Research Of Autonomous Underwater Vehicle Local Path Planning Methods

With the rising status of China's ocean, autonomous underwater vehicle is also playing an increasingly important role. It plays an irreplaceable role in the development of marine resources, gathering military intelligence. Due to the particularity and complexity,autonomous navigation capabilities work environment is the key to its intelligent nature.Therefore, the path planning technology development largely demonstrated its degree of intelligence.Due to the complex nature of the work environment, autonomous underwater vehicle is difficult to obtain all the environmental information before planning. Therefore local path planning methods are more suitable for autonomous underwater vehicle. Based on the analysis of traditional artificial potential field method and the rapid expansion of the basic foundation of the random tree algorithm deeper research conducted its shortcomings.First, the actual situation may be encountered in moving obstacle, moving obstacle modeling discussed in detail, including the uniform motion and random motion of these two types, and to predict the moving obstacle against analyzed; then environmental modeling three-dimensional space, a detailed description, including access to environmental data and their processing.Secondly, its deficiencies exist in the traditional two improved methods. The first is the case when an obstacle and end in close proximity can not reach the finish line; another underwater vehicle suffered as a repulsive gravitational equal and opposite direction under the circumstances resulting in staggnant underwater vehicle before. For both cases, the first modification of the adjustment factor increased repulsion function to solve this problem;second is by looking for potential field with minimal to avoid deadlock.Then, in the rapid expansion of the basic principles and the basis of a random tree algorithm steps described, points out some shortcomings of the algorithm exist. Extended way too fragmented for this defect of the algorithm has been modified , a fast random tree algorithm is biased target , the algorithm has a bias in the search path when , and from the principle of the algorithm is analyzed in this bias a characteristic, and basic and improved algorithm for the simulation analysis . Based on the analysis of dynamic obstacle avoidance strategy, autonomous underwater vehicle use improved algorithm for path planning in dynamic environments and simulation.Finally, the two algorithms are mixed in a three-dimensional path planning. For real workspace autonomous underwater vehicle, and elaborated representation and real-time updates to detect the depth of the data obtained, and the distance of obstacles and get through the underwater vehicle sonar to obtain a three-dimensional environment can exercise free space. In the analysis of the artificial potential field method and basis of the rapid expansion of the advantages and disadvantages of the random tree algorithm , the algorithm proposed the idea of a mixed use of hybrid algorithm partially completed underwater vehicle path planning problem in three-dimensional environment,the method was confirmed by simulation effectiveness.