Research On A Dead Reckoning Method Aided By Dynamic Model Of AUV

As the application of Autonomous Underwater Vehicle (AUV) in the sea development becomes more and more wide, the requirement for the precision of the navigation system of AUVs becomes more and more high. The precision of navigation system of AUVs is one of the important factors that depends whether AUV can complete mission and return safely. At present, one of the most widely-used navigation methods for AUV is the traditional dead reckoning method, which means that the current position is reckoned from the initial position step by step using the data gained from the velocity sensor as Doppler Velocity Log (DVL) and the attitude sensor. Therefore, traditional dead reckoning method relies highly on velocity sensor as Doppler Velocity Log (DVL) and attitude sensor. Situations will occur where DVL fails to work or measurements of DVL are discarded due to decreased quality when AUV works underwater. For example, when the height of AUV is too low that it has been out of the measurement range of DVL, or when the sea bottom is silt that has strong wave absorbing ability, we can not gain effective data from DVL, which will result in increasing error for dead reckoning. Therefore, it is important to find a kind of method that can effectively improve the precision of AUV navigation system and decrese the time for AUV to be proofreaded on surface. When the navigation system of AUV is in the absence of DVL measurements, the navigation error of traditional dead reckoning method will increase highly. To solve this problem, the thesis reports a new dead reckoning method aided by dynamic model of AUV, which is called MADR (Model-Aided Dead Reckoning) for short. The research object of the thesis is one AUV which is independently developed by Harbin Engineering University.According to the demand of the research, the main tasks of the thesis include several aspects, which are shown below.Firstly, the thesis needs to establish the dynamic model of AUV, which is expressed as kinematics equations in six degrees of freedom (6 DOF), and validate whether the dynamical model is correct.Secondly, the thesis needs to expound the theory and realization of the dead reckoning method aided by dynamic model of AUV based on the theory and realization of the traditional dead reckoning method, as well as the disposal methods for sensor data of the navigation system of AUV.Thirdly, the thesis needs to design a software system to realize the dead reckoning method aided by dynamic model of AUV base on QNX operating system, which includes the design of the entire software system as well as the designs of every part of the entire system.Lastly, the performance of the method of the dead reckoning method aided by dynamic model of AUV need to be evaluated using real data collected by AUV in one of the trials in the lake. What is more, the experimental results of the dead reckoning method aided by dynamic model of AUV are also given, which should be compared with the experimental results of the traditional dead reckoning method.The research on the dead reckoning method aided by dynamic model of AUV of the thesis indicates that during velocity sensor as DVL dropouts or failures, the dynamical model of AUV can efficiently aid the dead reckoning system of AUV instead of velocity sensor as DVL, which shows promise to improve the robustness and redundancy of the underwater navigation system of AUVs and decrese the time for AUV to be proofreaded on surface. The research can satisfy the demand of AUV, which has character in long work-time and high concealment. The dead reckoning method aided by dynamic model of AUV has broad application foreground, which can be applied to the systems that are absent in velocity sensor, as well as systems where redundancy and integrity is important.