Research On Underwater Aided Navigation Base On The Full Tensor Gravity Gradient Matching Method

Because underwater vehicle needs to be a long time under the water for safety,positioning is very important. Passive navigation theories and methods of use of theEarth’s physical features to aid inertial navigation is a hot concerned domestic in foreignresearch institutions and academic research, the gravity and gravity gradient data isstability, anti-interference, high reliability advantages, and gravity gradient device doesnot need to transmit to the outside or receive information, therefore gravity gradientautonomous navigation is a completely passive way.Three key technologies of underwater gravity gradient aided navigation systemsinclude the gradiometer, preparation of the gravity gradient map and effective matchingalgorithm. In this paper, take advantage of five separate component information of thefull tensor gravity gradient, full tensor gravity gradient underwater matching aidednavigation methods are researched.Firstly, in order to obtain the gravity gradient data, the gravity gradient mapspreparation techniques were studied, and the factors affecting the gravity gradientreference maps prepared were analyzed, Based on terrain elevation forward method isprepared, which can meet the needs of navigation accuracy. the gravity gradientreference map, feature extraction and analysis of the gravity gradient were applied for thefive independent components, which were based on a single component of the gravitygradient matching algorithm;Then, a full tensor gravity gradient weighted fusion matching methods simulationhave been done, applying MAD algorithm, experiments results show the matching effectunder different matching measurement noise, take advantage of the full tensor gravitygradient, weighted fusion matching can improve the positioning accuracy;Finally, a full tensor gravity gradient matching method is proposed which based onthe similarity measure, utilizing the full tensor gravity gradient measurements and theestimated value, the cosine similarity measure to the full tensor gravity gradient are fusedeffectively, simulation have been down, the full tensor weighted fusion matching methodand the full tensor gravity gradient fusion matching were compared.