Real-time Path Planning System Based On PSO For Underwater Vehicles

With more and more attention on ocean resources nowadays, people are having higher expectation of the resource exploration and development completed by undersea vehicles, thus higher requirements of intelligence and autonomous navigation are needed. Path planning technique is one of its intelligent navigation control essential technologies, which has direct relations with undersea vehicle intelligence level. The paper focuses on the research of particle swarm optimization application in path planning problem, especially the obstacle avoidance method in dynamic obstacles environment. Using the characteristics of rapid solving rate, strong global optimization ability of PSO, the efficiency of the designed path planning method is enhanced. The research is developed according to the order of First theory then application, first static environment then dynamic environment, first 2D then 3D environment". The main research contents and achievements include:The application value of PSO in path planning problem is evaluated according to the following aspects: parameters setting, optimization principle, application performance and environment adaptability. Then the characteristics and improvement ways of PSO are summarized.In order to guarantee the good performance of PSO, the optimization principle with the application background is analyzed, and an improving PSO algorithm based on the equilibrium distribution parameter is proposed. According to the analysis of PSO disadvantages in micro-view of a particle composition, an equilibrium distribution parameter is introduced to guarantee the diversity in PSO evolutionary process. A self-exploration flying scheme is then proposed. Simulation results show the improvements of the proposed method.Global planning is the base of the underwater vehicle path planning task. A global path planning algorithm based on particle swarm optimization in polar coordination space is proposed in this paper. On the basis of polar coordination workspace modeling method, a cross-link data structure is designed which stores obstacle information according to its position in dimensional workspace and can realize the high-efficient storage and visit. Simulations are performed to show the stronger adaptability to environment, quicker convergence rate and higher solution precision of the proposed algorithm.Local path planning is discussed in known dynamic environments and un-known dynamic environments. An algorithm based on PSO to solve path planning problem in known dynamic environment and anther based on underwater vehicle watching window to solve the problem in un-known environment are proposed. In PSO path planning algorithm, obstacle avoidance conditions can be obtained through geometric model, and the path planning task can be converted to a multi-condition optimization problem. Then a suitable fitness evaluation function is established by comparing with the variable speed strategy and the variable heading strategy.Path planning in unknown environments is a difficult problem in the robot intelligent control. Motivated by the rolling window path planning algorithm, an AUV watching window local path planning algorithm is proposed which emphatically solves the obstacle-meeting problem in AUV navigation. The key steps of the algorithm include: environment modeling and prediction, environmental information survey and feedback by the watch window and local path planning in the window. The feasibility and the autonomic adaptability to un-known environment of the algorithm are verified and validated by simulations. The balance strategy between global path planning task and local path planning task is proposed. Based on this strategy, an AUV real-time path planning system is established which can achieve not only the global optimization but also the local path planning of real-time dynamic obstacle avoidance according to the sensor information. The reaching ability and safety of the system is analyzed, especially the algorithm performance when dynamic obstacle suddenly appears.Finally, with the 3-D seabed environment of AUV, a 3-D path planning algorithm based on PSO algorithm is proposed. The environment of 3-D seabed is described by the real depth data from the electronic chart, and the data structure for 3-D planning are defined. By using punish function and problem special heuristic knowledge, the designed algorithm can obtain optimal paths of different characteristics and is also adaptive to the obstacles bursting upon in 3-D environment. Simulations for the designed algorithms in 3D environment are performed. Lots of simulations of the designed algorithms show that the presented path planning algorithm makes full use of rapid convergence and good global searching capability of PSO, and can solve the path planning problem under static and dynamic, and 2D and 3D environments. It can be summarized that the algorithm can adapt to all kinds of complex and wicked environments, has better searching capability and rapid convergence, meets the requirements for autonomous navigation of underwater vehicle and has considerable value in safe navigation of underwater vehicles whether in theory research or in practice application.