Research On Prior Map Data Processing Based Terrain Aided Navigation Methods For Underwater Vehicles

Most underwater vehicles navigation systems are based on accurate low drift inertial system (INS) designed to sustain submerged operation for long period of time. Regardless of velocity aiding sensor and INS quality, the position accuracy decrease necessitating position updates. When GPS surface fixes are not allowed due to efficiency and covertness, terrain aided navigation (TAN) is a well suited method for submerged position updates. In this method, distance to sea bottom get by bathymetric sonar is added with the depth from the sea level given by pressure sensor to get the water depth, which is matched with prior terrain map to estimate position of the vehicle, to correct INS positional error.Terrain Elevation Matching (TEM) strategy is one of the key techniques in TAN, and up to now, many TEM have been developed for aircraft on land, well there is rare special TEM for underwater vehicles. Compared to TAN for aircrafts, there are some particularities of TAN for underwater vehicle, therefore, studying probable TEM algorithm with high applicability for underwater vehicles has become an important research topic, which is concerned in this paper.Traditional terrain linearization methods cause abvious linearization error, in this paper interpolation technique, a solution for sovling partial differential equations (PDEs), is used to take place linearization method. Nature neighbor interpolation (NNI) has good capability however the computation time is long, so fast nature neighbor interpolation (FNNI) is proposed. In two aspects, searching nature neighbor and computing shape function, the traditional method is improved. Results from experiment gives out that FNNI has good robot, high precision, and time complexity is decreased, the desire of real time computation is fulfilled.This paper focuses on multi beam echo sounder (MBE) which has particularly attractive advantages to single beam echo sounder (SBE), that are described in this paper, MBE is used as bathymetric sensor in terrain aid navigation for underwater vehicles. TEM problem is essentially a nonlinear state estimation problem with nonlinear state function and measure function, proposed. Nonlinear estimation method with probability criterion can directly deal with TEM model, and is able to avoid errors introduced by linearization of nonlinear model. Theoretically, this method shows superiority in solving TEM problems. Long time continual filter method is not agree with the case of underwater environment and vehicles characters, then maximum likelihood estimation (MLE) becomes a suitable underwater positioning method. Subsequently, the derived formulae and simulations both prove MLE's validity.