Research On Terrain Following Technology Of Plant Protection UAV Based On Millimeter Wave Radar

Compared with artificial or ground machinery,the plant protection UAV(unmanned aerial vehicle)has advantages in spraying efficiency,scope of application,saving water and medicine and reducing pollution.The UAV needs to keep a constant distance from the plant during the operation,and this process is the terrain following.The UAV in sloping fields or tea garden usually operated at a height of 0.5~3m above the crops.The UAV needed to maintain a stable terrain following to ensure the safety of flight and spray uniformity.The domestic plant protection UAV is mainly based on remote control multi-rotor,which is more suitable for the operation in the terrain with less fluctuation,and it is difficult to meet the needs of plant protection operation with fluctuating terrain.Research institutions and related enterprises at home and abroad are vigorously developing intelligent terrain following complex terrain.Therefore,terrain following is one of the key technologies of automatic plant protection UAV,which has very important research and application value.The study of terrain following is carried out in the following aspects.(1)Terrain following algorithm for plant protection UAVIn this paper,the application scenario of plant protection UAV,taking the most common sloping fields and tea plantations in plant protection environment as the research object,the millimeter wave radar is applied to the terrain following system,a terrain following method based on millimeter wave radar sensor is proposed.The method includes six parts: basic parameters of terrain calculation,kalman filter,multi-radar altitude information fusion algorithm,gap identification algorithm,gradient memory algorithm and fuzzy PID control algorithm.In the sloping scene the pre millimeter wave radar was used for slope judgment.When the slope was small,the height of the differential GPS and the height of the ground penetrating millimeter wave radar were filtered by kalman filter,which can improve the accuracy.When the slope changed beyond the threshold,the height of the pre millimeter wave radar and ground penetrating millimeter wave radar were fused by multi-radar information fusion,which can improve the response speed.According to the height difference between the tea tree and the gap in the flat tea garden scene,the gap identification algorithm is used to filter the gap between the tea trees and only follow the height change of the tea tree.In the sloping field of tea plantation,the vertical velocity of UAV is calculated according to the slope.The change of the gradient and the positive and negative gradient are used to judge whether the tea tree gap is detected or not,and the vertical velocity is kept unchanged by the gradient memory algorithm when the gap is detected,UAV does not respond to the height of the gap only to follow the gradient of the tea plantation.Finally,according to the characteristics of the plant protection UAV variable system,the fuzzy PID algorithm is designed to control the height of UAV.(2)Plant protection UAV terrain following systemThree millimeter-wave radars were installed on the "W730S" platform.The millimeter wave radar,attitude sensor,DGPS sensor and data fusion module(PICO-CV01)were selected to form a data acquisition module for the terrain following height of plant protection UAV.The positive and negative angle of elevation is used to decide whether the UAV is advancing or backward.Therefore,we choose the height data of the front radar or the post radar,and adjust the height data of the front radar or the post radar through the angle of pitch angle.Millimeter wave radar sensor and data fusion module(PICO-CV01)use serial communication protocol to acquire the latest detection height from high data acquisition module and complete data preprocessing work such as extraction and analysis of high data.Then the corresponding relationship between high data and topographic data is established according to the validity of the highly detected distance values.The basic parameters of slope,plant height and time needed for uplift are used to describe the landform features.The formulas for calculating the basic parameters of slope,height of tea tree and control height are given.(3)Plant protection UAV terrain following softwareThe terrain following software is divided into three parts,that is,the height detection section,the height control part and the exception handling part.The exception handling function is distributed in the height detection code and the height control code,and is responsible for monitoring the running state of each sub module in the whole terrain following system.Multithread and software timer are applied to ensure that the height detection part and the height control part operate independently and satisfy the real-time requirement of UAV terrain following system.The height detection section is responsible for the exploration of the terrain,including pre flight self-examination,data acquisition,and data processing.The height control part is responsible for deciding the suitable height control instruction based on environmental information,terrain information,aircraft status information and other factors and restrictions.And completed the design of terrain following computer system software,we can set the radar parameters at the interface of the software,check the terrain following system,set the flight height,according to the operation scene selection hilly or flat tea plantation or sloping tea garden to follow,displaying real-time height data.(4)Feasibility and error verification of terrain followingEight experiments are carried out to verify the accuracy of the terrain following and the effectiveness of the whole terrain following the system.The first experiment calibrated the millimeter-wave radar and verified the accuracy of millimeter-wave radar ranging above the plant.The second experiment verified that spraying and propeller rotation at high speed did not affect the height detection of millimeter-wave radar.The third experiment verifies that millimeter-wave radar can ensure the accuracy and stability of altitude detection in dynamic flight.The fourth experiment verifies that the fuzzy PID control algorithm can improve the response capability and stability of the system.The fifth experiment verifies the accuracy and effectiveness of the kalman filter algorithm in a small gradient scenario.The sixth experiment verifies that the multi-radar altitude information fusion algorithm can respond to the change of slope ahead of time and ensure the accuracy of terrain following.The seventh experiment validates the validity of the tea tree gap identification algorithm,which can neglect the height variation of the gaps in the identification process to ensure the accuracy of terrain following.The eighth experiment verifies the feasibility of gradient memory algorithm.When the tea tree gap is detected,the accuracy of terrain following is guaranteed by keeping the vertical velocity unchanged.Experiments show that the terrain following system based on the millimeter-wave radar sensor can maintain the plant height unchanged in the complex terrain such as the slope and the tea plantation so as to ensure the environment adaptability and spraying effect of the UAV.This paper validates the effectiveness of the terrain following algorithm based on three millimeter wave radar sensors.