Research On The Method Of Optimal Energy Path Planning Of Natural Energy-Driven USV

Natural energy-driven unmanned surface vessel is a new type of long-endurance surface vehicle.It can not only monitor marine environmental information in real time,but also capture natural energy such as solar energy and wind energy.Compared with conventional unmanned surface vessel,it has higher endurance.As a core issue in the research of natural energy-driven unmanned surface vessel,path planning directly affects the level of endurance.Therefore,it is of great significance to carry out research on the optimal energy path planning method of natural energy-driven unmanned surface vessel.This paper takes the "Wave Rider" as the research object.Considering the influence of ocean currents,wind and other marine environments on energy consumption and the natural energy such as wind and solar energy that can be captured during navigation,an energy-efficient optimal path planning method is proposed.The main research contents are as follows:The energy system of the "Wave Rider" was introduced to analyze the main sources of energy consumption and capture of natural energy-driven unmanned surface vessel from the perspective of hardware,and determined the optimization goal of the optimal energy path planning.The research on the optimal energy path planning method was divided into the research on the optimal energy global path planning method and the research on energy saving local path planning method.Then introduced the common global path planning algorithm and local path planning algorithm.Aiming at the energy consumption of natural energy-driven unmanned surface vessel,the influence of marine environmental factors such as ocean current and wind on the energy consumption was analyzed.Then an energy consumption model was established.Analyzed the influence of marine environmental factors such as wind,solar radiation and atmospheric clouds on the natural energy capture of natural energy-driven unmanned surface vessel,then a natural energy capture model was established.Aiming at the energy optimal global path planning problem of natural energy-driven unmanned surface vessel,a grid marine environment model of ocean current,wind,solar radiation and atmospheric clouds was established by using the grid method and combined with marine environment forecast data.The Dijkstra algorithm and Theta* algorithm are improved by combining the grid marine environment model with the energy consumption model and natural energy capture model of the natural energy-driven unmanned surface vessel to study the energy optimal global path planning algorithm of the natural energy-driven unmanned surface vessel in the static marine environment.Considering the problem of dynamic changes in the marine environment,the dimensionality of the grid marine environment model is extended and a 3D grid marine environment model is established by combining the marine environment forecast data.Under the three-dimensional grid marine environment model,the energy-optimal global path planning algorithm in static ocean environment is improved,and the energy optimal global path planning algorithm for natural energy-driven unmanned surface vessel in dynamic ocean environment is studied.The effectiveness of the algorithm is verified by conducting simulation comparison tests with the conventional global path planning algorithm under various working conditions.Aiming at the energy-saving local path planning problem of the natural energy-driven unmanned surface vessel,based on the Asynchronous Advantage Actor-Critic(A3C)algorithm,the navigation state of the natural energy-driven unmanned surface vessel was analyzed and the state space and action space was designed.The neural network structure of actor and critic was designed with Long Short Term Memory(LSTM).Using the energy consumption model and natural energy capture model to design the reward function,a energy-saving local path planning algorithm that can fully consider the marine environment information was propose.Through simulation experiments under various working conditions,the effectiveness of the algorithm was proved.