In this chapter we have not dealt with uncertainties on the AUV perception of the world. It has been assumed that the vehicle had either an a priori comprehensive knowledge of its environment (chapters 2, 3 and 4) or had a limited visibility (chapter 5). In both cases, precise location of C-obstacles was assumed. This hypothesis is not very realistic as underwater sensors have limited performance. In (Petres, 2007) a dilation of C-obstacles is proposed to improve the safety of the trajectories. This simple method is easy to implement practically but it does not explicitly include uncertainties about sensor specifications. Further work on FM based trajectory planning for AUV using an information space representation would be promising.
On the other hand, uncertainties about AUV configurations have not been considered in the presented trajectory planning methods. This is not very realistic as accurate underwater navigation is still a challenging issue. Nonetheless, concurrent mapping and localization (CML) techniques exist for AUV navigation (Tena Ruiz et al., 2004). This topic is out of the scope of this chapter but further development would benefit from a joint navigation approach for AUV that would couple CML and advanced trajectory planning techniques.
Was this article helpful?