A miniature automatic weeding device for rice fields and a control method thereof

Abstract

The application discloses a kind of paddy field micro automatic weeding equipment and control method thereof, belong to agricultural machinery control technical field.The equipment includes floating cavity, airflow generating device, flow guide pipe, buoyancy adjusting device, driving device, power supply device, positioning device and control system and water quality detector.Combination of intelligent control system and airflow injection technology, utilize vortex effect to remove weeds in rice field, and dynamically adjust operating state to adapt to different water conditions.Control method includes vortex effect analysis, key parameter set determination, trajectory planning generation, weed dense distribution point identification and best power output calculation and other steps.The application realizes efficient, environmentally friendly automatic weeding, significantly reduces manual labor intensity, avoids chemical pollution, and provides an optimized solution for weed control in rice fields.

Description

Technical Field

[0001] This invention belongs to the field of agricultural machinery control technology, specifically, it relates to a miniature automatic weeding device for paddy fields and its control method. Background Technology

[0002] The growth of weeds in paddy fields has a significant impact on the growth and development of rice. They compete with rice for nutrients, water, and sunlight, directly leading to reduced rice yield and quality. Currently, the main methods for controlling weeds in paddy fields are manual weeding and chemical weeding.

[0003] Manual weeding, as a traditional method, has the advantages of being intuitive and leaving no chemical residue, but it is labor-intensive and relatively inefficient. In large-scale planting scenarios, manual weeding often requires a large investment of manpower, which may not be able to fully meet the actual needs.

[0004] Chemical weeding can achieve high weed control efficiency by using chemical agents, but this method has certain limitations. For example, the use of chemical agents may pollute the environment and disrupt the ecological balance. At the same time, some agents may cause phytotoxicity to rice, affecting its normal growth and final yield.

[0005] Against this backdrop, developing an efficient and environmentally friendly rice paddy weeding device has become an urgent problem to be solved. This device needs to reduce the intensity of manual labor while avoiding adverse effects on the environment and crops, thus providing a more optimized solution for rice paddy weed control. Summary of the Invention

[0006] The purpose of this invention is to provide a micro-automatic weeding device for paddy fields and its control method, which mainly solves the problems of high labor intensity, environmental pollution from chemical agents, and low efficiency of traditional machinery in existing paddy field weed removal methods.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A miniature automatic weeding device for paddy fields includes a floating cavity that floats on the water surface, an airflow generator disposed inside the floating cavity, a guide pipe connected to the outlet end of the airflow generator and extending out from the bottom of the floating cavity, a buoyancy adjustment device disposed inside the floating cavity, a drive device disposed at the tail end of the floating cavity, a power supply device disposed inside the floating cavity, and a positioning device and a control system disposed inside the floating cavity.

[0009] Furthermore, in this invention, the output end of the airflow generating device is designed as a porous diffuser, and the outside of the porous diffuser is covered with a removable filter screen; the guide pipe is made of flexible material, one end of which is sealed to the porous diffuser, and the other end extends to the outside of the floating cavity and points towards the underwater area of ​​the paddy field; a pressure sensor and a flow meter are installed between the porous diffuser and the guide pipe of the airflow generating device. The pressure sensor is used to detect the output pressure of the airflow generating device, and the flow meter is used to measure the actual flow rate of the airflow. A water quality detector extending to the underwater area of ​​the paddy field is also installed in the floating cavity to detect the turbidity of the paddy field water. The pressure sensor, flow meter and water quality detector transmit the collected data to the control system in the floating cavity.

[0010] Furthermore, in this invention, the buoyancy adjustment device consists of several cavities filled with a liquid medium of adjustable density, and the inflow and outflow of the liquid medium are controlled by built-in valves and diaphragm pumps.

[0011] Furthermore, in this invention, the top of the floating cavity is also provided with a solar panel connected to the power supply device.

[0012] Based on the above-mentioned automatic weeding device, the present invention also provides a control method for a micro automatic weeding device for paddy fields, comprising the following steps:

[0013] S1, through the control system, analyzes the formation law of vortex effect in water based on the water flow characteristics, and dynamically adjusts the equipment operation status by monitoring water quality parameters in real time through a water quality analyzer;

[0014] S2, based on the vortex intensity distribution of the integrated water vortex effect and the water quality parameter change curves obtained from the dynamic water quality detection, determine the key parameter set for equipment operation;

[0015] S3, using the set of key parameters as input variables, generate a device operation trajectory planning scheme based on the improved particle swarm optimization algorithm;

[0016] S4. Based on the location information and operational requirements in the equipment operation trajectory planning scheme, identify the dense distribution points of weeds in the underwater area of ​​the paddy field;

[0017] S5. Establish a data-guided, mechanism-driven weeding effect evaluation model, input the dense weed distribution points into the weeding effect evaluation model, and calculate the optimal power output for equipment operation;

[0018] S6, by adjusting the jet intensity of the airflow generator through the optimal power output, the automatic weeding task in the underwater area of ​​the paddy field is completed.

[0019] Furthermore, in step S1, the analysis process of the eddy effect in the water body is as follows:

[0020] S11, Calculate the airflow velocity distribution at the nozzle outlet of the guide pipe, and determine the intensity index V of the vortex effect in the water body based on the airflow velocity distribution. t , where V t The mathematical expression is:

[0021]

[0022] Where V0 is the initial airflow velocity at the nozzle outlet of the guide tube, d is the distance between the nozzle and the water surface, and k1 and k2 are the airflow attenuation coefficient and diffusion coefficient, respectively.

[0023] S12, through the strength index V t The vortex intensity distribution function F(V) is obtained. t Its expression is:

[0024] F(V t )=∫V t dt

[0025] In the formula, t is the time variable;

[0026] The process of dynamically adjusting the equipment's operating status is as follows:

[0027] Dissolved oxygen concentration (DO), turbidity (T), and conductivity (EC) collected by a water quality analyzer are used as input features. These input features are normalized, and an adaptive fuzzy clustering algorithm is used to calculate the water quality characteristic change label Q. w The water quality parameter change curve C of the dynamic adjustment model (2) is calculated based on the water quality characteristic change labels. w (t):

[0028] C w (t)=Q w ·(DO(t)+T(t)+EC(t)).

[0029] Further, in step S2, the set of key parameters is:

[0030] P key ={V t_max C w_avg ,Δt}

[0031] Among them, V t_max C represents the maximum value of the vortex intensity. w_avg Δt represents the average value of the water quality parameter variation curve, and Δt represents the time interval of equipment operation.

[0032] Furthermore, in step S3, the process of generating the trajectory planning scheme is as follows:

[0033] S31, Preprocess the key input parameters: complete the standardization of the key parameter set and the correlation analysis of trajectory planning;

[0034] S32 improves upon the traditional particle swarm optimization algorithm to address the dynamic nature of the underwater environment in paddy fields.

[0035] S33, establish the objective function of the running trajectory, and establish the constraint objective of minimizing the comprehensive weeding efficiency and energy consumption;

[0036] S34 iteratively solves the objective function under constraints, and finally outputs the trajectory planning scheme.

[0037] Furthermore, in step S4, the method for identifying densely distributed weed locations is as follows:

[0038] Calculate the gradient of water quality parameter changes along the trajectory of the computing device. Based on the aforementioned gradient change, the criteria for determining densely distributed weed points (4) are constructed:

[0039]

[0040] Wherein, θ is a preset threshold; the area that meets the determination condition is defined as a densely distributed weed point.

[0041] Furthermore, in step S5, the process of calculating the optimal power output is as follows:

[0042] Obtain the area S and water depth H of the densely distributed weeds, and combine them with the airflow vortex intensity distribution function F(V) t Establish the objective function E of the weed control effect evaluation model. eff ;

[0043] E eff =S·H·F(V t )

[0044] The optimal power output P for device operation is calculated based on the objective function. opt :

[0045] P opt =Eeff / (η·Δt)

[0046] Where η is the energy conversion efficiency of the device; when the optimal power output exceeds the rated power range of the device, the power output is corrected, and the correction process is as follows:

[0047] P corr =P opt ·(1-α)

[0048] Where α is the power correction coefficient, and its value ranges from [0, 0.5].

[0049] The design principle of this invention lies in using an airflow generator to produce high-speed gas, which is then sprayed into the paddy field through a guide pipe. This causes the water in the paddy field to change from clear to turbid, thus blocking sunlight from reaching the water surface and preventing underwater weeds from photosynthesizing. By maintaining the paddy field water in a turbid state for a certain period, the underwater weeds will naturally die due to the inability to photosynthesize. If the device provided by this invention is installed from the initial planting of rice seedlings, weeds in the paddy field will never grow. Furthermore, after the weeds die naturally underwater, they decompose into nutrients, directly fertilizing the field and providing the rice with more nutrients, promoting better rice growth.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] (1) The equipment provided by this invention is compact in size, with a minimum size of about 20cm in length and 5cm in width, allowing it to move freely and operate flexibly in paddy fields. As it is an automatic weeding device, an appropriate number of devices can be deployed according to the size of the paddy field and the urgency of weeding, enabling multiple devices to work simultaneously, forming a cluster effect. The weeding efficiency and frequency can be freely adjusted. Compared to conventional manual weeding and chemical weeding, the larger the paddy field area, the more obvious the technological and cost advantages.

[0052] (2) The present invention removes weeds by spraying air into the water to make the water turbid and block sunlight, thereby cutting off the conditions for weeds to carry out photosynthesis. It is a fully automatic method that does not require manual labor or chemical agents. It is highly efficient, avoids the pollution of the environment caused by chemical agents, protects the ecological balance, reduces the risk of pesticide damage to rice, and ensures the quality and yield of rice.

[0053] (3) This invention effectively combines artificial intelligence with agricultural machinery, promoting the technological, intelligent, and automated development of agriculture. At different stages of rice growth, the water depth in the paddy field varies, the row spacing between adjacent rice plants differs, the area of ​​sunlight naturally reaching the water surface also varies, and the frequency and growth rate of weeds also differ. In other words, the natural environment of the paddy field changes as the rice grows. Therefore, to achieve continuous and effective weed control, the weeding equipment also needs to be able to adapt to changes in the paddy field environment. This invention utilizes the algorithm built into the control system to determine the optimal operating state of the equipment based on the natural environmental parameters of the paddy field, and then adjusts the parameters of the airflow generator and buoyancy adjustment device to match the current natural environment of the paddy field. This allows it to meet the weeding needs of rice at different stages, greatly improving the adaptability and work efficiency of the equipment.

[0054] (4) This invention combines a positioning device with a control system, which allows for the customization of the equipment's movement trajectory in paddy fields and enables automatic pathfinding. Compared to existing mechanical weeding equipment, the level of automation and intelligence is greatly improved.

[0055] (5) This invention utilizes solar energy to supplement battery power, realizing the sustainable use of energy, reducing the operating cost of the equipment, and has good economic and environmental benefits. Attached Figure Description

[0056] Figure 1 This is a schematic diagram of the external structure of the micro automatic weeding device for paddy fields according to the present invention.

[0057] Figure 2 This is a schematic diagram of the internal structure of the micro automatic weeding device for paddy fields according to the present invention.

[0058] Figure 3 This is a flowchart of the control method for the micro automatic weeding device for paddy fields according to the present invention.

[0059] The attached figures are labeled as follows:

[0060] 1. Floating cavity; 2. Airflow generator; 3. Guide pipe; 4. Buoyancy adjustment device; 5. Drive device; 6. Power supply device; 7. Positioning device and control system; 8. Porous diffuser; 9. Water quality analyzer; 10. Solar panel. Detailed Implementation

[0061] The present invention will be further described below with reference to the accompanying drawings and embodiments. The embodiments of the present invention include, but are not limited to, the following embodiments.

[0062] Example

[0063] like Figure 1 and Figure 2 As shown, the micro automatic weeding device for paddy fields disclosed in this invention includes: a floating cavity 1, an airflow generating device 2, a guide pipe 3, a buoyancy adjusting device 4, a driving device 5, a power supply device 6, a positioning device, a control system 7, a water quality analyzer 9, and a solar panel 10.

[0064] The floating cavity 1 is the core component of the equipment, designed to float and operate stably on the water surface. It integrates multiple functional modules for overall control and operation of the equipment. A solar panel 10 is mounted on the top of the floating cavity 1 to convert solar energy into electricity, providing power to the equipment.

[0065] An airflow generator 2 is installed inside the floating cavity 1 to generate high-speed airflow. This embodiment uses a small vortex air pump (power range 50-150W), which features small size, stable airflow, and low noise, meeting the miniaturization requirements of the equipment. Specifically, the output end of the airflow generator 2 is designed as a porous diffuser 8, covered with a removable filter screen to prevent impurities from entering and causing blockage. A pressure sensor and flow meter are installed between the porous diffuser 8 and the guide pipe 3 to monitor the airflow pressure and flow rate in real time. The guide pipe 3 is made of flexible material, with one end sealed to the porous diffuser 8 and the other end extending outside the floating cavity 1 and pointing towards the underwater area of ​​the paddy field. The design of the guide pipe 3 ensures that the airflow can be effectively delivered underwater. The jet airflow generates a vortex effect, making the water turbid and blocking sunlight, thereby cutting off the conditions for weeds to photosynthesize and thus eliminating weeds.

[0066] The buoyancy adjustment device 4 consists of several chambers filled with a liquid medium of adjustable density. The inflow and outflow of the liquid medium are controlled by built-in valves. The buoyancy adjustment device 4 is used to dynamically adjust the buoyancy of the equipment in water, ensuring stable operation under different water depths and conditions. Multiple chambers can independently adjust the amount of liquid medium injected, achieving step-by-step fine-tuning of buoyancy. For example, when the equipment operates at different water depths (such as the water level in paddy fields at different growth stages of rice), by controlling the increase or decrease of liquid in some chambers, the target buoyancy can be precisely matched, avoiding sudden changes in buoyancy caused by adjusting a single chamber. The chambers are symmetrically distributed around the geometric center of the equipment (e.g., front-to-back, left-to-right layout). By differentially controlling the liquid volume in each chamber, the center of gravity of the equipment can be dynamically adjusted. For example, when the equipment tilts due to weed accumulation on one side, increasing the liquid mass in the other chamber restores its horizontal posture and prevents it from tipping over. Each chamber is equipped with an electromagnetic proportional valve (response time ≤ 0.5s), whose opening degree (0-100%) is independently controlled by the control system based on parameters such as water depth and flow velocity fed back by the positioning device and water quality analyzer. For example, when an increase in water depth is detected, the control system commands the valve to open, injecting a liquid medium (e.g., density 1.05g / cm³) into the chamber. 3 The equipment uses a brine solution to increase its weight and reduce buoyancy; conversely, it discharges liquid to increase buoyancy. The equipment incorporates a miniature diaphragm pump, connected to all chambers via a main pipeline, enabling cross-chamber dispensing of liquid media. For example, when adjusting the longitudinal balance of the equipment, the pump draws liquid from the front chamber into the rear chamber, altering the front-to-rear weight distribution to accommodate attitude shifts caused by water flow impact.

[0067] The drive unit 5 is located at the tail of the floating cavity 1 and is used to propel the equipment to move in the paddy field. The specific form of the drive unit 5 can be a propeller or a jet propulsion device to ensure that the equipment can navigate autonomously.

[0068] A power supply unit 6 is located inside the floating cavity 1 to provide power to the equipment. The power supply unit 6 includes a battery pack and a power management system. A solar panel 10 is installed on the top of the floating cavity 1 to convert solar energy into electrical energy, achieving self-sufficient energy supply. A positioning device and control system 7 are used to determine the location of the equipment and control its operation. The control system receives data from pressure sensors, flow meters, and water quality analyzers 9, analyzes and makes decisions, and dynamically adjusts the equipment's operating status.

[0069] The water quality analyzer 9 extends into the underwater area of ​​the paddy field to monitor water quality parameters in real time (such as dissolved oxygen concentration DO, turbidity T, conductivity EC, etc.); the data collected by the water quality analyzer 9 is transmitted to the control system to optimize the equipment's operating status.

[0070] The equipment achieves automated weeding in rice paddies through an intelligent operating process. For example... Figure 3 As shown, the specific operation process is as follows: First, based on the water flow characteristics (the jet airflow causes water flow, while paddy field water itself is generally static), the control system analyzes the formation law of the vortex effect in the water body and calculates the airflow velocity distribution at the nozzle outlet of guide pipe 3; based on the airflow velocity distribution, the intensity index V of the vortex effect in the water body is determined. t , where V t The mathematical expression is:

[0071]

[0072] Where V0 is the initial airflow velocity at the nozzle outlet of the guide pipe, d is the distance between the nozzle and the water surface, and k1 and k2 are the airflow attenuation coefficient and diffusion coefficient, respectively. The intensity index V... t The vortex intensity distribution function F(V) is obtained. t Its expression is:

[0073] F(V t )=∫V t dt

[0074] In the formula, t is the time variable.

[0075] Dissolved oxygen concentration (DO), turbidity (T), and conductivity (EC) collected by the water quality analyzer 9 are used as input features. These input features are normalized, and an adaptive fuzzy clustering algorithm is used to calculate the water quality characteristic change label Q. w Based on the water quality characteristic change labels, the water quality parameter change curve C of the dynamic adjustment model is calculated. w (t):

[0076] C w (t)=Q w·(DO(t)+T(t)+EC(t)).

[0077] Among them, the water quality change label Q w The calculation process requires the integration of the core logic of the Adaptive Fuzzy Clustering (AFC) algorithm. Its core is to generate quantitative labels reflecting changes in water quality characteristics by dynamically adjusting clustering parameters, handling uncertainties, and tracking time-varying data. The following are the specific calculation steps (using time series data t1, t2, ..., t...). n For example:

[0078] First, the time series data for dissolved oxygen (DO), turbidity (T), and conductivity (EC) are normalized to eliminate dimensional differences. A common method is min-max normalization (assuming no outliers in the data):

[0079]

[0080] Where X represents any one of the parameters DO, T, and EC, X(t) is the original value at time t, and min(X) and max(X) are the minimum and maximum values ​​of that parameter in the historical data. The normalized data range is [0,1], denoted as: X(t) = [DO...]. norm (t),T norm (t),EC norm (t)].

[0081] Initialization of Adaptive Fuzzy Clustering (AFC): The "adaptive" aspect of AFC is primarily reflected in dynamically determining the initial clustering parameters (cluster number, fuzzy index, m), rather than pre-setting fixed values. The specific method is as follows:

[0082] (1) Automatically determine the number of clusters c

[0083] By analyzing the distribution characteristics (such as density and dispersion) of normalized data, the optimal number of clusters can be selected using the silhouette coefficient method or the DB index method.

[0084] Silhouette coefficient: Calculates the average distance (cohesion) between each sample and samples in the same cluster and the average distance (dissociation) between each sample and the nearest sample in a different cluster. The larger the silhouette coefficient s(i), the better the clustering effect.

[0085] DB index: measures the ratio of inter-cluster distance to intra-cluster distance. The smaller the index, the better the clustering effect. Finally, the c that maximizes the silhouette coefficient (or minimizes the DB index) is selected as the initial number of clusters (e.g., c=3, corresponding to the three categories of "clean", "lightly polluted" and "heavily polluted").

[0086] (2) Dynamically set the fuzzy index m

[0087] The fuzziness index m (m>1) controls the "fuzziness" of membership: the larger m is, the smoother the membership distribution (more fuzzy); the smaller m is, the closer the membership is to hard clustering (more stringent). AFC dynamically adjusts m based on the data noise level.

[0088] If the data is noisy (e.g., the turbidity T fluctuates wildly), increase m (e.g., m = 2.5) to allow for more fuzzy membership assignments;

[0089] If the data is stable (e.g., the conductivity EC changes slowly), decrease m (e.g., m = 1.5) to enhance the discriminative power of the clusters.

[0090] Iterative optimization of clustering model: AFC updates cluster centers v iteratively by minimizing the objective function J. j and membership matrix u ij (where i is a sample and j is a cluster), the objective function is defined as:

[0091]

[0092] Where ||·|| is the Euclidean distance, v j The initial value can be randomly selected or set based on the data distribution to represent the cluster center of the j-th class.

[0093] The iterative process is as follows:

[0094] 1. Calculate the membership degree u ij :

[0095]

[0096] 2. Update cluster center v j :

[0097]

[0098] 3. Repeat steps 1-2 until the change in J is less than the set threshold (e.g., 10). 5 (or reach the maximum number of iterations.)

[0099] Dynamically adjust clustering parameters (adaptive core):

[0100] The key to AFC's "adaptive" nature lies in dynamically adjusting clustering parameters based on new data to cope with the time-varying nature of water quality (such as seasonal changes and sudden pollution):

[0101] (1) Dynamic adjustment of the number of clusters c

[0102] When the new data X(t) n +1) When inputting, calculate its distance to all existing cluster centers:

[0103] If all distances are greater than a threshold (e.g., 1.5 × average intra-cluster distance), then add a new cluster (c ← c + 1) and set X(t) to the threshold. n +1) serves as the initial center of the new cluster;

[0104] If no new samples are added to a cluster for an extended period (e.g., no samples have a membership degree exceeding 0.5 for 10 consecutive time points), then that cluster is deleted (c←c-1).

[0105] (2) Dynamic adjustment of the fuzzy index m

[0106] Adjust m by monitoring the local density of the data (such as the average distance of the nearest neighbor samples):

[0107] If the local density is low (data is sparse), increase m to allow for more fuzzy membership degrees;

[0108] If the local density is high (in the data set), decrease m to enhance the discriminative power of the cluster.

[0109] Generate water quality change label Q w (t)

[0110] Based on the membership matrix u ij (t) and cluster center v j (t), Q w (t) is defined as the weighted sum of the membership degree of each cluster and the water quality state represented by the cluster. The specific steps are as follows:

[0111] 1. Define the water quality state value of the cluster: based on the cluster center v j The original parameters (after denormalization) corresponding to (t) are assigned a state value s for each cluster j. j (e.g., "clean" s1 = 0.8, "lightly polluted" s2 = 1.2, "heavily polluted" s3 = 1.5);

[0112] 2. Calculate Q w (t):

[0113]

[0114] Where u ij (t) represents the membership degree of sample i to cluster j at time t (i can be omitted if it is a single sample input).

[0115] Q w The larger the value of (t), the stronger the trend of water quality towards "pollution" or "fluctuation"; conversely, the smaller the value, the more likely it is to be "stable" or "clean".

[0116] Based on the eddy intensity distribution of the combined water vortex effect and the water quality parameter variation curves obtained from dynamic water quality monitoring, the key parameter set for equipment operation is determined; the form of the key parameter set is as follows:

[0117] P key ={v t_max C w_avg ,Δt}

[0118] Among them, V t_max C represents the maximum value of the vortex intensity. w_avg Δt represents the average value of the water quality parameter variation curve, and Δt represents the time interval of equipment operation.

[0119] A device trajectory planning scheme is generated based on a set of key parameters.

[0120] First, the key input parameters are preprocessed, and the key parameter set P is... key ={V t_max C w_avg Convert ,Δt} into a normalized form that the algorithm can process, and analyze its relationship with trajectory planning:

[0121] Maximum vortex intensity V t_max This reflects the maximum disturbance capability of the nozzle's jet airflow on the water body, directly affecting the equipment's operating efficiency in different areas (e.g., V). t_max The larger the size, the wider the area it can cover for weeds;

[0122] Average value C of water quality parameter variation curve w_avg Characterizing the overall water quality stability of paddy fields (C w_avg The higher the elevation, the more drastic the water quality fluctuations (potentially corresponding to areas with dense weeds);

[0123] Time interval Δt: The time step for continuous operation of the equipment, which needs to be matched with the time characteristics of the vortex effect (such as the vortex decay period) to avoid repeated operations or missed areas.

[0124] By using linear standardization (such as z-score normalization), the three values ​​are unified to the range [0,1], denoted as... As input variables for the particle swarm optimization algorithm.

[0125] Traditional particle swarm optimization (PSO) searches for the optimal solution through information sharing among particles. However, considering the dynamic nature of the underwater environment in paddy fields (such as water flow disturbance and changes in weed distribution), the following improvements are needed:

[0126] (1) Representation of particle position and velocity

[0127] Particle position X i : Represents the trajectory point of the device during the time interval Δt, defined as X i ={(x1,y1,t1),(x2,y2,t2),…,(x k ,y k ,t k)}, where (x j ,y j Let t be the two-dimensional coordinates of the j-th time point. j = t0 + j·Δt (t0 is the initial time);

[0128] Particle velocity V i : Represents the rate of change of the trajectory points, defined as X i ={Δx1,Δy1,Δx2,Δy2,…,Δx k ,Δy k}, with constraints |Δx j |≤vm ax· Δt、|Δy j |≤v max ·Δt(v max (This refers to the maximum moving speed of the equipment).

[0129] (2) Dynamic inertia weight and acceleration coefficient

[0130] An adaptive weight adjustment strategy is introduced to balance the capabilities of global search (exploring new regions) and local search (refining the trajectory):

[0131] Inertia weight ω: initially set to ω max =0.9, decreasing linearly with the number of iterations g to ω min =0.4, the formula is: (G max (Maximum number of iterations);

[0132] Acceleration coefficients c1 and c2: respectively control the particle's trajectory towards its historical optimal value (p). best ) and global optimum (g best The learning weights are dynamically adjusted to c1. (The early stage focuses on overall exploration, while the later stage focuses on local optimization).

[0133] 3. Objective Function Construction: Combining Weeding Efficiency and Energy Constraints

[0134] The goal of trajectory planning is to minimize device power consumption while covering areas with dense weeds. The objective function is f(X). i The sum of the following four sub-objectives is defined as:

[0135] (1) Weed coverage F cov

[0136] The trajectory coverage identification requires densely distributed weed points (satisfying) (The area). Let S be the set of densely weeded areas. weed The trajectory covers an area of ​​S. path ,but:

[0137]

[0138] Weight w1 = 0.4 (prioritizes covering weedy areas).

[0139] (2) Path smoothness F smooth

[0140] To avoid frequent equipment turning (reducing mechanical wear), utilize the curvature of the trajectory points. Calculate and take the mean curvature but:

[0141]

[0142] Weight w2 = 0.2 (The maximum average curvature allowed by the device).

[0143] (3) Energy consumption cost F energy

[0144] Based on optimal power output P opt Calculate the total energy consumption E based on the time interval Δt. total =∑P opt ·Δt, after normalization:

[0145]

[0146] Weight w3 = 0.3 (E max (Maximum battery capacity of the device).

[0147] (4) Vortex effect matching degree F vortex

[0148] Ensure the trajectory matches the vortex intensity distribution F(Vt) (e.g., in V). t_max Increased stay time in the area is defined as:

[0149]

[0150] Weight w4 = 0.1(Δt) j For the trajectory in V t_max (Duration of stay in the area).

[0151] Final objective function:

[0152] f(X i )=w1·F cov +w2·F smooth +w3·F energy +w4·F vortex

[0153] 4. Optimize the iterative process

[0154] The iterative steps of the improved particle swarm optimization algorithm are as follows (with N particles, G... max (Taking the next iteration as an example):

[0155] (1) Initialize the particle swarm: Randomly generate the initial positions of N particles. and speed Ensure the location is within the paddy field operation area (x∈[0,L]y∈[0,W], where L and W are the length and width of the paddy field);

[0156] (2) Calculate the initial fitness: For each particle i, calculate the objective function value. Record individual historical best (Corresponding to the maximum f value), globally optimal

[0157] (3) Iterative update (loop from g=1 to G) max ):

[0158] a. Update speed: (r1, r2 are random numbers in the range [0, 1]).

[0159] b. Speed ​​limit: For Beyond ±v max The components of Δt are truncated to boundary values;

[0160] c. Update location: And correct the location of the boundary violation (e.g., if it exceeds the rice paddy area, reflect it back to the boundary);

[0161] d. Update the optimal value: If but like but

[0162] (4) Termination condition: When the number of iterations reaches G max or g best If the fitness value does not change significantly for 10 consecutive iterations (e.g., change < 0.01), the iteration stops.

[0163] 5. Trajectory planning scheme output

[0164] The final output is the globally optimal particle g. best Corresponding trajectory Includes the following information:

[0165] Time series location: coordinates within each Δt Ensure coverage of areas with dense weeds and that the path is smooth;

[0166] Speed ​​command: Movement speed of adjacent trajectory points No more than v max .

[0167] Operation mode: in V t_max Reduce speed (extend residence time) in the area (where the vortex intensity is greatest) to enhance weed control.

[0168] Finally, based on the location information and operational requirements in the trajectory planning scheme, the equipment calculates the gradient of water quality parameter changes along the operating trajectory. Determination criteria for dense weed distribution points based on gradient changes. (where θ is a preset threshold), and the area that meets the judgment condition is defined as a densely distributed weed point.

[0169] After identifying densely distributed weeds, a data-driven mechanism-based weed control effectiveness evaluation model is established. The densely distributed weeds are input into the model to calculate the optimal power output for equipment operation. The area S and water depth H of the densely distributed weeds are obtained, and the objective function E of the weed control effectiveness evaluation model is established by combining this with the airflow vortex intensity distribution function F(Vt). eff Its form is E eff =S·H·F(V t The optimal power output P for device operation is calculated based on the objective function. opt Its form is P opt =E eff / (η·Δt) (where η is the energy conversion efficiency of the equipment); when the optimal power output exceeds the rated power range of the equipment, the power output is corrected, and the correction formula is P. corr =P opt • (1-α) (where α is the power correction coefficient, with a value range of [0, 0.5]). By adjusting the jet intensity of the airflow generator 2, the automatic weeding task in the underwater area of ​​the paddy field is completed.

[0170] The actual operating scenario of the equipment is as follows: In a paddy field environment, the equipment floats on the water surface through the floating cavity 1. The solar panel 10 absorbs sunlight and converts it into electrical energy, which is stored in the power supply device 6 to provide continuous power support for the equipment. After the equipment is started, the airflow generator 2 starts working, generating a high-speed airflow. The airflow is diffused through the porous diffuser 8 and transmitted to the underwater area of ​​the paddy field through the guide pipe 3. The design of the guide pipe 3 ensures that the airflow can accurately act on the underwater area, using the vortex effect to make the water turbid and block sunlight, thereby cutting off the conditions for weeds to photosynthesize and thus clearing weeds. Pressure sensors and flow meters monitor the pressure and flow rate of the airflow in real time and transmit the data to the positioning device and control system 7. The control system analyzes water quality parameters such as dissolved oxygen concentration (DO), turbidity (T), and conductivity (EC) collected by the water quality analyzer 9, and dynamically adjusts the equipment's operating status to adapt to the operational needs under different water conditions. The buoyancy adjustment device 4 changes the buoyancy of the equipment by adjusting the density of the liquid medium in the cavity, ensuring that the equipment maintains stable operation under different water depth conditions. According to the trajectory planning scheme, the drive unit 5 propels the equipment to move along the predetermined path to complete the comprehensive weeding task in the underwater area of ​​the paddy field.

[0171] Calculations show that 1-3 devices are needed for one acre of paddy field. This device is equipped with two 6000mAh batteries, allowing it to operate continuously for about 5 hours without sunlight. With sunlight, it can operate continuously. When the battery is low, the device can automatically move to a sunny area to recharge when its power level drops below 10%. Alternatively, it can be equipped with a wireless network and a monitoring terminal to send an alarm to the monitoring terminal, reminding it to manually charge the battery or replace it. The device itself has a very low cost; once invested, it can be used repeatedly for a long time without further investment. Compared to manual weeding, it is more efficient and cheaper; compared to chemical pesticides, it is more environmentally friendly, non-toxic, and harmless.

[0172] Meanwhile, the water depth in the paddy field changes according to the growth stage of the rice. This technology can adjust the intensity of the output airflow by changing the power intensity of the airflow generator, thereby achieving the effect of adaptive adjustment of airflow intensity according to water depth. This avoids the problem of excessive airflow intensity in shallow water or insufficient airflow intensity in deep water, thus ensuring that the paddy field water is always in a suitable turbidity state.

[0173] As can be seen from the above embodiments, the micro-automatic weeding device and its control method for paddy fields of the present invention have significant technical advantages and practical application value: the device achieves automated weeding in the underwater area of ​​paddy fields through an intelligent control system and air jet technology, significantly improving operational efficiency; it eliminates the need for chemical agents, avoiding environmental pollution and ecological imbalance; it achieves energy self-sufficiency through solar panels 10, reducing equipment operating costs; and the buoyancy adjustment device 4 and intelligent control system ensure stable operation of the device under different water depths and conditions. In summary, the present invention provides a highly efficient, environmentally friendly, and adaptable micro-automatic weeding device and its control method for paddy fields, offering an optimized solution for weed control in paddy fields.

[0174] The above embodiments are merely one of the preferred embodiments of the present invention and should not be used to limit the scope of protection of the present invention. Any modifications or refinements made to the main design concept and spirit of the present invention that are not of substantial significance, but solve the same technical problem as the present invention, should be included within the scope of protection of the present invention.

Claims

1. A control method for a micro-automatic weeding device for paddy fields, characterized in that, Includes the following steps: S1, the control system analyzes the formation law of vortex effect in water based on the water flow characteristics, and dynamically adjusts the equipment operation status by monitoring water quality parameters in real time through a water quality analyzer; the analysis process of vortex effect in water is as follows: S11, Calculate the airflow velocity distribution at the nozzle outlet of the guide pipe, and determine the intensity index V of the vortex effect in the water body based on the airflow velocity distribution. t , where V t The mathematical expression is: in, Let be the initial airflow velocity at the nozzle outlet of the guide tube, d be the distance between the nozzle and the water surface, and k1 and k2 be the airflow attenuation coefficient and diffusion coefficient, respectively. S12, through the strength index V t The vortex intensity distribution function F(V) is obtained. t Its expression is: The process of dynamically adjusting the equipment's operating status is as follows: Dissolved oxygen concentration (DO), turbidity (T), and conductivity (EC) collected by a water quality analyzer are used as input features. These input features are normalized, and an adaptive fuzzy clustering algorithm is used to calculate the water quality characteristic change label Q. w The water quality parameter change curve C of the dynamic adjustment model is calculated based on the water quality characteristic change labels. w (t): C w (t)=Q w ·(DO(t)+T(t)+EC(t)); S2, based on the vortex intensity distribution of the integrated water vortex effect and the obtained water quality parameter variation curves, determine the key parameter set for equipment operation; S3, using the aforementioned key parameter set as input variables, a device operation trajectory planning scheme is generated based on an improved particle swarm optimization algorithm; wherein, the generation process of the operation trajectory planning scheme is as follows: S31, Preprocess the key input parameters: complete the standardization of the key parameter set and the correlation analysis of trajectory planning; S32 improves upon the traditional particle swarm optimization algorithm to address the dynamic nature of the underwater environment in paddy fields. S33, establish the objective function of the running trajectory, and establish the constraint objective of minimizing the comprehensive weeding efficiency and energy consumption; S34, under constraints, iteratively solves the objective function and finally outputs the trajectory planning scheme; S4. Based on the location information and operational requirements in the equipment operation trajectory planning scheme, identify the dense distribution points of weeds in the underwater area of ​​the paddy field; S5. Establish a data-guided, mechanism-driven weeding effect evaluation model. Input the dense weed distribution points into the weeding effect evaluation model to calculate the optimal power output for equipment operation. The specific process is as follows: Obtain the area S and water depth H of the densely distributed weeds, and combine them with the airflow vortex intensity distribution function F(V) t Establish the objective function E of the weed control effect evaluation model. eff ; E eff =S·H·F(V t ) The optimal power output P for device operation is calculated based on the objective function. opt : P opt =E eff / (η·Δt) Where η is the energy conversion efficiency of the device; when the optimal power output exceeds the rated power range of the device, the power output is corrected, and the correction process is as follows: P corr =P opt ·(1-α) Where α is the power correction coefficient, and its value ranges from [0, 0.5]. S6, by adjusting the jet intensity of the airflow generator through the optimal power output, the automatic weeding task in the underwater area of ​​the paddy field is completed.

2. The control method for a micro-automatic weeding device for paddy fields according to claim 1, characterized in that, In step S2, the set of key parameters is as follows: in, This represents the maximum value of the vortex intensity. Δt represents the average value of the water quality parameter variation curve, and Δt represents the time interval of equipment operation.

3. The control method for a micro-automatic weeding device for paddy fields according to claim 2, characterized in that, In step S4, the method for identifying densely distributed weed locations is as follows: Calculate the gradient of water quality parameter changes along the trajectory of the computing device. C w Based on the changing gradient, the criteria for determining dense weed distribution points are constructed as follows: C w >θ Wherein, θ is a preset threshold; the area that meets the determination condition is defined as a densely distributed weed point.

Images