基于过程神经网络的移动机器人RTK-GPS 导航定位研究

Abstract

China's self-developed Beidou 3 system has been fully operational and has achieved global positioning. In order to further improve the ground mobile robot terminal satellite navigation and positioning function, data reception for mobile robot parsing of high-frequency oscillation random disturbance signal and the system of higher order nonlinear dynamic influence on navigation and positioning accuracy, a process neural network is used the time-varying characteristics of building dynamic adaptive RTK-GPS positioning algorithm. Using existing satellite positioning terminal input and output data of the neural network model is constructed, using the dynamic error data as samples to train the neural network model correction, the satellite signal interference or process of loss of lock using the trained neural network to predict the output divergence to suppress the position and velocity error, and thus improve the accuracy of positioning and navigation. Experimental results show that the proposed method is still suitable and effective for improving the positioning accuracy under the condition that the positioning interference noise is unconstant, and can significantly reduce the error of positioning results, especially when the number of observable satellites changes.