基于微机电系统与径向基函数的水下航行器姿态检测

为实现水下航行器自主运动控制，准确掌握水下航行器的运动姿态至关重要。微机电系统惯性传感器因具有体积小、成本低和功耗低等特点，被广泛应用于姿态检测系统当中。由于海底的工作环境复杂，故水下航行器在实验过程中设定有一定的非合作性；因此设计采用径向基函数神经网络算法对采集数据进行训练与分类。实验结果表明，微机电系统惯性传感器结合径向基函数神经网络算法对水下航行器的基本行为识别的平均召回率可达94%。

Attitude Detection of Underwater Vehicle Based on MEMS and Radial Basis Function

In order to realize the autonomous motion control of the underwater vehicle, it is very important to master the motion attitude of underwater vehicle accurately. Current micro-electro-mechanical system (MEMS) inertial sensors have small size and low cost. It is widely used in attitude detection systems because of its low power consumption and other characteristics. Due to the complex working environment of the seabed, the underwater vehicle is set to be non-cooperative. Radial basis function neural network algorithm is used to train and classify the detection data. The experimental results show that the average recall rate of micro-electro-mechanical system inertial sensor equipped with radial basis function neural network algorithm for basic behavior recognition of underwater vehicle is 94%, which has certain engineering practical significance.