The Path Planning Of Unmanned Surface Vehicle Based On Artificial Potential Field And Ant Colony Algorithm

In recent years, with the continuous improvement of the comprehensive national strength, China has been embroiled in maritime disputes with some neighboring countries, especially in the South China Sea. In order to deal with the complex and changeable environment of sea, we should develop the technology of marine intelligent equipments to protect our maritime sovereignty. Unmanned Surface Vehicles(USV) have the features of small size, flexibility and high-level intelligence. As an important part of the marine transportation system, the research on the path planning of USV has great significance to be widely used in the military and civilian fields. Based on the current technologies of USV path planning, this paper has done some deep researches.This paper analyzes the advantages and disadvantages of different algorithms of path planning and finally chooses the artificial potential field(APF) method and ant colony algorithm(ACA). This paper introduces the fundamental principles of APF and ACA respectively firstly, and then proposes an improved algorithm combining the two algorithms. The main work of this paper is as follows:At first, aiming at the problems in the practical application, a new APF algorithm of path planning is proposed with the improvements of field function, dynamic parameter adjustment, etc., which solves the GNRON problem and the local minimum problem of the traditional APF algorithm.Secondly, a new APF-ACA algorithm is proposed with the improvements of the update rule of pheromones and the heuristic information function, in which the heuristic information is controlled by the distance between the ant and the goal as well as the artificial field. In the guidance of artificial field, the convergence rate of the algorithm is accelerated and the shortcoming of blindness of the traditional ACA in the early stage of path search is overcome.As to the improvements of the update rule of pheromones, the shortest and longest routes after each iteration will be awarded and penalized accordingly, combined with the maximum and minimum ant system, which can narrow the range of searching the optimal path and improve the convergence rate without emerging the phenomenon of "premature".Finally, the proposed algorithms are simulated in the platform of Matlab2010b, compared with the traditional AFP method and ACA. The experimental results show that the proposed algorithms have more smooth planned paths and faster convergence rate. As a conclusion, the proposed algorithms are effective and feasible.