With the development of ocean source exploitation and ocean space utilization, UUV (Unmanned Underwater Vehicle) has been paid more and more attention by engineers from all over the world. UUV must come back to the underwater workstation when assignments have been done, or it needs to download missions, transform data and charge up batteries. Automatic underwater recovery is used widely for the requirements of covert, safety and low cost. In this dissertation, flat-carrying recovery is adopted, that is a special location for UUV recovery is set on the back of the underwater workstation, where SBL (Short Base Line) sonar, guiding lights and fixed compartments are installed on. UUV comes down to the recovery location automatically according to the controllers and decision-making programs, which are decided in advance, and then the vehicle moves with the underwater workstation after being fixed to the flat. Meanwhile, mission download, data transforming and charging up are finished.During the recovery, UUV comes down to the flat from infinite broad environment to near-wall region, that is, the characteristics of hydrodynamic change greatly, and the rule is so complex that it is not persuasive for the controller capability if simulating with the UUV motion model under infinite broad region. Considering no mature theory exists, "Fluent", a CFD (computational fluid dynamics) software, is used for numerical simulation calculating the hydrodynamic forces at different height in UUV recovery. Then the data coming from computing by Fluent is analyzed regressively, and the interference calculation formulas are fitted after the rules of the hydrodynamic are summarized. Finally, the 6-DOF model of UUV flat-carrying recovery is built based on the motion characteristics.Considering the characteristics of nonlinear, weak-observation and strong couple, the automatic control of UUV is very difficult, so the traditional control method can not satisfy the automation performance under the complex environment, especially the unknown near-wall region. To deal with these problems, adaptive controller based on DOB (Nonlinear Disturbance Observer) is presented as an inner loop for UUV recovery. DOB not only can compensate the unknown inferences, but also eliminate the high frequency noise from sonar by the low pass filter in DOB.At the end of this dissertation, some simulation schemes are simulated with the near wall UUV motion model. By comparing the simulation results with adaptive controller without DOB and PID controller, the performances of adaptive controller based on DOB for UUV recovery, such as robust, self-tuning, and high-frequency noise eliminating, are verified. |