Abstract: With the rise of artificial intelligence, ships are gradually becoming intelligent. Ship path planning, as the foundation of ship intelligence, has become a current research hotspot. To achieve the planning of ship collision avoidance paths, a hybrid ant colony algorithm and artificial potential field for optimal safety path planning was proposed. The model makes up for the disadvantages of the slow convergence rate and the easy local optimum. At the same time, the ant colony algorithm in turn makes up for the poor global search ability of the artificial potential field method, so that the algorithm can achieve accurate path planning. The test results show that in a simple environment, the hybrid ant colony algorithm, improved tangent and Dijkstra algorithm, and improved fast extended random tree * algorithm iterated about 7, 25, and 40 times respectively before starting to converge; In complex environments, the improved tangent and Dijkstra algorithms do not converge, while the improved fast expanding random tree * algorithm and hybrid ant colony algorithm iterate about 40 and 8 times respectively to begin convergence. It is clear that the mixed ant colony algorithm converges fast and can obtain the optimal path in the shortest time. The number of unsafe path points for the optimal path in a simple environment is 6, 1, and 0, respectively, for the improved tangent and Dijkstra algorithm, the improved fast expanding random tree * algorithm, and the hybrid ant colony algorithm; The number of unsafe path points for the optimal path in a complex environment is 9, 3, and 0, respectively; It can be seen that the path planned by the hybrid ant colony algorithm has higher security. The above results show that the mixed ant colony algorithm can effectively shorten the search time of optimal paths while improving the security of pathways.
Keywords: Ant colony algorithm, artificial potential field, path planning, path smoothing strategy, inland river navigation, ship avoiding collision
