Multi-rotor Control System Applied In Farmland Information Acquisition

Agricultural aviation technology is a hotspot of precision agriculture.Multi-rotor UAV has good adaptability to irregular terrain.It is widely used in agricultural information and land resources monitoring.Multi-rotor can timely and accurately collect information of crop growth,which provides convenience for production managers to make quick,accurate decisions.It is an important part of precision agriculture and significant for promoting the modernization of agriculture.The purpose of this research was to develop multi-rotor control system applied in farmland information acquisition.The flight control system was composed of multi-rotor platform,embedded controller and multiple pose sensors.Fuzzy-PI(Proportional and Integration)deviation correction method and extended Kalman filter based on the hardware system designed in this paper were proposed to measure the multi-rotor attitude and provide accurate attitude information for attitude control.Double closed-loop expert PID control strategy was designed to control flight attitude.Altitude control based on multi-sensor measure was presented to control flight altitude and keep the multi-rotor flying in certain height.Flash memory was adopted to store the control system parameters.A wireless data transmission system was designed to transmit the multi-rotor status information to the host computer.The contents of this paper were as follows.(1)Open source projects on multi-rotor were introduced.The flight control system construction,attitude estimation and control strategy of the open source projects were investigated and searched to provide reference for flight control system study in this thesis.APM flight controller was selected as the test object from open source projects.Configuration and flight test results of the APM controller were given in this section.(2)The construction of multi-rotor applied in farmland information acquisition was presented.The multi-rotor was electric powered.Body frame,battery,motors,electronic speed controllers and propellers of the multi-rotor were selected.The settings and the communication interfaces of the main controller and sensors in the control system were introduced.The main controller of the system was STM32F407.I2 C,USART,SPI interfaces were used for communications between the main controller and the sensors.The timers in STM32F407 received the remote control input signals and output motor control signals.The characteristics,working principles and specific settings of gyroscope and accelerometer MPU6050,compass AK8975 and barometer MS5611 were introduced.(3)According to the rotor-layout is X-type,the force analysis,movement analysis and control principle of multi-rotor were presented.The navigation coordinate system and the body coordinate system were defined.The relationship between the two coordinate systems was described.Quaternion method was used to represent the attitude of the multi-rotor.As the feedback part in the closed-loop of multi-rotor attitude control,accurate and stable attitude measurement is especially important.According to the characteristics of the sensors,calibrations were taken to improve the measure accuracy.Fuzzy-PI deviation correction multi-sensor information fusion method based on quaternion was proposed.The static and dynamic test results indicated that the attitude measure system had a good real-time performance.It can satisfy the accuracy requirement of multi-rotor attitude measure and had quickly respondence to the multi-rotor movement.In the long period static measurement,the measure angle did not produce static drift or divergence.(4)Kalman filter is a classical recursive algorithm based on minimum variance estimation,and it has been applied in multi-sensor attitude measure.The hardware of the multi-sensor attitude measure system was composed of STM32F407 controller,6-axis motion process sensor MPU6050 and 3-axis electronic compass AK8975.Extended Kalman filter estimation method based on the hardware system designed in this thesis was proposed.The filter compensated the attitude measure error caused by gyro drift by accelerometer and magnetometer measure information.Through experimental comparisons,the optimized value combination of covariance in Kalman filters was achieved.The system did not produce static drift or divergence in the static measurement.In the dynamic experiment,the measure error was less than fuzzy-PI deviation correction method.But because of the large computation,Kalman filter update rate was slower than the fuzzy-PI deviation correction method.(5)The multi-rotor control system needs to manage multiple tasks including attitude estimation,input and output signals process,control algorithm operation.Task scheduling method was adopted to arrange the system operations.The input signals were processed by validity checking,control input conversion,low-pass filter,lock and unlock switch operations.Because multi-rotor flight with high-speed motors rotation,its stability is vulnerable to natural wind,electromagnetic interference and other factors.The double closed-loop expert PID control strategy in the attitude control was presented.Appropriate PID parameters were selected by engineering debugging method.Expert control method was introduced to adjust the control strategy when the multi-rotor status changed in control process.The control system was testified by anti-interference and step input tests.Its outdoor farmland flight data was also collected.The data showed that the double closed-loop expert PID control strategy made the system stable,controllable,and have anti-disturbance ability.It can satisfy the attitude control requirement of multi-rotor applied in farmland information acquisition.(6)In the height control,MS5611 barometer and US-100 ultrasonic sensor were adopted as height measure sensors.A multi-sensor fusion method of barometer,ultrasonic sensor and accelerometer was designed to estimate the multi-rotor vertical speed and height.The single closed-loop PID based on barometer vertical velocity control strategy and the double closed-loop based on ultrasonic height control strategy were proposed.The experiment data showed that the throttle input can effectively adjust the height position of the multi-rotor.The vertical velocity measure and the height measure based on multi-sensor fusion method of barometer,ultrasonic and accelerometer can satisfy the multi-rotor height control requirement.The single closed-loop PID vertical speed control strategy and double closed-loop height control strategy can effectively control the multi-rotor vertical movement and keep it in a certain height.The height control system can meet the height control requirements of multi-rotor applied in farmland information acquisition near ground.(7)The control system of the multi-rotor applied in farmland information acquisition should have saving system parameters function and wireless transmit the multi-rotor status information.Flash W25Q32 was used to save sensors offset values and control parameters.FATFS was used for users to modify the system parameters according to the flight environment.At the same time,a wireless data transmission system was adopted to deliver the multi-rotor status information to the host computer.In order to ensure the correction and reliability of wireless transmission data,a data communication protocol was designed to avoid receiving data errors caused by packet loss.