Underwater Vehicle Maneuvering Identification

Maneuverability of underwater vehicle has a close relationship with navigation safety.Maneuverability prediction is an important task at the underwater vehicle design stage to analyze and improve maneuverability.Computer simulation based on mathematical model is one of the most popular and effective methods. To apply this method, accurately determining the hydrodynamic derivatives and maneuverability indexes in the mathematical models is vital to improve the prediction accuracy. With the development of measurement technique and system identification technique, the method of system identification combined with maneuvering tests has achieved rapid development in terms of identification of hydrodynamic coefficients.In this thesis, a novel system identification method, Support Vector Machines (SVM),is applied in research of underwater vehicle system combined with maneuvering tests.Firstly, the correctness of SVM method is verified by using simulation method. Combined with the zigzag simulation test of an underwater vehicle, SVM is used to identify the hydrodynamic coeficients in mathematical models of underwater vehicle maneuvering motion. Comparing predictions of underwater vehicle maneuvering motion with results of simulation test, the correctness of SVM method is verified. Secondly, SVM method is applied to analyze the results of captive model tests. And oblique towing test and pure sway tests of an underwater vehicle are taken as examples. The predicted results at different test conditions are compared with the test results and the feasibility of SVM method in analyzing the results of captive model tests is preliminarily demonstrated. Finally, SVM method is applied to identify the hydrodynamic coefficients in the motion model combined with the zigzag test of an underwater vehicle. By using the identified model, zigzag maneuvers are predicted. The feasibility and validity of SVM method in identification of underwater vehicle system based on free-running model zig-zag tests are verified.