Research On Obstacle Avoidance And Side Spray Technology Of Plant Protection UAV

Plant protection mechanization is the key part of agricultural production.As one of the plant protection machines,the plant protection UAV have obvious advantages over manual plant protection and ground plant protection machines in terms of flexible adaptation to the operating plot and efficient spraying.The environment of the plant protection UAV operation usually distributes various kinds of obstacles that greatly threaten flight safety.Currently,the plant protection UAV usually use the single type environmental sensing sensor to detect the operating environment.And the obstacle avoidance function is implemented by single preset command such as hover alert or emergency landing.Although some plant protection UAV can avoid obstacles through bypass by pre-setting obstacle avoidance paths.However,the plant protection UAV will reserve a safe distance of 2～5 m from the obstacles during bypass.And the current research on plant protection UAV spraying all focus on the spraying without obstacle.Under the condition that the plant protection UAV uses bypass to avoid obstacles,there is no targeted spraying around the obstacles.This will result in insufficient droplet coverage around obstacles,requiring a secondary manual replenishment spraying to avoid affecting the crop growth around the obstacle.In order to improve the autonomous operation capability and plant protection operation effect of the plant protection UAV in obstacle environment,and meet the requirements of autonomous spraying,real-time obstacle avoidance and precise operation.This article conducts research from two aspects of the plant protection UAV: obstacle avoidance and increasing the droplet coverage around obstacles during obstacle avoidance bypass.For the plant protection UAV obstacle avoidance,the research mainly involves field obstacle information recognition,obstacle avoidance algorithm improvement,obstacle avoidance system and flight platform design and construction.For increasing the droplet coverage around obstacles during obstacle avoidance bypass,the study of the plant protection UAV side spray technology was mainly conducted.The detailed research contents and main results are as follows.(1)To improve the recognition accuracy of obstacles in the field of the plant protection UAV,millimeter wave radar and monocular camera were used as the environment sensing sensors of the plant protection UAV.Spatial fusion model and multithreaded time synchronization model were built to realize the fusion of both spatial and temporal data.Using trees and utility poles in the field as common obstacles,this data fusion model was used for obstacle data collection.The contour enhancement and extraction of obstacles from the data fusion model was performed by weighted average grayscale processing,Canny edge detection and mathematical morphology processing.The data fusion model detects obstacles with a maximum error of 8.2% for distance,17.3% for width and 18.5% for height.Clear and smooth contours of the obstacle can be obtained from the obstacle images.It can provide an obstacle basis for the plant protection UAV obstacle avoidance decision.(2)In this article,from the requirements of efficiency and path smoothness of the plant protection UAV obstacle avoidance,an improved A* obstacle avoidance algorithm was proposed through dynamic heuristic function,search point selection strategy optimization,and inflection point number optimization.Compared with the traditional A* obstacle avoidance algorithm,the improved A* obstacle avoidance algorithm proposed in this study for the operational environment and obstacle avoidance requirements of the plant protection UAV was more purposeful in path search.It can at least reduce planning time by 68.4%,search grids by 74.9% and inflection points by 20.7% with very little or no increase in path length.(3)Based on the millimeter wave radar and monocular camera data fusion model and the improved A* obstacle avoidance algorithm,the obstacle avoidance system and flight platform were studied and built to realize the actual obstacle avoidance of the plant protection UAV.The coordinate system conversion and attitude description were carried out by combining the structure form and parameters of the plant protection UAV,and the control model of the whole UAV was established.By simplifying the control model,an inner and outer loop control strategy was proposed.A two-layer control system based on NVIDIA Jetson TX2 airborne computer(primary controller)and Pixhawk4 flight controller(secondary controller)was designed.The plant protection UAV obstacle avoidance system and flight platform were designed and built based on the two-layer control system as the core,and actual obstacle avoidance flight experiments were conducted.The results show that the plant protection UAV obstacle avoidance flight platform constructed in this article can integrate obstacle recognition under millimeter wave radar and monocular camera data fusion,path planning by improved A* obstacle avoidance algorithm,and flight control by obstacle avoidance path following into one system.During the actual obstacle avoidance flight experiment,the minimum deviation between actual obstacle avoidance flight trajectory and path planning trajectory was 0.1 m,the maximum deviation was 1.4 m.And the minimum distance between the UAV and obstacle was 1.6 m,the maximum distance was 2.8 m.(4)In order to improve the droplet coverage around obstacles during the plant protection UAV obstacle avoidance process,this article conducts research on the side spray technology used to increase the droplet coverage around obstacles,and proposes a design method for the side spray device.The research explored the relationship between the plant protection UAV obstacle avoidance bypass path and the side spray device spraying pattern to the droplet coverage around the obstacles.An accurate measurement method of the plant protection UAV flight path and parameters were proposed.According to the accurate measurement method,the flight path of plant protection UAV during obstacle avoidance was measured,and the position of plant protection UAV relative to obstacles was obtained.The spray dispersal angle design of nozzle and nozzle model selection of the side spray device were carried out combining the relative positions.The plant protection UAV rotor wind field simulation based on the Lattice Boltzmann Method was proposed.According to the rotor wind field distribution,the spatial layout design of the side spray device on the plant protection UAV was carried out.The spraying pattern suitable for the side spray device was explored,and comparison experiments were conducted between the side spray device and non-side spray device under different spraying environment and flight parameters.According to the experiment results,the relationship between the effective spraying width and coverage uniformity of the side spray device with different flight parameters combination was analyzed,and the regression model and linear relationship were obtained.Compared with the condition without side spray device,using side spray device can effectively improve the droplet coverage around obstacles during obstacle avoidance under same flight parameters combination.The effective spraying width was increased by minimum of 6.35%,maximum of 35.32%,and average of 15.25%.The results verify the effectiveness and rationality of the side spray device design method in this article.The error between the predicted and experimental values of droplet coverage uniformity and effective spraying width of the established regression model and linear relationship of the side spray device at different flight altitude,speeds and the interaction between the two is no more than15%.This regression model and linear relationship can guide the spraying operation of the side spray device during obstacle avoidance.In summary,the research in this article improves the environmental perception ability and obstacle avoidance ability of the plant protection UAV in obstacle environments,and provides a new solution for the safe and autonomous flight of the plant protection UAV.It improves the problem of insufficient droplet coverage around obstacles when the plant protection UAV adopts bypass way to avoid obstacles,and improves the effect of plant protection operation.The research in this article can lay the foundation for the similar research and development of the same type device.And also can provide new ideas to improve and promote the application of the plant protection UAV.