INTEGRATION OF SPATIAL DATABASE SYSTEMS AND SAMPLING-BASED PATH PLANNING FOR OPTIMIZING MARITIME NAVIGATION

Abstract

This paper addresses the challenge of representing configuration spaces for sampling-based path planning in maritime navigation scenarios by using data-efficient vector maps stored in spatial database systems. The proposed approach optimizes the performance of fundamental algorithm operations in large environments, such as collision checking, by using spatial indexing to efficiently reduce the number of geometries that need to be evaluated by computationally expensive spatial predicates. An implementation combining the SpatiaLite database system, standardized electronic navigational charts (ENC) map features, and the widely used Open Motion Planning Library (OMPL) demonstrates practical applicability. The implemented system provides collision-free paths for maritime navigation, includes a graphical user interface, and is incorporated to a system for autonomous surface vehicles. Simulations show that the implementation supports multiple planning algorithms in generating valid paths in four representative large-scale maritime environments: a cluttered archipelago, a river inlet, a peninsula, and a fjord transit.