Self-driving system and method of agricultural plant protection unmanned aerial vehicle

Abstract

The invention relates to the technical field of unmanned aerial vehicle self-driving systems, and provides a self-driving system and method for an agricultural plant protection unmanned aerial vehicle, and the system comprises an environment sensing terminal, a path planning terminal, a flight control terminal, a spraying execution terminal, a task management terminal, a communication coordination terminal, and an operation evaluation terminal. The spraying execution terminal is used for adjusting the nozzle opening mode, the spraying width and the pumping pressure output according to the flight speed, the flight path height and the crop distribution characteristics in the flight state data, variable spraying of different areas is achieved, and a spraying record packet is generated to be used by the operation evaluation terminal; and the operation evaluation terminal is used for fusing the spraying record packet, the flight path data and the crop structure information, calculating the spraying coverage rate, the repeated spraying rate and the spraying leakage area distribution, and obtaining an operation evaluation result. The invention has the effect of improving the continuity of plant protection operation and the stability of plant protection operation in a complex farmland environment.

Description

Technical Field

[0001] This invention relates to the technical field of unmanned aerial vehicle (UAV) autonomous driving systems, specifically to an autonomous driving system and method for an agricultural plant protection UAV. Background Technology

[0002] With the advancement of precision agriculture and smart farm construction, agricultural drones are increasingly being used in pesticide spraying and foliar fertilizer application. Currently, most agricultural drones operate using semi-automatic flight or simple route replication, and the quality, uniformity of spraying, and efficiency largely depend on the pilot's experience. In complex farmland environments, factors such as irregular plot shapes, uneven crop growth, and frequent wind disturbances make it difficult for traditional agricultural drones to achieve highly consistent spraying and precise operational assessment while ensuring safety.

[0003] In existing technologies, some solutions optimize path planning and formation coordination for drone-based plant protection operations. For example, CN106997209B discloses a method and system for spraying plant protection drones. By acquiring GPS boundary information of the target operation area, it determines the spraying range and formation of multiple drones, generating operation paths for master and slave drones to improve operational efficiency. CN113371200B proposes that a main control system coordinates drones and autonomous vehicles to achieve integrated linkage of pesticide application and charging, thereby improving continuous operation capabilities. These solutions focus more on path planning and supply support based on plot boundaries and drone configuration, but generally treat the operation environment as static, lacking fine-grained modeling of factors such as crop structure changes and local wind disturbances; the generated flight paths are mostly pre-planned, lacking a closed-loop autonomous driving mechanism combined with real-time operation quality feedback, and have weak capabilities in identifying and correcting spray uniformity and missed areas. CN108303148B proposes a precision spraying system for agricultural drones. By installing a weighing sensor module and flow meter on the pesticide tank, combined with a spraying control system, the system monitors the quality and flow rate of the pesticide solution to improve application accuracy. While existing solutions offer some precision in adjusting spray volume and width, they often rely on fixed operating parameters and lack joint modeling with the drone's real-time flight status, perturbation acceleration, and local wind field characteristics. Furthermore, nozzle-level adjustments typically only consider altitude, speed, or simple environmental variables, failing to dynamically predict and compensate for the spatial deviation of the spray trajectory in complex wind fields.

[0004] Therefore, it is necessary to provide an autonomous driving system for agricultural plant protection drones, which enables collaborative work of terminals such as environmental perception, path planning, flight control, spraying execution, task management, communication coordination, and operation evaluation within the same airborne platform. In particular, it is necessary to introduce a correction mechanism to achieve more uniform spraying and better coverage consistency in complex farmland and wind conditions, overcoming the shortcomings of existing technologies such as dispersion, fragmentation, and lack of fine-grained compensation. Summary of the Invention

[0005] The purpose of this invention is to address the aforementioned shortcomings by proposing an autonomous driving system and method for agricultural plant protection drones.

[0006] The present invention adopts the following technical solution:

[0007] An autonomous driving system for agricultural plant protection drones includes an environmental perception terminal, a path planning terminal, a flight control terminal, a spraying execution terminal, a task management terminal, a communication coordination terminal, and an operation evaluation terminal.

[0008] The environmental sensing terminal is used to acquire multi-source environmental information of the farmland operation area and preprocess various growth parameters of crops to form an operation environment parameter package.

[0009] The path planning terminal is used to comprehensively analyze the farmland boundary shape, safety height constraints and spray uniformity requirements based on the operation environment parameter package and the operation strategy issued by the task management terminal, generate a path planning instruction set containing a track point sequence and obstacle avoidance path, and send it to the flight control terminal.

[0010] The flight control terminal is used to execute autonomous flight control according to the flight path planning instruction set, and to perform attitude correction, wind resistance compensation and speed adjustment based on environmental change information, while feeding back flight status data to the spraying execution terminal and the task management terminal.

[0011] The spraying execution terminal is used to adjust the nozzle opening mode, spray width and pump pressure output according to the flight speed, flight path altitude and crop distribution characteristics in the flight status data, so as to realize variable spraying of different areas and generate a spraying record package for the operation evaluation terminal to use.

[0012] The task management terminal is used to configure strategies, switch stages and divide tasks for the overall operation process, and generate return or respray instructions based on the remaining battery data and operation progress in the flight status data, while coordinating the working rhythm of the path planning terminal and the spraying execution terminal.

[0013] The communication coordination terminal is used to realize bidirectional communication and timing synchronization between terminals, and to complete the broadcasting of job information and conflict avoidance interaction in multi-machine collaborative scenarios.

[0014] The operation evaluation terminal is used to integrate the spraying record package, flight trajectory data and crop structure information to calculate the spraying coverage, repeated spraying rate and distribution of missed spraying areas, and obtain the operation evaluation results.

[0015] Optionally, the environmental sensing terminal includes an environmental data acquisition module, a crop structure analysis module, and a parameter package generation module; the environmental data acquisition module is used to acquire multi-source environmental information of the farmland operation area; the crop structure analysis module is used to perform structured analysis of crop density, inter-plant spacing ratio, and canopy hierarchy based on the multi-source environmental information to remove noise points and extract effective crop parameters; the parameter package generation module is used to fuse and encode the analyzed crop parameters with the environmental information to form an operation environment parameter package.

[0016] Optionally, the path planning terminal includes a boundary recognition module, a trajectory generation module, and an obstacle avoidance analysis module. The boundary recognition module is used to identify effective farmland operation areas based on the plot outline information in the operation environment parameter package, and to perform regularization processing on irregularly shaped plots. The trajectory generation module is used to comprehensively analyze the spray overlap rate, safe height, and spray width distribution according to the operation strategy issued by the task management terminal, and generate a sequence of trajectory points that meet the farmland shape constraints. The obstacle avoidance analysis module is used to perform obstacle avoidance correction and real-time adjustment of the trajectory using obstacle distribution and wind field disturbance information, and to send the corrected trajectory planning instruction set to the flight control terminal.

[0017] Optionally, the flight control terminal includes an attitude control module, a wind resistance compensation module, and a flight status feedback module. The attitude control module is used to autonomously adjust the pitch, roll, and yaw of the UAV according to the trajectory planning instruction set to maintain trajectory tracking accuracy. The wind resistance compensation module is used to predict crosswind offset in real time based on environmental change information and execute speed gain or angle gain compensation strategies to improve flight stability. The flight status feedback module is used to feed back flight status data to the spraying execution terminal and the task management terminal to support spraying compensation and task scheduling.

[0018] Optionally, the spraying execution terminal includes a spray volume control module, a spray width adjustment module, and an adaptive spray width disturbance compensation module. The spray volume control module is used to adjust the pump pressure output and nozzle opening mode according to the flight speed, flight path altitude, and crop distribution characteristics to achieve regional variable spraying. The spray width adjustment module is used to adjust the spray width according to different vegetation densities so that the spray coverage range adapts to the actual canopy structure. The adaptive spray width disturbance compensation module is used to perform real-time trend back-calculation of the discrete trajectory of the spray based on the micro-perturbation acceleration changes in the flight status data, predict the spatial offset of the spray particles under wind conditions, and perform spray width offset correction at the nozzle level to achieve high consistency coverage of the spray under wind disturbance.

[0019] Optionally, the task management terminal includes a task allocation module, a process scheduling module, and a return-to-base decision module. The task allocation module is used to decouple the overall operation objective, divide the operation area into several flight segments, and assign priorities. The process scheduling module is used to coordinate the execution rhythm of the path planning terminal and the spraying execution terminal, realizing the phase switching between trajectory planning, spraying adjustment, and flight feedback. The return-to-base decision module is used to determine whether to trigger the return-to-base or respraying process based on the remaining battery information in the flight status data and the current operation progress, and generate corresponding instructions.

[0020] A self-pilot method for an agricultural plant protection drone, applied to the self-pilot system of an agricultural plant protection drone as described above, the self-pilot method for the agricultural plant protection drone includes:

[0021] S1: Acquire multi-source environmental information of the farmland operation area and preprocess various growth parameters of crops to form an operation environment parameter package;

[0022] S2, based on the operational environment parameter package and the operational strategy issued by the task management terminal, comprehensively analyze the farmland boundary shape, safety height constraints and spray uniformity requirements, generate a flight path planning instruction set containing a flight path sequence and obstacle avoidance path, and send it to the flight control terminal;

[0023] S3, execute autonomous flight control according to the flight path planning instruction set, and perform attitude correction, wind resistance compensation and speed adjustment based on environmental change information, while feeding back flight status data to the spraying execution terminal and the task management terminal.

[0024] S4. Based on the flight speed, flight path altitude and crop distribution characteristics in the flight status data, adjust the nozzle opening mode, spray width and pump pressure output to achieve variable spraying of different areas and generate a spraying record package for use by the operation evaluation terminal.

[0025] S5, configures strategies, switches stages and divides tasks for the overall operation process, and generates return or respray instructions based on the remaining battery data and operation progress in the flight status data, while coordinating the working rhythm of the path planning terminal and the spraying execution terminal.

[0026] S6 enables bidirectional communication and timing synchronization between terminals, and completes job information broadcasting and conflict avoidance interaction in multi-machine collaborative scenarios.

[0027] S7. By integrating the spraying record package, flight trajectory data and crop structure information, the spraying coverage rate, repeated spraying rate and distribution of missed spraying areas are calculated to obtain the operation evaluation results.

[0028] The beneficial effects achieved by this invention are:

[0029] 1. By combining environmental perception terminals, path planning terminals, flight control terminals, spraying execution terminals, task management terminals, communication coordination terminals, and operation evaluation terminals, the system can complete the entire chain of processing, including environmental perception, trajectory planning, attitude control, spraying adjustment, task scheduling, and operation quality evaluation, within the same platform. This helps avoid the data fragmentation and inconsistent control rhythm caused by distributed subsystems in traditional technologies, thereby enhancing the continuity of operations and the real-time performance of parameter responses. This facilitates stable, autonomous, and highly consistent plant protection operations in complex farmland environments.

[0030] 2. By setting up the environmental data acquisition module, crop structure analysis module, and parameter package generation module in the environmental sensing terminal, the system can perform multi-source fusion and structured processing of crop density, canopy structure, inter-plant gaps, and environmental information. This is beneficial for providing more refined scene cognition results than traditional single-sensor solutions, thereby improving the adaptability to heterogeneous farmland and the accuracy of path planning and spraying adjustment. This helps to overcome the problem of operational strategy distortion caused by rough environmental modeling in existing technologies.

[0031] 3. By setting the boundary recognition module, trajectory generation module, and obstacle avoidance analysis module in the path planning terminal, the system can perform regular processing of irregularly shaped plots and perform dynamic trajectory generation and obstacle avoidance correction in combination with spray width requirements, operation strategies, and wind field disturbances. This helps to avoid the shortcomings of traditional preset routes in dealing with complex boundaries and local wind disturbances, thereby improving trajectory feasibility, spray overlap rate control capability, and path safety, which is conducive to achieving precise trajectory planning and improved spray consistency.

[0032] 4. By setting the attitude control module, wind resistance compensation module, and flight status feedback module in the flight control terminal, the system can perform dynamic attitude correction and wind resistance compensation under wind disturbance, speed fluctuation, and changes in the acceleration of the aircraft's micro-disturbance. This is beneficial to significantly improve the stability of trajectory tracking and the accuracy of speed control, and thus provide more reliable flight status parameters to the spraying execution and mission management terminal in real time. This helps to overcome the problem of spraying drift and mission rhythm disorder caused by the lag in the response of traditional flight control to disturbances.

[0033] 5. By setting the spray volume control module, spray width adjustment module, and adaptive spray width disturbance compensation module in the spray execution terminal, the system can perform spray width adjustment and dynamic compensation for spray offset according to flight speed, flight path altitude, crop structure, and micro-disturbance acceleration characteristics. This helps to solve the problem that traditional spraying systems rely solely on static parameters and cannot cope with the discrete trajectory offset of spray under wind disturbances. This improves spray consistency and effective coverage, thereby facilitating fine-grained spray width control and high-quality variable spraying, overcoming the limitations of existing technologies that lack spatial offset compensation.

[0034] 6. By setting up the task allocation module, process scheduling module, and return-to-base decision module in the task management terminal, the system can execute task segmentation, phase coordination, and return-to-base / re-spraying decisions in real time based on the operation progress, remaining power, and spraying status. This helps overcome the problems of low efficiency and easy omissions in traditional manual judgment of return-to-base and operation phase switching, thereby realizing closed-loop scheduling and stable execution of the task process, which helps improve the overall machine operation efficiency and ensure spraying integrity and operation continuity.

[0035] To further understand the features and technical content of the present invention, please refer to the following detailed description and drawings of the present invention. However, the drawings provided are for reference and illustration only and are not intended to limit the present invention. Attached Figure Description

[0036] Figure 1 This is a block diagram showing the overall structure and connections of the present invention;

[0037] Figure 2 This is a schematic diagram of the autonomous driving system of an agricultural plant protection drone according to the present invention;

[0038] Figure 3 This is a schematic diagram of the internal structure of the drone equipped with the autonomous driving system in this invention;

[0039] Figure 4 This is a schematic diagram of the method flow for an agricultural plant protection drone's self-driving method according to the present invention;

[0040] Figure 5 This is a schematic diagram illustrating the relationship between perturbation acceleration and predicted jet space offset in another embodiment of the present invention;

[0041] Figure 6 This is a matrix heatmap of the liquid spraying space offset under different disturbance conditions in another embodiment of the present invention;

[0042] Explanation of reference numerals in the attached figures:

[0043] 1. Environmental perception terminal; 2. Path planning terminal; 3. Flight control terminal; 4. Spraying execution terminal; 5. Task management terminal; 6. Communication coordination terminal; 7. Operation evaluation terminal; 8. Agricultural plant protection drone body; 81. Spraying components. Detailed Implementation

[0044] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can understand the advantages and effects of the present invention from the content disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and various details in this specification can also be modified and changed based on different viewpoints and applications without departing from the spirit of the present invention. Furthermore, the accompanying drawings of the present invention are for simple illustrative purposes only and are not depictions of actual dimensions; this is stated in advance. The following embodiments will further describe the relevant technical content of the present invention in detail, but the disclosed content is not intended to limit the scope of protection of the present invention.

[0045] Example 1: Combining Figure 1 As shown, this embodiment provides an autonomous driving system for an agricultural plant protection drone. An autonomous driving system for an agricultural plant protection drone, combined with... Figure 2 and Figure 3 As shown, the system includes an environmental perception terminal 1, a path planning terminal 2, a flight control terminal 3, a spraying execution terminal 4, a task management terminal 5, a communication coordination terminal 6, and a job evaluation terminal 7. In this embodiment, the environmental perception terminal 1, path planning terminal 2, flight control terminal 3, spraying execution terminal 4, task management terminal 5, communication coordination terminal 6, and job evaluation terminal 7 are all mounted inside the agricultural plant protection drone body 8. The agricultural plant protection drone body 8 is equipped with a spraying component 81, which is used to achieve adaptive spraying operations based on the control of the spraying execution terminal 4.

[0046] The environmental sensing terminal 1 is used to acquire multi-source environmental information of the farmland operation area and preprocess various growth parameters of crops to form an operation environment parameter package.

[0047] The path planning terminal 2 is used to comprehensively analyze the farmland boundary shape, safety height constraints and spray uniformity requirements based on the operation environment parameter package and the operation strategy issued by the task management terminal 5, generate a path planning instruction set containing a track point sequence and obstacle avoidance path, and send it to the flight control terminal 3.

[0048] The flight control terminal 3 is used to execute autonomous flight control according to the flight path planning instruction set, and to perform attitude correction, wind resistance compensation and speed adjustment based on environmental change information, while feeding back flight status data to the spraying execution terminal 4 and the task management terminal 5.

[0049] The spraying execution terminal 4 is used to adjust the nozzle opening mode, spray width and pump pressure output according to the flight speed, flight path altitude and crop distribution characteristics in the flight status data, so as to realize variable spraying of different areas and generate a spraying record package for use by the operation evaluation terminal 7.

[0050] The task management terminal 5 is used to configure strategies, switch stages and divide tasks for the overall operation process, and generate return or respray instructions based on the remaining battery data and operation progress in the flight status data. At the same time, it coordinates the working rhythm of the path planning terminal 2 and the spraying execution terminal 4.

[0051] The communication coordination terminal 6 is used to realize bidirectional communication and timing synchronization between terminals, and to complete the broadcasting of job information and conflict avoidance interaction in multi-machine collaborative scenarios.

[0052] The operation evaluation terminal 7 is used to integrate the spraying record package, flight trajectory data and crop structure information to calculate the spraying coverage rate, repeated spraying rate and the distribution of missed spraying areas, and obtain the operation evaluation results.

[0053] Optionally, the environmental sensing terminal 1 includes an environmental data acquisition module, a crop structure analysis module, and a parameter package generation module; the environmental data acquisition module is used to acquire multi-source environmental information of the farmland operation area; the crop structure analysis module is used to perform structured analysis of crop density, inter-plant spacing ratio, and canopy hierarchy based on the multi-source environmental information to remove noise points and extract effective crop parameters; the parameter package generation module is used to fuse and encode the analyzed crop parameters with the environmental information to form an operation environment parameter package.

[0054] Optionally, the path planning terminal 2 includes a boundary recognition module, a trajectory generation module, and an obstacle avoidance analysis module. The boundary recognition module is used to identify effective farmland operation areas based on the land contour information in the operation environment parameter package, and to perform regularization processing on irregularly shaped land parcels. The trajectory generation module is used to perform comprehensive analysis on the spraying overlap rate, safe height, and spray width distribution according to the operation strategy issued by the task management terminal 5, and to generate a sequence of trajectory points that meet the farmland shape constraints. The obstacle avoidance analysis module is used to perform obstacle avoidance correction and real-time adjustment of the trajectory using obstacle distribution and wind field disturbance information, and to send the corrected trajectory planning instruction set to the flight control terminal 3.

[0055] Optionally, the flight control terminal 3 includes an attitude control module, a wind resistance compensation module, and a flight status feedback module; the attitude control module is used to autonomously adjust the pitch, roll, and yaw of the UAV according to the trajectory planning instruction set to maintain trajectory tracking accuracy; the wind resistance compensation module is used to predict crosswind offset in real time based on environmental change information and execute speed gain or angle gain compensation strategies to improve flight stability; the flight status feedback module is used to feed back flight status data to the spraying execution terminal 4 and the task management terminal 5 to support spraying compensation and task scheduling.

[0056] Optionally, the spraying execution terminal 4 includes a spray volume control module, a spray width adjustment module, and an adaptive spray width disturbance compensation module. The spray volume control module is used to adjust the pump pressure output and nozzle opening mode according to the flight speed, flight path altitude, and crop distribution characteristics to achieve regional variable spraying. The spray width adjustment module is used to adjust the spray width according to different vegetation densities so that the spray coverage range adapts to the actual canopy structure. The adaptive spray width disturbance compensation module is used to perform real-time trend back-calculation of the spray discrete trajectory based on the micro-perturbation acceleration changes in the flight status data, predict the spatial offset of spray particles under wind conditions, and perform spray width offset correction at the nozzle level to achieve high consistency coverage of the spray under wind disturbance conditions.

[0057] Optionally, the task management terminal 5 includes a task allocation module, a process scheduling module, and a return-to-base decision module. The task allocation module is used to decouple the overall operation objective, divide the operation area into several flight segments, and assign priorities. The process scheduling module is used to coordinate the execution rhythm of the path planning terminal 2 and the spraying execution terminal 4, realizing the phase switching between trajectory planning, spraying adjustment, and flight feedback. The return-to-base decision module is used to determine whether to trigger the return-to-base or respraying process based on the remaining battery information in the flight status data and the current operation progress, and generate corresponding instructions.

[0058] In an exemplary implementation, this system forms a closed-loop operational chain following the sequence of "environmental perception, path planning, flight control, spraying execution, operation evaluation, and task management and scheduling." Each terminal exchanges data through a timing synchronization mechanism provided by the communication coordination terminal. After system startup, the environmental perception terminal first collects multi-source environmental data from airborne sensors, remote image sources, and ground monitoring points, and completes crop structure analysis and parameter package generation, enabling the path planning terminal to perform trajectory analysis based on structured operational environment information.

[0059] After the task management terminal completes the configuration of the operation strategy, the path planning terminal generates a preliminary flight path based on the shape of the work area, obstacle distribution, safe altitude, and spray width constraints. The obstacle avoidance analysis module further uses wind field disturbance trends to dynamically correct the flight path, and the correction results are then sent to the flight control terminal. When performing autonomous flight, the flight control terminal calls the attitude control module and wind resistance compensation module in real time to counteract attitude deviations caused by crosswinds, updrafts, or sudden disturbances, and continuously feeds back flight status data to the spraying execution terminal and the task management terminal.

[0060] During the spraying process, the spray volume control module adjusts the pump pressure and nozzle opening mode according to the real-time flight speed and vegetation density; the spray width adjustment module realizes active width adjustment based on changes in the canopy structure; and the adaptive spray width disturbance compensation module predicts the actual offset trajectory of the sprayed particles based on the acceleration micro-disturbance and airflow disturbance trend in the flight status data, and performs spray width offset calibration at the nozzle level to ensure consistency of variable spraying under different wind field conditions.

[0061] During the operation, the task management terminal dynamically executes process scheduling and return-to-base decisions based on the remaining battery power and operation progress information in the flight status data, and forms a two-way linkage with the path planning terminal and the spraying execution terminal. For example, when the battery power is lower than the threshold, the return-to-base decision module will generate a return-to-base instruction; and when there is a missed spray area in the operation area, the task management terminal can generate a supplementary spray instruction and reschedule the path planning terminal to generate a local flight path.

[0062] The spraying execution terminal continuously generates spraying record packages. After the operation is completed, the operation evaluation terminal integrates the spraying records, flight trajectory, and crop structure parameters to calculate the spray coverage, repetition rate, and distribution of missed areas, obtaining the operation evaluation results. These evaluation results are not only used to summarize the current task but also serve as a reference for path planning optimization for the next task, enabling experience-based operation improvement.

[0063] In a further preferred embodiment, each terminal uses a unified timestamp identifier provided by the communication coordination terminal to achieve accurate mapping of cross-module data packets across different work cycles. The communication coordination terminal can allocate communication time slots in multi-drone collaborative operation scenarios to broadcast operational intent information, detect and avoid flight path conflicts, enabling multiple drones to maintain safe distances and operational efficiency within the same area.

[0064] To enhance system robustness, in complex scenarios such as drastic wind speed changes, sudden shifts in ambient light, or temporary sensor obstruction, the environmental perception terminal can trigger a degradation processing strategy. This strategy uses the previous moment's operational environment parameter package to fill in missing data areas, ensuring the availability of the trajectory planning and spraying execution terminals even with incomplete input. Simultaneously, in the event of sudden disturbances, the flight control terminal and spraying execution terminal can enter a buffer mode based on scheduling instructions from the mission management terminal, temporarily suspending non-critical actions to ensure flight safety.

[0065] In another optional implementation, the system can be remotely monitored via a ground station or mobile monitoring equipment. The task management terminal can upload operation progress, spraying statistics, and path tracking errors to the ground station, enabling operators to monitor the operation results in real time. Simultaneously, the operation evaluation terminal can export spraying heat maps, coverage statistics, and repeat spraying analysis charts after the task is completed, providing visualized data support for farmland management.

[0066] Through the collaboration of the above-mentioned processing links and modules, the agricultural plant protection drone autonomous driving system of this embodiment can not only achieve autonomous operation in complex farmland environments, but also perform variable spraying based on crop structure differences, maintain spray width consistency under wind disturbance conditions, and realize automatic scheduling of operation processes through task management terminals, thereby significantly improving the efficiency, uniformity and resource utilization of farmland spraying operations.

[0067] An autonomous driving method for agricultural plant protection drones, applied to the autonomous driving system of an agricultural plant protection drone as described above, combined with... Figure 4 As shown, the self-driving method of the agricultural plant protection drone includes:

[0068] S1: Acquire multi-source environmental information of the farmland operation area and preprocess various growth parameters of crops to form an operation environment parameter package;

[0069] S2, based on the operational environment parameter package and the operational strategy issued by the task management terminal, comprehensively analyze the farmland boundary shape, safety height constraints and spray uniformity requirements, generate a flight path planning instruction set containing a flight path sequence and obstacle avoidance path, and send it to the flight control terminal;

[0070] S3, execute autonomous flight control according to the flight path planning instruction set, and perform attitude correction, wind resistance compensation and speed adjustment based on environmental change information, while feeding back flight status data to the spraying execution terminal and the task management terminal.

[0071] S4. Based on the flight speed, flight path altitude and crop distribution characteristics in the flight status data, adjust the nozzle opening mode, spray width and pump pressure output to achieve variable spraying of different areas and generate a spraying record package for use by the operation evaluation terminal.

[0072] S5, configures strategies, switches stages and divides tasks for the overall operation process, and generates return or respray instructions based on the remaining battery data and operation progress in the flight status data, while coordinating the working rhythm of the path planning terminal and the spraying execution terminal.

[0073] S6 enables bidirectional communication and timing synchronization between terminals, and completes job information broadcasting and conflict avoidance interaction in multi-machine collaborative scenarios.

[0074] S7. By integrating the spraying record package, flight trajectory data and crop structure information, the spraying coverage rate, repeated spraying rate and distribution of missed spraying areas are calculated to obtain the operation evaluation results.

[0075] Example 2: Building upon Example 1, this example provides an algorithm for predicting spray spatial offset and correcting spray width bias in an adaptive spray amplitude disturbance compensation module, based on joint modeling of micro-perturbation acceleration and wind field characteristics. This algorithm utilizes the triaxial micro-perturbation acceleration changes recorded in flight state data, combined with wind field shear characteristics and asymmetric parameters of spray particle size distribution, to infer the trend of the discrete spray trajectory in real time. This allows for the determination of the spray amplitude bias at the nozzle level, ensuring that the spray coverage profile after wind disturbance approximates the ideal coverage profile in windless conditions. Combined with... Figure 5 As shown, Figure 5 This study demonstrates the trend of lateral spray offset as a function of perturbation acceleration, used to verify the dominant role of perturbation acceleration in offset prediction within the adaptive spray amplitude perturbation compensation algorithm. Combined with... Figure 6 As shown, Figure 6 The statistical trend of the lateral displacement of the sprayed liquid under the combined conditions of micro-perturbation acceleration and wind speed / wind shear intensity is displayed in two-dimensional matrix form to verify the effectiveness of the displacement prediction model based on the comprehensive perturbation characteristics.

[0076] In this embodiment, the adaptive jet amplitude disturbance compensation module uses a sliding time window. The three-axis acceleration of the aircraft at the current moment is read from the flight status data, in units of [unit]. and the corresponding benchmark steady acceleration The perturbation acceleration vector is defined as:

[0077] ;

[0078] in, It can be obtained from the sliding mean of low-frequency acceleration components or steady-state cruise acceleration over a recent period, and is used to characterize the baseline gravity and maneuver components without instantaneous disturbances. In the lateral offset prediction of the jet ejection discrete trajectory, the focus is mainly on the lateral perturbation component perpendicular to the jet amplitude direction. This component can be obtained by... The lateral axis is obtained by projecting it onto the body coordinate system.

[0079] To characterize the perturbation effect of the wind field on the trajectory of the jet particles, this embodiment introduces a local wind shear strength index. This can be obtained by interpolation from airborne wind speed estimation modules or ground weather station data, representing the gradient of wind speed with altitude near the nozzle's height. On the other hand, to reflect the influence of spray spectral width structure on offset sensitivity, this embodiment introduces a spray particle size distribution asymmetry coefficient. This coefficient, obtained during the calibration phase by statistically analyzing the droplet size distribution of the nozzle and calculating its skewness, reflects the degree of bias in droplet size towards either "smaller droplets" or "larger droplets." Studies have found that under the same wind shear conditions, The larger the spray pattern, the more uneven the response of the spray profile to disturbances, and the more prone the edge areas are to "voids" or excessive overlap. Therefore, The modulation factor used to correct the intensity of the spray amplitude bias.

[0080] In the window Internally, structural disturbance comprehensive characteristic quantity for:

[0081] ;

[0082] in, This is the scale factor for the effect of wind shear on acceleration disturbances. This is the asymmetry factor for the spray particle size. This reflects the strength of the overall migration trend of the jet particle swarm under the current conditions of perturbation acceleration and wind shear. When When the spray spectral width is close to zero, the effect of the shift is relatively symmetrical. It is mainly determined by acceleration and wind shear; while when Even when taking a larger value, Even smaller droplets at the edges can be significantly deflected by the wind field. It will still be amplified nonlinearly.

[0083] Based on the comprehensive characteristics of the disturbance This embodiment presents a predictive model for the lateral spatial offset of the spray. It assumes the effective drop height from the nozzle to the crop canopy is... The horizontal combined velocity of the body relative to the airflow is The displacement of a droplet caused by lateral disturbance per unit time can be expressed by a nonlinear response function as:

[0084] ;

[0085] in, This is the disturbance response sensitivity coefficient. To prevent the use of tiny positive numbers with a denominator of zero, in this formula, This is used to compress the tendency of the offset to increase sharply with the intensity of the disturbance to a limited range, so that the predicted value does not grow indefinitely under extreme wind field conditions, while retaining high sensitivity to the offset change under moderate disturbance conditions. The sign indicates the direction of the offset of the spray pattern relative to the planned coverage centerline, and the absolute value indicates the expected offset magnitude.

[0086] In order to map the spray offset to the nozzle spray width offset command, the adaptive spray width disturbance compensation module in this embodiment performs each window... Internal calculation of the spray width offset angle of the nozzle level Or spray width offset Taking spray angle compensation as an example, the following relationship can be used:

[0087] ;

[0088] in, For symbolic functions, To prevent division by zero of tiny positive numbers, the formula maps the predicted lateral displacement to an angular offset in the spray pattern, ensuring the jet falls back near the target crop rows as it descends in the wind. For structures using multi-nozzle arrays, the adaptive spray pattern disturbance compensation module can calculate separately for each nozzle. Alternatively, differential offsets can be applied at both ends of the array to keep the centroid of the overall spray pattern on the planned trajectory.

[0089] In practical implementation, the above algorithm uses a periodic... Sliding evaluation window Each window is updated based on the latest flight status data and wind field estimates. , and The updated bias command is then sent to the nozzle control drive unit via the spray execution terminal. To prevent frequent fluctuations in the spray width command due to instantaneous measurement noise, the adaptive spray width disturbance compensation module can adjust the value within a continuous window. Smoothing filtering is performed, and the current instruction remains unchanged when the magnitude of the bias change is below the threshold. New bias adjustment is only performed when the disturbance trend continues and exceeds the threshold.

[0090] By introducing the asymmetry coefficient of spray particle size distribution This unexpected parameter, coupled nonlinearly with micro-perturbation acceleration and wind shear intensity, enables the adaptive spray amplitude disturbance compensation algorithm in this embodiment to distinguish between "offset dominated by geometric attitude disturbance" and "offset caused by droplet spectral width structure sensitivity." Under the same flight trajectory, it performs targeted spray amplitude offset correction for different nozzles and different crop sections, thereby maintaining a highly consistent spray coverage effect in farmland environments with significant wind field disturbances and crop structure differences.

[0091] To clarify the meaning of each parameter in the above jet amplitude disturbance compensation algorithm, this embodiment further supplements the explanation of some parameters. Effective drop height. This is the real-time vertical distance from the nozzle exit position to the crop canopy. This parameter dynamically changes with the UAV's altitude control strategy and can be jointly corrected by the onboard altimeter and visual estimation module. Relative airflow velocity. This is the resultant velocity between the forward velocity of the UAV and the velocity components of the airflow after the wind field is decomposed within the system. Its physical meaning lies in characterizing the "effective motion medium velocity" of the jet particles falling in the air; under headwind conditions... The reduction makes the spray more susceptible to lateral disturbances, especially under tailwind conditions. If the volume increases, the sensitivity of the sprayed liquid to the same level of crosswind field decreases.

[0092] Disturbance response sensitivity coefficient To adjust hyperparameters, its function is to control Nonlinear perturbation synthesis characteristics The amplification speed is relatively large. It will make It produces a significant response to even small perturbations, while smaller ones... This makes the model more conservative, suitable for scenarios with large nozzle spray volume and high coverage stability. Parameters and For small positive numbers, generally take The magnitude is intended to avoid abnormally amplified biased predicted values ​​when the denominator approaches zero, while maintaining the continuity of the model within the normal range.

[0093] In addition, wind shear scale factor Amplification factor for particle size asymmetry It can be obtained through experimental calibration, among which The weight of wind shear in the disturbance characteristics is determined, and This is used to demonstrate the amplifying effect of the asymmetry in spray particle size distribution on the offset of the spray edge region.

[0094] In this embodiment, to further illustrate the lateral offset of the sprayed liquid... offset angle with spray pattern The actual correction process is illustrated below with an example calculation and correction method.

[0095] Suppose a sliding window Inside, the environmental perception terminal detected a lateral disturbance caused by a sudden gust of wind, and the lateral component of the micro-disturbance acceleration provided by the flight control terminal was... The local wind shear strength is The asymmetry coefficient of the spray particle size is Substituting the comprehensive characteristic of the disturbance:

[0096] ;

[0097] If the calibration is taken , Then we can obtain:

[0098] The disturbance amplitude based on acceleration and wind shear is approximately:

[0099] ;

[0100] The magnification factor for particle size distribution asymmetry is:

[0101] ;

[0102] Final perturbation composite characteristic:

[0103] ;

[0104] Further substituting this into the lateral offset prediction model (assuming effective fall height) relative airflow velocity , ):

[0105] ;

[0106] Pick We can obtain:

[0107] ;

[0108] ;

[0109] This result means that, under the current conditions of wind disturbance and acceleration fluctuations, the sprayed liquid is expected to drift outward by about 35 cm relative to the planned coverage centerline.

[0110] To map the offset to the spray width correction, this embodiment uses a nozzle layer angle offset method. Substituting into the compensation angle calculation formula:

[0111] ;

[0112] like ,but:

[0113] ;

[0114] This offset angle indicates that the nozzle spray pattern needs to be rotated approximately 9.4° in the opposite direction (against the wind) to counteract the tendency of the sprayed liquid to drift laterally in the wind field, so that its actual landing point is realigned with the planned spray centerline.

[0115] During the correction process, the adaptive spray amplitude disturbance compensation module sends an update command to the spray execution terminal. The nozzle servo driver will then dynamically adjust the spray amplitude angle within a closed-loop response period of approximately 50–100 ms. To avoid frequent nozzle jitter under continuous small disturbances, the module adjusts the spray amplitude angle within a continuous window. Smoothing can be performed, such as by averaging the most recent three windows or using a threshold trigger mechanism. For example, when the offset angle changes three consecutive times are all less than... When the system maintains the current spray width command, it will remain unchanged; when a sudden gust of wind causes the offset angle to exceed the set threshold... If necessary, immediately execute the correction command.

[0116] After receiving the new spray width offset command, the nozzle spray profile will gradually approach the ideal coverage bandwidth under windless conditions during subsequent spraying. Actual measurements show that this algorithm can effectively reduce spray width offset error by approximately 25%–40% in environments with wind speed fluctuations of 2–6 m / s, significantly improving the problems of missed and repeated spraying, thereby enhancing the uniformity of pesticide application.

[0117] Example 3: Based on Examples 1 and 2, this example provides a hardware platform architecture and program execution method for an agricultural plant protection drone autonomous system. This example aims to illustrate that the system of the present invention can be deployed on an onboard control platform and a ground monitoring platform of a drone with computing, sensing, execution, and communication capabilities, enabling functions such as environmental perception, path planning, flight control, spraying execution, task management, and operation evaluation to be implemented through software programs running on a processor.

[0118] The autonomous driving system in this embodiment can run on an UAV onboard hardware platform including a processor, memory, accelerometer, vision acquisition device, wind speed detection module, electronic speed controller (ESC) driver, nozzle control unit, and communication module. The processor is used to call program instructions stored in the memory to sequentially execute various functional modules of the environmental perception terminal, path planning terminal, flight control terminal, spraying execution terminal, task management terminal, communication coordination terminal, and operation evaluation terminal. The memory may include cache, solid-state storage, or on-chip SRAM, used to store data structures such as operational environment parameter packages, trajectory planning instruction sets, spraying record packages, flight status data, and spray width offset calculation results. The accelerometer and wind speed detection module are used to generate the perturbation acceleration and wind field information required for the algorithm input in Embodiment 2; the vision acquisition device is used to provide the data required for crop structure analysis and plot boundary identification; the ESC and nozzle control unit are used to implement flight attitude control and spray width correction, respectively.

[0119] In this embodiment, the UAV onboard control platform can run a real-time operating system (RTOS) to schedule the execution cycles of each functional terminal. For example, the environmental perception task can run at a cycle of 50–100 ms to ensure the real-time performance of sensor data; the spray width offset correction task can run at a sliding window cycle given in Embodiment 2; the flight control task can perform pitch, roll, and yaw attitude adjustments within a high-speed cycle of 10–20 ms; and the spraying execution and path planning tasks can use a longer cycle of 100–200 ms to perform strategic decision-making. Under the scheduling rules of the task management terminal, each task exchanges data through a shared data area or DMA channel. The unified timestamp mechanism provided by the communication coordination terminal is used to ensure data synchronization between different terminals and reduce spraying control deviations caused by cross-module delays.

[0120] In another implementation, the system can be deployed partially or entirely on a ground station or remote server. For example, the operation assessment terminal and some path planning functions can run on the ground station to utilize stronger computing power to calculate spray coverage, analyze areas of missed or excessive spraying, and optimize flight paths for complex farmland geometries. The UAV and the ground station can exchange data via a wireless link. The ground station is responsible for storing spray record packets, flight trajectories, and crop structure information, generating operation reports, coverage heat maps, and other results, and then distributing the spray compensation strategy to the UAV for continued execution.

[0121] To further illustrate the software configuration of the system of the present invention, this embodiment provides a computer program product for implementing the above-mentioned functions. This program product is stored in a non-transitory computer-readable storage medium, such as flash memory, solid-state drive, EEPROM, or cloud storage. When the processor invokes the program instructions, it sequentially executes the environmental perception processing flow, the trajectory generation flow, the attitude control flow, the jet amplitude disturbance compensation flow, the task scheduling flow, and the job evaluation flow. The program may include code segments for executing the jet amplitude disturbance compensation algorithm of Embodiment Two, responsible for calculating the micro-disturbance acceleration, wind shear, particle size asymmetry parameters, and nonlinear offset prediction results, and thereby generating a jet amplitude offset correction command that is transmitted to the nozzle control unit.

[0122] In a further embodiment, the above program can adopt a modular structure to adapt to agricultural equipment with different drone models or different numbers of nozzles through configuration. For example, when there are many nozzles, array bias control logic can be enabled in the program to achieve differential correction for multiple nozzles; when wind disturbances are strong, it can automatically switch to a high-sensitivity mode to enhance the weight of disturbance characteristics; when GPS signals are weak or visual conditions are poor, trajectory control stability can be maintained through redundant inertial navigation and altimeter information. In multi-drone collaborative scenarios, the program can also call the broadcast mechanism of the communication coordination module to exchange information on operation boundaries, spraying progress, and trajectory conflict to achieve regional collaborative spraying.

[0123] Through the above description of the hardware platform structure and program execution method, this embodiment demonstrates that the agricultural plant protection drone autonomous system of the present invention can operate stably in a standard airborne hardware platform and ground station environment, and can realize core functions such as environmental perception, path planning, flight control and spraying compensation through software modules, making the technical solutions described in Embodiment 1 and Embodiment 2 industrially feasible and deployable.

[0124] The content disclosed above is only a preferred and feasible embodiment of the present invention, and is not intended to limit the scope of protection of the present invention. Therefore, all equivalent technical changes made based on the content of the present invention specification and drawings are included within the scope of protection of the present invention. Furthermore, the elements therein can be updated as technology develops.

Claims

1. An autopilot system of an agricultural plant protection unmanned aerial vehicle, characterized in that, The system comprises an environment sensing terminal, a path planning terminal, a flight control terminal, a spraying execution terminal, a task management terminal, a communication coordination terminal and a work evaluation terminal. The environment sensing terminal is configured to acquire multi-source environment information of a farmland work area, and to preprocess various growth parameters of crops to form a work environment parameter package. The path planning terminal is configured to perform comprehensive analysis on farmland boundary shape, safety height constraint and spraying uniformity requirement based on the work environment parameter package and a work strategy issued by the task management terminal, to generate a flight path planning instruction set containing a flight path point sequence and an obstacle avoidance path, and to send the flight path planning instruction set to the flight control terminal. The flight control terminal is configured to perform autonomous flight control according to the flight path planning instruction set, to perform attitude correction, wind resistance compensation and speed adjustment based on environment change information, and to feed back flight state data to the spraying execution terminal and the task management terminal. The spraying execution terminal is configured to adjust nozzle opening mode, spraying width and pump pressure output according to flight speed, flight path height and crop distribution characteristics in the flight state data, to realize variable spraying in different areas, and to generate a spraying record package for the work evaluation terminal. The task management terminal is configured to perform strategy configuration, stage switching and task division on an overall work process, to generate a return instruction or a supplementary spraying instruction according to residual power data and work progress in the flight state data, and to coordinate work rhythm of the path planning terminal and the spraying execution terminal. The communication coordination terminal is configured to realize bidirectional communication and time sequence synchronization among the terminals, and to complete work information broadcasting and conflict avoidance interaction in a multi-machine cooperation scenario. The work evaluation terminal is configured to fuse the spraying record package, flight trajectory data and crop structure information, to calculate spraying coverage, repeated spraying rate and missed spraying area distribution, and to obtain a work evaluation result. 2.The self-driving system of an agricultural plant protection unmanned aerial vehicle of claim 1, wherein The environment sensing terminal comprises an environment data acquisition module, a crop structure analysis module and a parameter package generation module. The environment data acquisition module is configured to acquire multi-source environment information of a farmland work area. The crop structure analysis module is configured to perform structural analysis on crop density, inter-plant spacing ratio and canopy level based on the multi-source environment information, to eliminate noise points and extract effective crop parameters. The parameter package generation module is configured to fuse and encode the analyzed crop parameters and environment information to form a work environment parameter package. 3.The self-driving system of an agricultural plant protection unmanned aerial vehicle according to claim 2, wherein, The path planning terminal comprises a boundary identification module, a flight path generation module and an obstacle avoidance analysis module. The boundary identification module is configured to identify an effective farmland work area based on land contour information in the work environment parameter package, and to perform regular processing on irregular land blocks. The flight path generation module is configured to perform comprehensive analysis on spraying overlap rate, safety height and spraying distribution according to a work strategy issued by the task management terminal, to generate a flight path point sequence satisfying farmland shape constraint. The obstacle avoidance analysis module is configured to perform obstacle avoidance correction and real-time adjustment on the flight path by using obstacle distribution and wind field disturbance information, and to send the corrected flight path planning instruction set to the flight control terminal. 4.The self-driving system of an agricultural plant protection unmanned aerial vehicle of claim 3, wherein, The flight control terminal comprises an attitude control module, a wind resistance compensation module and a flight state feedback module; the attitude control module is used for autonomously adjusting the pitch, roll and yaw of the unmanned aerial vehicle according to a flight path planning instruction set, so as to maintain the flight path tracking accuracy; the wind resistance compensation module is used for predicting the side wind deviation amount in real time based on the environmental change information, and performing speed gain or angle gain compensation strategy, so as to improve the flight stability; the flight state feedback module is used for feeding back the flight state data to the spraying execution terminal and the task management terminal, so as to support the spraying compensation and task scheduling. 5.The self-driving system of an agricultural plant protection unmanned aerial vehicle of claim 4, wherein, The spraying execution terminal comprises a spraying amount regulation module, a spraying width adjustment module and an adaptive spraying width disturbance compensation module; the spraying amount regulation module is used for adjusting the pump pressure output and the nozzle opening mode according to the flight speed, the flight path height and the crop distribution characteristics, so as to realize the regional variable spraying; the spraying width adjustment module is used for adjusting the spraying width according to different vegetation densities, so that the spraying coverage range is adapted to the actual canopy structure; the adaptive spraying width disturbance compensation module is used for performing real-time trend backstepping on the spraying particle discrete trajectory based on the micro-disturbance acceleration change in the flight state data, predicting the spatial deviation amount of the spraying particles under the wind field condition, and performing the spraying width bias correction at the nozzle level, so that the spraying liquid realizes the high consistency coverage under the wind disturbance. 6.The self-driving system of an agricultural plant protection unmanned aerial vehicle of claim 5, wherein, The task management terminal comprises a task allocation module, a process scheduling module and a return decision module; the task allocation module is used for decoupling the overall operation target, dividing the operation area into a plurality of flight segments and allocating priorities; the process scheduling module is used for coordinating the execution rhythm of the path planning terminal and the spraying execution terminal, so as to realize the stage switching among the flight path planning, the spraying adjustment and the flight feedback; the return decision module is used for judging whether the return or the supplementary spraying process is triggered based on the residual power information and the current operation progress in the flight state data, and generating corresponding instructions.

7. The self-driving method of the agricultural plant protection unmanned aerial vehicle, applied to the self-driving system of the agricultural plant protection unmanned aerial vehicle according to claim 6, characterized in that, The self-driving method of the agricultural plant protection unmanned aerial vehicle comprises: S1, obtaining multi-source environmental information of a farmland operation area, and preprocessing various growth parameters of crops to form an operation environment parameter package; S2, based on the operation environment parameter package and the operation strategy issued by the task management terminal, comprehensively analyzing the farmland boundary shape, the safety height constraint and the spraying uniformity requirement, generating a flight path planning instruction set containing a flight path point sequence and an obstacle avoidance path, and sending the flight path planning instruction set to the flight control terminal; S3, performing autonomous flight control according to the flight path planning instruction set, and performing attitude correction, wind resistance compensation and speed adjustment based on environmental change information, while feeding back flight state data to the spraying execution terminal and the task management terminal; S4, adjusting the nozzle opening mode, the spraying width and the pump pressure output according to the flight speed, the flight path height and the crop distribution characteristics in the flight state data, so as to realize variable spraying in different regions, and generating a spraying record package for the operation evaluation terminal. S5, the whole operation process is configured, the stage switching and task division are carried out, and the remaining power data and the operation progress in the flight state data are used to generate the return instruction or the supplementary spraying instruction, and the working rhythm of the path planning terminal and the spraying execution terminal is coordinated; S6, the two-way communication and timing synchronization between the terminals are realized, and the operation information broadcasting and conflict avoidance interaction are completed in the multi-machine cooperation scene; S7, the spraying record package, the flight trajectory data and the crop structure information are fused, the spraying coverage, the repeated spraying rate and the missed spraying area distribution are calculated, and the operation evaluation result is obtained.

Images