一种多时间尺度融合的视觉辅助惯性定姿算法
2020-06-19张琳曾成王羿帆
张琳 曾成 王羿帆



摘 要: 针对基于微机电系统的惯性传感器中,陀螺仪在姿态解算过程中存在误差累积的问题,提出一种多时间尺度融合的视觉辅助惯性定姿算法。利用视觉传感器在不同时间尺度的角度变化量,不断修正并更新惯性器件的角度累积误差,对姿态角进行多次校正,进而输出较高精度的姿态角。实验结果表明,采用多时间尺度融合的视觉辅助惯性定姿算法,相对于单时间尺度的融合算法,航向角误差降低了27%,俯仰角误差降低了23%,横滚角误差降低了63%,角度误差能够保持在2°以内,满足定姿定位的需求。
关键词: 姿态解算; 多时间尺度融合; 惯性测量单元; 数据采集; 姿态角校正; 角度误差更新
中图分类号: TN820.2+2?34; TP212.9 文献标识码: A 文章编号: 1004?373X(2020)12?0014?04
Abstract: In allusion to the error accumulation in the attitude resolving of the gyroscope in the inertial sensor based on micro?electro mechanical system, a visual aided inertial attitude determination algorithm based on multi?time scale fusion is proposed. By means of the angle variation of the vision sensor at different time scales, the angle cumulative error of the inertial device are corrected and updated continuously, and the attitude angle is corrected for many times to output the attitude angle with higher precision. The experimental results show that, in comparison with the single?time scale fusion algorithm, the visual aided inertial attitude determination algorithm based on the multi?time scale fusion reduces the course angle error by 27%, the pitch angle error by 23%, and the roll angle error by 63%. The angle error can be able to keep within 2° to meet the needs of attitude?determination and positioning.
Keywords: attitude resolving; multi?time scale fusion; inertial measurement unit; data acquisition; attitude angle correction; angle error renewal
0 引 言
基于微机电系统传感器的惯性测量单元(Inertial Measurement Unit,IMU)在惯性定位技术中得到了广泛应用,尤其是卫星信号不能穿透建筑物的室内环境[1]。由于IMU中的陀螺仪存在较高的漂移误差,而且捷联惯导系统在解算过程中采用积分的方法,导致内部误差和外部误差不断累积,并且缺乏自我纠正的手段,远远达不到高精度室内定位的需求。视觉传感器误差不随时间累积,成本低,但比较依赖光照条件和可视环境[2]。将视觉传感器和惯性传感器的信息进行优势互补:求解载体的姿态采用以惯性传感器为主,视觉传感器为辅的数据融合方式[3]。在无视觉输出信息或视觉信息不充足的情况下,IMU能够持续提供载体的姿态变化;在视觉信息充足的情况下,视觉信息抑制IMU的姿态发散,校正惯性器件的参数,保证载体的姿态精度。……
