Obstacle Avoidance Of Autonomous Underwater Vehicle

Since the 21st century, the research of ocean by human has reached a new period. With increasing reduction of earth resource, global warming and diversity development of naval force, more and more attentions have been paid to ocean, so the research of underwater vehicle has increased. Autonomous underwater vehicle's development is outstanding, which is used in oceanic science research, exploration and naval activity. For the complexity of sea bottom and the uncertainty of oceanic environment, the obstacle avoidance of AUV becomes a focus problem. How to design obstacle avoidance algorithm is my work's core.Firstly, classification, development and application of AUV are introduced, and path planning and real time control of obstacle avoidance are presented.Secondly, path planning is achieved by layered genetic algorithm. The environment model is set by the grid square of 16×16, whose length is 1. Two obstacle forms are used to simulate float and wreck respectively. The path from start point to end point is calculated for the two obstacle forms. Thirdly, an estate control equation of "REMUS" in two-dimension perpendicular plane is constructed, which contains sea flow factor. The estate variables include relative velocity in perpendicularity, angular velocity of pitch, pitch angle and depth. Lyapunov two quadratic form optimization feedback control algorithm is chosen and forward-look sonar simulation formation of sea bottom is designed. The real time control algorithm of obstacle avoidance contains real time obstacle avoidance algorithm and control algorithm, which are both designed in discrete time. In the real time obstacle avoidance algorithm, the depth input of next moment is calculated by forward-look sonar simulation formation of sea bottom in every discrete time step. If there is no obstacle, the depth input is set as above sea bottom 3 m. if there are obstacles, the depth input has been set along with obstacle height which assures the minimum distance between AUV and obstacle is 3 m. In control algorithm, the estate variables of the forward moment are used as the initial estate variables of next moment, and the optimization feedback control algorithm responds the depth input. So the estate variables of all simulation time will be acquired. According to three depth input forms, the simulations are done for rectangle and abnormity obstacles in different sea flow. The simulation results show that obstacle avoidance results are different for different depth input forms; the depth input is chosen suitably for abnormity obstacles, and the obstacle avoidance result is satisfying.Last, path planning and real time control for obstacle avoidance are integrated. If there is no new obstacle or the height of new obstacle is less than the known obstacle in path planning, AUV will track the planning path, that is the path planning depth is used as depth input, and the optimization feedback control algorithm will responds the depth input. If there is a new obstacle and its height is more than the known obstacle in path planning, the real time control algorithm of obstacle avoidance will starts. After the obstacle avoidance, AUV still start to tract the planning path. From four simulations, we can infer that the safety of AUV can be assured by the algorithm integrated by path planning and real time control for obstacle avoidance in the condition of uncertain environment.