无人机姿态视觉导航位置测量仿真研究  

Simulation Research on Attitude Measurement Based on Visual Navigation for Unmanned Aerial Vehicle

作　　者：黄颖辉[1] HUANG Ying - hui(College of Electronics and Information, Nantong University, Nantong Jiangsu 226019, China)

机构地区：[1]南通大学电子信息学院,江苏南通226019

出　　处：《计算机仿真》2017年第4期51-54,60,共5页Computer Simulation

基　　金：江苏省政策引导类计划-前瞻性联合研究项目(产学研合作)(BY2015047-08)

摘　　要：无人机姿态视觉导航位置测量可满足无人机在复杂环境下的导航需求。针对姿态视觉导航位置测量系统中存在姿态测量误差累积的问题,提出了一种基于最小二乘支持向量机左逆的无人机姿态视觉导航位置测量方法。方法根据"内含传感器"左逆原理,假设在无人机内部存在一个以姿态为输入的内含传感器子系统,通过证明该子系统左可逆性,得到姿态的左逆模型。然后利用最小二乘支持向量机辨识左逆模型,同时将模型串联到原系统之后,构成姿态测量的伪线性系统,即可实现无人机姿态的测量。仿真结果表明,所提方法姿态测量精度高,性能稳定,能有效满足一般无人机姿态测量的要求。The attitude measurement based on visual navigation for unmanned aerial vehicle （UVE） can meet the demand of UVE navigation under complicated environment. To reduce the error accumulation of current attitude measurement systems in unmanned aerial vehicle, a novel attitude measurement strategy based on least squares sup- port vector machines left inversion is proposed. According to the ＂assumed inherent sensor＂ inverse theory, one may assume that, in UVE interior, there exists an ＂inherent sensor＂ subsystem whose inputs are exactly flight attitudes. The left - inevitability of the subsystem is proved and the left - inverse model is obtained. Then least squares support vector machine is used to identify the left -inverse model. Finally, by cascading the ＂assumed inherent sensor＂ in- version to the ＂assumed inherent sensor＂, the Attitude measurement of unmanned aerial vehicle can be obtained. Simulation results show that the novel method has higher precision and better stability performance, able to meet the requirements of general UAV attitude measurement.

关 键 词：无人机 视觉导航 姿态测量 左逆 最小二乘支持向量机 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计]