In this paper, authors proposed a practical motion estimation strategy to obtain the single static object position and the vehicle's motion parameters simultaneously by utilizing camera and IMU (Inertial Measurement Unit), and tried several classic nonlinear parameters estimation methods on this problem based on the Matlab simulation. A lot of preliminary calibration experiments have been done to fuse the data from IMU and camera, which included: camera calibration, IMU error modeling, and relative attitude calibration between IMU and camera. The obtaining of sensor's model made the simulation work be possible, and the PF algorithm can track the vehicle's motion parameters and obtain the target position very well. The accuracy is centimeter level.


IAUV(Intervention-Autonomous Underwater Vehicle) is a kind of special underwater vehicle which can perform the tasks autonomously, such as installing instruments on the sea floor, searching, locating, approaching and docking on the underwater device; retrieving geological/biological sampling etc (Wang,1996). Currently such tasks mostly are performed by ROV (Remotely Operated Vehicles), and only few intervention missions have be performed in the real sea environment by IAUV(Marani, Choi and Yuh, 2009). In order to perform the underwater intervention missions, imaging sonar, ultralow- light camera, Inertial Navigation System, Long Base Line System, doppler sonar need to be fixed on the vehicle(Evans and Redmond, 2003), and the general intervention mission can be performed following the six stages process. 1) Launching from R/V ship. 2) Searching and navigating to the worksite(∞ m- 25m). In this stage, sonars (image sonar, multibeam sonar or sidescan sonar) are used to search the suspected targets, navigation sensors (LBL(Long Base Line), doppler sonar, INS(Inertial Navigation System), image sonar and high precision camera(Stefan and Ian, 2004)) are used to track the vehicle's position and attitude. 3) Approaching the target(25m-2m).

This content is only available via PDF.
You can access this article if you purchase or spend a download.