The segmentation of unstructured roads, a key technology in self-driving technology, remains a challenging problem. At present, most unstructured road segmentation algorithms are based on cameras or use LiDAR for projection, which has considerable limitations that the camera will fail at night, and the projection method will lose one-dimensional information. Therefore, this paper proposes a road boundary enhancement Point-Cylinder Network, called BE-PCFCN, which uses Point-Cylinder in order to extract point cloud features directly and integrates the road enhancement module to achieve accurate unstructured road segmentation. Firstly, we use the improved RANSAC-Boundary algorithm to calculate the rough road boundary point set, training in the same parameters with the original point cloud as a submodule. The whole network adopts the encoder and decoder structure, using Point-Cylinder as the basic module, while considering the data locality and the algorithm complexity. Subsequently, we made an unstructured road data set for training and compared it with existing LiDAR(Light Detection And Ranging) semantic segmentation algorithms. Finally, the experiment verified the robustness of BE-PCFCN. The road intersection-over-union (IoU) was increased by 4% when compared with the best existing algorithm, reaching 95.6%. Even on unstructured roads with an extremely irregular shape, BE-PCFCN also currently has the best segmentation results.

Quad-frequency signals have thus far been available for all satellites of BeiDou-3 and Galileo systems. The major benefit of quad-frequency signals is that more extra-wide-lane (EWL) combinations can be formed with quad-frequency than with triple- or dual-frequency, of which the ambiguities can be fixed instantaneously in medium and long baselines. In this paper, the long-baseline positioning algorithm based on optimal triple-frequency EWL/wide-lane (WL) combinations of BeiDou-3 and Galileo is proposed. First, the theoretical precision of multi-frequency combinations of BeiDou-3 and Galileo is studied, and EWL/WL combinations with a small noise amplitude factor and a small ionospheric scalar factor are selected. Then, geometry-free methods are used to estimate the a priori precision of EWL/second EWL/WL signals for different combination schemes. Second, the double-differenced (DD) geometry-based function models of quad-frequency configurations and three different triple-frequency configurations are given, and the DD ionospheric delays are estimated as unknown parameters. In the end, the real BeiDou-3 and Galileo data are used to evaluate the positioning preference. The results show that, when using fixed EWL observations to constrain WL ambiguities, the proposed triple-frequency EWL/WL signals composed of (B1I,B3I,B2a) of BeiDou-3 and (E1,E5a,E6) of Galileo can achieve the same precision as the quad-frequency signals. Therefore, the method proposed in this article can realize long-baseline instantaneous decimeter-level positioning while reducing the dimension of matrix and improving calculation efficiency.