The assessment of PPP-B2b kinematic positioning performance under different scenarios will be an important reference for users and contribute to the expansion of the application scope and promotion of BDS-3. In this study, the kinematic positioning performance of the PPP-B2b service of BDS-3 under different scenarios was comprehensively analyzed by performing one vehicle experiment and one vessel experiment. Experimental results indicate that using the post-processing product released by the Helmholtz Centre Potsdam-German Research Centre for Geosciences (GFZ) as a reference, the PPP-B2b product had orbit correction accuracy up to the decimeter level and clock correction accuracy up to the sub-meter level. In the vehicle experiment of PPP-B2b, multipath-root mean square (MP-RMS) for BDS-3 B1/B2 frequencies was 41.3 cm, while that for GPS L1/L2 frequencies was 52.2 cm. The positioning accuracies of BDS-3 in the horizontal and vertical directions were 10.3 and 10.5 cm and those of the BDS-3+GPS combination were 5.6 and 4.9 cm, respectively. In the vessel experiment of PPP-B2b, the MP-RMS for BDS-3 B1/B2 frequencies was 52.5 cm, while that for GPS L1/L2 frequencies was 70.4 cm. The accuracy values of BDS-3 in the horizontal and vertical directions were 22.5 and 12.0 cm and those of the BDS-3+GPS combination were 9.7 and 5.1 cm, respectively. The overall impact of multipath effects on the kinematic positioning performance of PPP-B2b in the vessel environment was slightly greater than that in the vehicle environment.

Oceans are not only super-ecosystems but also strategic resource reservoirs, and thus, ocean monitoring is crucial. Unmanned surface vehicles (USVs) are new types of multifunctional unmanned platforms for ocean monitoring, and path planning plays a crucial role as a core technology in their operation. With the continuous increase in maritime traffic density and upgrading of navigation safety standards, traditional path planning methods are facing growing challenges in adapting to complex environments. In this study, a multidimensional improvement strategy is proposed to address the limitations of the bidirectional rapidly-exploring random tree star(Bi-RRT^*) algorithm in USV path planning. First, an adaptive step-size adjustment mechanism, based on environmental feature perception, is established; second, a key node selection strategy is designed; and finally, Bezier curves are used to smooth the generated path, producing a smoother trajectory that better meets the kinematic requirements of USVs. Simulation results show that the improved bidirectional RRT^* algorithm outperforms its traditional counterpart in terms of node-generation efficiency, overall performance, and path smoothness.