Exploration with loop-closing in depth-fixed navigation for autonomous underwater vehicle

Autonomous underwater vehicle (AUV) has to face many intractable problems while implementing indeed autonomy. In the past years, the method of simultaneous localization and mapping (SLAM) has prevailed and provided a feasible way to autonomous navigation. Loop-closing problems of AUV are discussed in this paper. Loop-closing is to identify that AUV has returned to the already visited areas, and localization error can be reduced by re-entering visited areas, so more accurate underwater feature maps can be obtained. We use the FastSLAM based on the particle filter to estimate AUV pose and obstacle objects (landmarks). The experimental results show that our approach is better than the approaches lacking the ability of actively closing loops.