The layout of seafloor datum points is the key to constructing the seafloor geodetic datum network, and a reliable underwater positioning model is the prerequisite for achieving precise deployment of the datum points. The traditional average sound speed positioning model is generally adopted in underwater positioning due to its simple and efficient algorithm, but it is sensitive to incident angle related errors, which lead to unreliable positioning results. Based on the relationship between incident angle and sound speed, the sound speed function model considering the incident angle has been established. Results show that the accuracy of positioning is easily affected by errors related to the incident angle; the new average sound speed correction model based on the incident angle proposed in this paper is used to significantly improve the underwater positioning accuracy.

Systematic error is one of the major factors that affect positioning accuracy owing to the changeable and complex nature of seawater environments. Based on a Global Navigation Satellite System-acoustic intelligent buoys system, whose acoustic array consists of a series of surface buoys, a single-difference method for underwater dynamic positioning is proposed to eliminate systematic error. Positioning configuration optimisation was addressed using dilution of precision (DOP). A simulation of DOP proved that for the single-difference method, a radiation network with a centre-difference reference point was superior to a regular polygon network. The positioning experiment showed that the novel method could effectively eliminate systematic error, improving vertical positioning accuracy from a metre- to a decimetre scale.