Coverage Path Planning for Ship Hull Based on Improved GBNN Algorithm

Robots for ship hull cleaning are getting increasing attention to reduce fuel consumption and protect the marine environment. Since the hull surface is a complex curved surface, coverage path planning on the curved surfaces needs to be studied. This article introduces a map-building method from the ship’s table of offsets based on connection groups, and then extends the Glasius bio-inspired neural network (GBNN) algorithm to 3D space. Based on the connection map, an improved GBNN algorithm is developed, which has a much smaller amount of data and computation. In addition, the ant colony optimization (ACO) algorithm is used for speeding up the escape from the dead zone and reduce the total path length. The simulation shows that the proposed algorithm realizes coverage path planning on curved hull surface with small computational quantity and low repetition rate, which can be applied to the hull cleaning robot in the future.