the actual speed does not change after the fan speed command is issued or the coolant flow regulating valve gets stuck, the fault diagnosis and early warning system determines that there may be problems such as sensor faults, heat dissipation component faults, or control system faults. At this time, the system immediately issues a fault early warning signal. The early warning signal is sent to the monitoring terminal of the orchard management center through the wireless communication module, and at the same time, the fault code and brief fault information are displayed on the robot operation panel to facilitate the operator or maintenance personnel to quickly locate the fault. Moreover, the system automatically activates emergency heat dissipation measures, such as starting a standby heat dissipation fan (the standby fan is powered by an independent power supply to ensure normal operation in case of main power failure), increasing the pressure of the coolant storage tank to increase the coolant flow rate (achieved by starting a high-pressure pump, and the pressure of the high-pressure pump can be increased to 1.5 times the normal working pressure within a short period of time), etc., to maintain the normal operation of the robot and reduce the downtime caused by heat dissipation system faults and the impact on orchard weeding operations.

[0054] The effects brought by this embodiment are as follows:

[0055] In the application scenario of this embodiment, the orchard intelligent weeding robot demonstrates excellent adaptive heat dissipation ability. Through the carefully designed sensor installation and calibration, as well as the dynamic data acquisition strategy, it can accurately capture the changes in heat dissipation requirements of the robot in the complex hilly orchard environment. The weighted coefficients optimized for the orchard environment in the sensor fusion algorithm make the judgment of heat dissipation requirements more in line with the actual situation, effectively avoiding the problems of insufficient or excessive heat dissipation, ensuring the stable operation of the internal electronic components of the robot under different environmental conditions, and significantly extending the service life of the components.

[0056] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. Therefore, in any aspect, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, it is intended to embrace all changes falling within the meaning and scope of the same elements of the claims in the present invention. Any reference signs in the claims should not be regarded as limiting the claimed rights.

Claims

1. An adaptive heat dissipation method for an orchard intelligent weeding robot, characterized in that: The method comprises the following steps: S1. Sensor installation and initialization: Install temperature sensors, humidity sensors and wind speed sensors on the surface and external positions of the heating components of the orchard intelligent weeding robot, perform initialization calibration, and establish a data transmission connection with the robot control system; S2, data collection: during the operation of the robot, each sensor collects temperature data of the internal heating components of the robot, humidity data of the surrounding environment, and wind speed data at predetermined time intervals, and transmits them to the robot control system; S3. Sensor fusion and heat dissipation demand judgment: The robot control system uses a sensor fusion algorithm to process the received data. The sensor fusion algorithm calculates the heat dissipation demand index Q based on the following formula: Wherein, T is the average temperature of the heating component in °C, H is the ambient humidity in %, S is the ambient wind speed in m / s, k1, k2, k3 are weighted coefficients pre-set according to the heat dissipation characteristics of the robot, and k1+k2+k3=1. The calculated Q value determines the heat dissipation requirement level of the current robot according to the preset heat dissipation requirement level classification standard; S4, heat dissipation control strategy adjustment: According to the heat dissipation demand level, the robot control system automatically adjusts the heat dissipation control strategy, including adjusting the fan speed, coolant flow and starting the auxiliary heat dissipation device. When the humidity exceeds the set threshold, the air cooling intensity is reduced; S5. Continuous monitoring and feedback: After the heat dissipation control strategy is adjusted, the heat dissipation status of the robot is continuously monitored through sensors, and the real-time monitoring data is compared and analyzed with the preset target heat dissipation parameters. The heat dissipation control strategy is adjusted again according to the deviation to form a closed-loop feedback control.

2. The adaptive heat dissipation method for an orchard intelligent weeding robot according to claim 1, characterized in that: In S1, sensor installation and initialization step, the installation position of the temperature sensor is determined according to the heat distribution model of the heat generating component, and the heat distribution model is constructed in the following manner: When the robot is in different working modes, including low-speed weeding, high-speed steering, and climbing operations, an infrared thermal imager is used to perform thermal imaging scans on key heat-generating components to obtain heat distribution data points; A clustering algorithm based on a neural network is used to analyze and process these data points, grouping data points with similar thermal characteristics into one category to determine the typical thermal distribution areas of heat-generating components under different working modes; According to the clustering algorithm of these typical heat distribution networks, an improved self-organizing map neural network SOM is used, and the neuron connection weight update formula is: Among them, w ij (t) represents the connection weight between the i-th input data and the j-th neuron at time t, α(t) and β(t) are learning rate functions, which decrease over time, and h ij (t) is the neighborhood function, which indicates the influence of the jth neuron on the i-th input data, g ik (t) is the global influence function, which represents the global influence of the kth neuron on the i-th input data, x i is the i-th input data, and K is the total number of neurons.

3. The adaptive heat dissipation method for an orchard intelligent weeding robot according to claim 1, characterized in that: In step S2, data collection, the data collection time interval of each sensor adopts a dynamic adjustment strategy, and the dynamic adjustment of the data collection time interval is determined according to the following formula: Among them, I is the actual data collection time interval in milliseconds, I0 is the initially set data collection time interval in milliseconds, Q is the currently calculated heat dissipation demand index, Q0 is the preset standard heat dissipation demand index, when the robot is in a stable working state and the heat dissipation demand level is low, the data collection time interval is once every 20 milliseconds, when the robot is in a high-load working state and the heat dissipation demand level is high, the data collection time interval is once every 5 milliseconds.

4. The adaptive heat dissipation method for an orchard intelligent weeding robot according to claim 1, characterized in that: In S3, the sensor fusion and heat dissipation demand judgment step, the weighting coefficients k1, k2, and k3 are determined as follows: Construct a multivariate test matrix with different combinations of temperature, humidity and wind speed to cover the various operating conditions that may occur in an orchard environment; For each combination, the finite element analysis method is used to simulate the heat dissipation process of the robot under this working condition to obtain the corresponding theoretical heat dissipation demand value; The heat dissipation effect data of the robot obtained from the actual test under the same working conditions are compared with the theoretical heat dissipation demand value. The weighted coefficients k1, k2, and k3 are adjusted through the least squares optimization algorithm to minimize the error between the calculated heat dissipation demand index Q and the actual heat dissipation demand. The finite element analysis method is based on an improved heat conduction equation solution algorithm. The correction term of ambient humidity on the heat conduction coefficient is introduced into the traditional heat conduction equation. The corrected heat conduction equation is: Where ρ is the material density, c is the specific heat capacity, T is the temperature, t is the time, k(T,H) is the thermal conductivity coefficient considering the influence of humidity, and its expression is k(T,H)=k0×(1-γH)×(1+δT), k0 is the initial thermal conductivity coefficient, γ and δ are correction coefficients determined according to the material and environmental characteristics, and q is the internal heat source term.

5. The adaptive heat dissipation method for an orchard intelligent weeding robot according to claim 1, characterized in that: In S4, the heat dissipation control strategy adjustment step, the fan speed is adjusted according to the following formula: N = N0 × (1 + λ × (Q-Q0)), where N is the adjusted fan speed in revolutions per minute, N0 is the initial fan speed in revolutions per minute, λ is the speed adjustment coefficient determined based on the fan characteristics and the robot heat dissipation system, Q is the current heat dissipation demand index, Q0 is the preset standard heat dissipation demand index, and a speed adjustment threshold ΔN is set. When the calculated speed adjustment exceeds ΔN, the fan speed is actually adjusted. When adjusting the fan speed, an optimization algorithm based on an acoustic model is used. The acoustic model is constructed based on the aeroacoustic principle of fan blades. By analyzing the relationship between the shape, number, speed of the fan blades and the airflow velocity factors and the noise generation, the fan speed adjustment is constrained and optimized.

6. The adaptive heat dissipation method for an orchard intelligent weeding robot according to claim 1, characterized in that: In S4, the cooling control strategy adjustment step, the coolant flow rate is adjusted using a fuzzy logic control method, and the input variables are defined as the cooling demand index Q and the current coolant temperature T. c The output variable is the coolant flow adjustment ratio F. A fuzzy control rule table is established. When the heat dissipation demand index Q is high and the current coolant temperature T c When the cooling demand index Q is low and the current coolant temperature T is high, the coolant flow adjustment ratio F is large. c When it is low, the coolant flow adjustment ratio F is small. The fuzzification process uses a triangle membership function to fuzzy the input variables. The reasoning process uses the Mamdani fuzzy reasoning method. The defuzzification uses the centroid method to obtain the accurate coolant flow adjustment ratio F. The actual coolant flow adjustment amount ΔV is calculated according to the coolant flow adjustment ratio F: ΔV=F×V0, where V0 is the initial coolant flow rate in liters / minute.

7. The adaptive heat dissipation method for an orchard intelligent weeding robot according to claim 1, characterized in that: In S4, the heat dissipation control strategy adjustment step, when the auxiliary heat dissipation device is started to use the heat dissipation fin deployment mechanism, the control of its deployment angle θ is based on the following formula: θ=arctan(μ×(Q-Q0)), where μ is the angle adjustment coefficient determined according to the structural characteristics of the heat dissipation fins and the heat dissipation requirements of the robot, Q is the current heat dissipation requirement index, and Q0 is the preset standard heat dissipation requirement index. A servo control algorithm based on position feedback is adopted. By installing an angle sensor on the heat dissipation fin, the deployment angle of the heat dissipation fin is monitored in real time and compared with the target deployment angle. The output torque of the servo motor is adjusted according to the deviation so that the heat dissipation fin can reach the predetermined deployment angle. The servo control algorithm adopts an improved PID controller, and its control law is: Where, u(t) is the output torque of the servo motor, K p , K i , K d are the proportional, integral and differential coefficients respectively, e(t) is the angular deviation, and ω(t) is the compensation term used to overcome the influence of friction and inertia of the system.

8. The adaptive heat dissipation method for an orchard intelligent weeding robot according to claim 1, characterized in that: In step S5, continuous monitoring and feedback, a predictive heat dissipation control strategy based on data trend analysis is adopted to perform time series analysis on continuously collected heat dissipation-related data including temperature, humidity, wind speed, and heat dissipation control parameters, build a data trend model, and use the autoregressive moving average ARIMA model to model the temperature data: Among them, T t is the temperature value at time t, and θ j are the autoregressive coefficient and the moving average coefficient respectively, L is the lag operator, ∈ t is a white noise sequence, p and q are the model orders, which are determined by fitting historical data. θ j , p, and q, and uses the constructed data trend model to predict the changing trend of the heat dissipation conditions in the future. When it is predicted that the temperature will rise rapidly in the future, the fan speed and coolant flow rate will be appropriately increased in advance, even if the current heat dissipation demand level has not yet reached a high level.

9. The adaptive heat dissipation method for an orchard intelligent weeding robot according to claim 1, characterized in that: A fault diagnosis and early warning system is established during the operation of the entire heat dissipation method. The system determines whether there are hidden dangers of failure in the heat dissipation system through comprehensive analysis of sensor data, heat dissipation control parameters and robot working status data. When the temperature data collected by the temperature sensor fluctuates abnormally and does not match other sensor data and heat dissipation control strategies, when the heat dissipation control parameters of fan speed and coolant flow cannot be adjusted according to the predetermined strategy, the fault diagnosis and early warning system determines that there may be sensor failure, heat dissipation component failure and control system failure. At this time, the system sends a fault warning signal to notify the operator to check and repair. At the same time, the heat dissipation control strategy is automatically adjusted and emergency heat dissipation measures are taken, including starting the backup cooling fan and increasing the pressure of the coolant reserve tank. The fault diagnosis and early warning system adopts an intelligent diagnosis algorithm based on multi-source data fusion. The algorithm maps sensor data, control parameter data and robot working status data to a high-dimensional feature space, and identifies and classifies different fault modes through the support vector machine SVM classifier. Its classification decision function is: Among them, x is the data sample to be diagnosed, x i is the training sample, y i is the category label of the training sample, α i is the Lagrange multiplier, K(x i ,x) is the kernel function, b is the bias term, and n is the number of training samples.

Images