저가형 센서를 이용한 수중정밀항법시스템 디자인

Abstract

This paper describes the implementation of precise underwater navigation system using multiple sensor fusion technique using low cost navigation sensors. Estimation of accurate positions and attitudes is a key issue to achieve qualified mapping results. To implement the precise underwater navigation for qualified mapping solutions, three strategies are selected. Firstly, sensor alignment is conducted to enhance a standalone DR(dead-reckoning) performance. Secondly, outliers due to multi-path are rejected and are not used in the fusion routine. Thirdly, outlier removed absolute position data is fused with DR algorithm results to prevent cumulative integration error of DR algorithm. Least square algorithm is used to identify the alignment matrix parameter such as alignment angle error and lever arm of 600kHz DVL(Doppler velocity log) sensor to the vehicle body coordinate for precise DR implementation. In field experiment, LBL is fused to provide the absolute positions every second. To prevent error accumulation of DVL DR routine, LBL measured positions are fused by using sensor fusion algorithm known as covariance intersection method.ping results. To implement the precise underwater navigation for qualified mapping solutions, three strategies are selected. Firstly, sensor alignment is conducted to enhance a standalone DR(dead-reckoning) performance. Secondly, outliers due to multi-path are rejected and are not used in the fusion routine. Thirdly, outlier removed absolute position data is fused with DR algorithm results to prevent cumulative integration error of DR algorithm. Least square algorithm is used to identify the alignment matrix parameter such as alignment angle error and lever arm of 600kHz DVL(Doppler velocity log) sensor to the vehicle body coordinate for precise DR implementation. In field experiment, LBL is fused to provide the absolute positions every second. To prevent error accumulation of DVL DR routine, LBL measured positions are fused by using sensor fusion algorithm known as covariance intersection method.