Abstract:

Unmanned undersea vehicle (UUV) is a tool widely used in underwater detection. It is the crucial factor for UUV to work well in deep underwater environment. With a view to the problems of the capability of robot to percept the underwater situation compared with fish, an model that can identify the flow characteristic of the water is presented. The control of robot must consume large amounts of energy to maintain the position and status traditionally, which is passive and dangerous for UUV to cope with the challenge in the water. Therefore, it is very important for underwater vehicles to recognize current and explore ocean further. Studies on the lateral line system show that it can distinguish the flow of water and improve the motion of the fish. Aimed at a torpedo UUV, the mechanism and capability of lateral line system is studied and realized using the pressure data of vehicle calculated by CFD software. The linear discriminant analysis and support vector machine methods are selected to establish the flow sensing classification model. The tests demonstrate that the method can identify different flows around the vehicle effectively and can provide a new view for the study of underwater vehicle during the development and use of ocean.

Key words: flow perception, lateral line, linear discriminant analysis, support vector machine, unmanned undersea vehicle