基于循环神经网络的车载DR状态估计

由于噪声的不确定性和自身的非线性特征,通过航位推算系统(DR)精确地估计车辆的状态是实际车辆组合导航中最困难的部分。提出了一种基于循环神经网络的方法,和传统的扩展卡尔曼滤波(EKF)方法相比,该方法不仅提高了系统定位的准确性和自适应抗干扰能力;而且不需要模型的具体解析形式,避免了复杂的 Jacobian 矩阵的计算,算法更简单,也更加易于实现。为了检验其有效性,将两种方法分别对车辆 DR 导航系统进行滤波仿真,仿真结果进一步表明该神经网络方法明显优于 EKF 方法,是车载 DR 导航中一种更理想的非线性滤波方法。

State-estimation of Vehicle Dead-reckoning System Based on Recurrent Neural Network

With indefinite noises and nonlinear characteristics, real-time estimating states of the dead reckoning (DR) unit is much more difficult than that of the other measuring sensors, which used in vehicle integrated navigation systems. Compared with the well known extended Kalman filter (EKF), a recurrent neural network was proposed for the solution, which not only improves the location precision, the adaptive ability of resisting disturbances, but also avoids calculating the analytic derivation and Jacobian matrices of the nonlinear system model. In order to test the performances of the recurrent neural network, these two methods were used to estimate states of the vehicle DR navigation system. Simulation results show the recurrent neural network is superior to the EKF and is a more ideal filtering method for vehicle DR navigation.