Research On Long Baseline Localization For AUV

Recent years, underwater acoustic sensor networks(UASNs) have been increasingly gaining concerns. As a core technology of UASNs, localization approaches of energy conserving and high accuracy for autonomous underwater vehicle(AUV) are desideratum. Sea bed transponder based long baseline(LBL) localization system is a proper option. It meets the goal of intelligence, energy efficiency, independency and high accuracy.Analysis of advantages and disadvantages of classical localization algorithms for AUV are given. Further, with reviewing recent researches of LBL system, this paper figures out critical problems that need solving. That is: time-synchronization is hard to achieve in underwater environment; round-trip ranging(RTR) technology causes large ranging error; sound speed uncertainty adds inaccuracy to distance estimation but LBL methods are sensitive to ranging precision. To address these, this work proposes time-synchronization free approach in shallow water(depth < 500m): self-localization of autonomous underwater vehicles with accurate sound travel time solution(SL-STTS). In SL-STTS, measurements of angle of arrival for acoustic signals and orientation for the AUV are mainly analyzed to get the one-way travel time. A rational method of using the mathematical expectation to approximate the sound speed is presented. Then we use the Levenberg-Marquardt algorithm(LMA) to optimize the distance estimates and localize the AUV. Simulation results show that the performance of our proposed algorithm meets the Cramér-Rao bound well. Additionally, under the measurement noise of time and angles, the root mean square error of SL-STTS is decreased by about 8% ~ 79% compared with the counterparts. And the average distance estimation error of SL-STTS is declined by 42% compared with the round trip ranging technology.In order to reduce the number of transponders used, a method called sequence-correlation localization algorithm based on signal screening(SCLSS) is proposed. SCLSS divides the location region into multiple separate virtual grid intersections by a grid-type measure to assist localization. Based on signal screening, SCLSS filters the task transponders and the corresponding task intersections which actually assist to localize the AUV. To improve location performance, according to the given parameters, it derives new intersections to isolate the centroid of the task derived intersections with the Pearson correlations as the weights. The isolated centroid is considered as the location where the AUV most probably resides. Simulation results show that our proposed methods have 49% higher location performance than the counterpart with fewer transponders. And under the effect of travel time measurement noise and angle measurement noise, our algorithm performs 35% better than the counterpart.