基于智能滤波技术的抗高过载陀螺信号处理方法

Anti-high-overload gyro signal processing method based on intelligent filtering technology

陀螺作为捷联惯性导航系统的关键传感器，其测量精度直接决定了整个系统的性能和精度指标。针对舰船高过载环境下捷联惯性导航系统陀螺输出信号出现畸变的问题，提出一种基于 BP 神经网络技术的陀螺信号智能模拟滤波方法。该方法根据系统加速度计输出值对舰船运动状态进行判断，当其输出小于设定阈值时，视为非过载环境，此时将陀螺输出用于导航计算并作为 BP 神经网络在线训练样本，以保证网络参数与当前舰船运动态势的一致性；否则视为进入高过载环境，并利用之前最新训练好的BP神经网络模拟当前陀螺信号输出，保证捷联惯性系统的平稳工作。采用智能模拟的优点是：数据并行计算速度快，不需要改变系统硬件条件。半物理仿真试验结果表明：该方法在加速度计输出为5～50g的高过载环境下，可有效改善陀螺输出信号出现畸变的问题，实现舰船运动状态的实时模拟。

Under high-overload ship environment, the gyro signal in strapdown inertial navigation system may have the problem of distortion. To solve this problem, an intelligent gyro signal filtering method based on BP neural network is proposed. This method takes the accelerometer output values as the threshold to judge whether the high overload happens or not. If the output is less than the setting threshold, it means a non-overload environment, in this case the gyro outputs are used for navigation calculation and used to train BP neural networks online, to ensure that the network parameters are kept consistent with the current situation of the ship movements; otherwise it means high overload happened, and the latest trained BP neural networks are used to replace the gyro signal outputs, which can ensure the strapdown inertial navigation system works smoothly. Adopting intelligent filtering technology has lots of advantages, such as fast speed in parallel calculating, and not having to change the system hardware environment. The simulation results show that this method can effectively improve the problem of the gyro signal distortion under high overload environment with 5g to 50g accelerometer outputs, and achieve real time calculating of ship movements.