北斗_GNSS应用与产业化

CSNC2010第一届中国卫星导航学术年会 北京

顾及姿态角更新的低成本车载GPS/INS组合导航算法

吴富梅1,2 聂建亮3 何正斌3

(1.信息工程大学，测绘学院，河南，郑州 450052；

2.西安测绘研究所，陕西，西 安 710054； 3. 长安大学 地测学院，西安 710054)

摘 要：低成本INS系统的元件误差严重影响INS导航精度。本文针对车载系统，提出一种顾及姿态角更新的低成本车载GPS/INS组合导航算法。首先在GPS/INS组合导航Kalman滤波方程基础上，给出两种姿态角更新的观测方程。然后给出了利用GPS测速确定航向角的原理，并且对低成本车载INS系统的俯仰角和翻滚角进行了分析，指出由INS随机误差造成的俯仰角和翻滚角误差比其本身量值要大，建议令俯仰角和翻滚角数值保持不变。利用实测算例确定了不同速度下的航向角精度，并且验证了本文提出的算法的有效性，相对于基于位置、速度组合的Kalman滤波，导航精度有明显提高。

关键词：姿态角更新；GPS/INS组合导航；Kalman滤波；GPS测速

Low-cost GPS/INS Integrated Navigation Algorithm in Land

Vehicle System Considering Attitude Update

WU Fu-mei1,2 NIE Jian-liang3 HE Zheng-bin3

(1.Institute of Surveying and Mapping , Information Engineering University , 450052, China;

2. Xi’an Research Institute of Surveying and Mapping, 710054, China;

3. School of Geological and Surveying Engineering, Chang’an University, Xi’an 710054 , China;)

Abstract：The INS instrument errors can degrade the navigation precision badly. In this paper, a new algorithm considering attitude update in GPS/INS integrated navigation for land vehicle is presented. Two kinds of attitude observation equations are given based on GPS/INS loose navigation. And the principle of yaw angle determined by GPS velocity is introduced and the pitch and roll angles of low-cost INS in land vehicle are analyzed. It is suggested that pitch and roll angles can be set consts when the errors caused by instrument errors are bigger than themselves. By the actual data calculation, the precision of yaw angle determined by GPS velocity is comfirmed and it is shown that the new algorithm can improve the navigation accuracy greatly compared with the Kalman filtering based on position and velocity.

Keywords: attitude update, GPS/INS integrated navigation, Kalman filtering, velocity determination by GPS

1 引言

按照载体的运行状态，INS姿态角更新分为静基座姿态角初始对准和动基座姿态角更新。顾名思义，静基座初始对准就是载体在静止的状态下完成姿态角对准的过程，而动基座姿态角更新就是载体在运动的状态下对INS姿态角进行更新的过程[1]。

目前对于静基座初始对准的研究成果较多，早期利用古典控制法先进行水平对准后进行方位对准；后来随着滤波技术的发展，Kalman滤波被引入到初始对准中；随着精度要求的提高，为抵制噪声干扰影响提高系统的抗差性，H∞滤波技术应用到初始对准中[2]；顾及状态误差的影响，利用神经网络训练样本补偿系统误差的影响[3]；为缩短对准时间，利用遗传算法全局性和并行性的特点，在较

1