Data-driven Based Automatic Routing For Unmanned Ship And Surface Vehicle

As unmanned marine transportation platform, unmanned ship and surface vehicle can undertake wide-range and low-cost task for a long time. In recent years, with the improvement of ship intelligent level and the development of unmanned driving technology, unmanned ship and surface vehicle has been tentatively applied in both military and civilian fields and achieved some success. The research of unmanned ship and surface vehicle is multi-discipline coordinated and multi-field integrated program with certain innovation and prospection. In this paper, data driven based automatic routing for unmanned ship and surface vehicle is researched.With the implementation of national Big Data strategy and the rise of Data Science, data is gradually transforming from a sort of simple process object into a kind of strategic resources; data driven-based analysis and research methods are also profoundly changing the traditional exploratory way of scientific research, and become a new pattern to promote the development of modern society and the advancement of science and technology. Traditional marine construction of infrastructure and information could provide the function of supervision and service; meanwhile, massive data has been accumulated.In the context of Data Technology era, based on the massive trajectory data and data driven-based method, the main target of this paper is to explore a new automatic routing method for unmanned ship and surface vehicle. The main work of this paper is listed as follows:Trajectory data simplification is one of the important component and key technology in data driven-based automatic routing method for unmanned ship and surface vehicle. Large sets of trajectory data create problems of storing, transmitting and processing. After appropriate pre-processing, e.g. data parse, data cleaning; a scheme and the implementation of the Douglas-Peucker algorithm are presented. As for simplification threshold, the solo parameter in algorithm, a new trajectory-based minimum ship domain evaluation method is proposed and acted as criteria for the simplification threshold determination. Finally, a validation is made to examine the effectiveness of the method and the rationality of the simplification threshold. The result indicates that the Douglas-Peucker algorithm can simplify trajectory data effectively and preserve trajectory characteristic reasonably.Based on the characteristic points extracted in the step of trajectory data simplification, according to the specifications of International Regulations for Preventing Collision at Sea and good seamanship, the indicator of turning point is set and turning points are discriminated from characteristic points. In view of the density disparities in different marine area, a hierarchical DBSCAN clustering algorithm is proposed that keeps the simplicity of the original algorithm with certain innovation to produce a more accurate cluster result. By setting similarity measurement with spatial proximity and turn similarity into consideration, the isolated turning nodes can be obtained by clustering turning points. Then, the linkage of between turning nodes is detected using the simplified trajectory data, thus a coherent chain modeled as a directed graph G= (TN, L) can be get, where TN refers to turning node describing the geometry of route network and L refers to linkage describing connectivity of the route network. Finally, the Ant Colony Optimization is used to find the optimal route. The example verification indicates that method proposed is effective and reasonable; the automatic generated route is compliance with the basic principle and specification of route planning.Traditional route safety validation is accomplished by visual assessment. In order to change this mode, IHO S63 standard Electronic Navigational Chart (ENC) data based route safety automatic validation method is proposed. First, after implementing Blowfish encryption algorithm, DSA digital signature algorithm, SHA-1 secure hash algorithm and CRC32 Cyclic Redundancy Check algorithm, the ENC data is parsed according to IHO S63 Data Protection Scheme. The potential risk is analyzed on the basis of Formal Safety Assessment framework and the corresponding risk control options are given. Finally, the IHO S63 standard ENC data is transformed into data complying with IHO S57 transfer standard and IHO S52 chart content and display standard and eventually converted into SENC. By extracting feature elements, e.g. soundings, point features, line features and area features, the route is validated automatically.Conclusions and future work are addressed at the end of this paper.