The Research Of Agricultural Vehicles' Attitude Measurement And Data Fusion Algorithms

Agricultural robot is one of the most effective methods to increase agricultural automation level in the world. It can meet the future agriculture of good harvest, safety, high efficiency, environmental protection, agronomic's high precision requirements. Automatic vehicle navigation technology is the basic problem. A navigation system is built based on GPS/DR and 3G technology. Key problems of sensor data fusion in navigation process, such as state estimation of agricultural vehicles, route tracking of agricultural vehicles, later evaluation of navigation information, was discussed as below:①Using global positioning system technology, combined with DR technology, a navigation system of low cost equipped on agricultural vehicles was developed. The sensors used in the system are GPS, Inertial navigation sensors such as electronic compass, fiber optic gyro. The software was developed with Labview. The software is used to configure serial port, extract vehicle attitude information and so on. Differential GPS Signal was used in the system to observe vehicle attitude, which was transmitted through the 3G network. The DGPS signal is used to correct the GPS data the mobile station collected.②Using the Simulink in the MATLAB, the dynamic model of field work machinery was developed. It includes:the front wheel side force calculation unit, the back wheel side force calculation unit, steady turning radius calculation unit, yaw angles gain steady-state response unit, Insufficient steering parameters calculation unit and so on. When function parameters and structure parameters is given to the calculation model, vehicle's yaw angles and centroid running trajectory can be calculated quickly.③Refers to multi sensor data fusion, using the Kalman filtering method based on iteration which has the minimum mean square error(MMSE), The attitude's optimal estimation of the agricultural vehicles can be obtained. Considering the state equation of the system is nonlinear equation, extended kalman filtering (EKF) method is used to solve this defect. The test result shows that the method is feasible.④The navigation system's work depends on different sensors, when using centralized filter method to deal with the signals sensors get, the damage of the subsystem can cause paralysis of the whole system. Distributed filter method is used here. A federal kalman filtering system is designed here, using this method, the system's fault tolerance is improved. At the same time, the amount of calculation of the system is reduced.⑤In the process of navigation of the vehicles, when GPS fails to get the signal, inertial navigation can be used instead of GPS navigation system. Inertial navigation system's error will accumulate over time. We use the optimal smooth theory to smooth the data get from sensors, the test results shows that this method is useful.⑥In the case of autonomous vehicles, special route is given to the vehicle to avoid scratching crops, reduce energy consumption, reduce cost of production A fuzzy controller is designed for linear tracking and curve tracking. To deduce system overshoot, the fuzzy controller's parameters can be adjusted. The test results shows that using this method, the system overshoot is deduced.