Unmanned Vehicle Transfer Alignment Of Strapdown Inertial Navigation System

With the improvement of science and technology, Unmanned Vehicle that based onadvanced technology and new scientific theory is more and more widely used in militaryand civil fields. Transportation, search and rescue, investigation, explosive handle andmany other functions of Unmanned Vehicle are need accurate navigation to realize, thusresearch on Unmanned Vehicle navigation technology become important subject. Theinitial alignment accuracy seriously interfere with the navigation accuracy of UnmannedVehicle strapdown inertial navigation system, so the initial alignment technology as thecore technology in the field of Unmanned Vehicle is necessary to conduct in-depthresearch.The prime condition of ensuring strapdown inertial navigation system normaloperation is the initial alignment, the accuracy and time of initial alignment impact onsystem work ability and fast reaction ability. Compared with the traditional self-aligned,transfer alignment needs less time and requirements of inertial sensors are lower so that itbecome research hot topic at home and abroad. Therefore, the subject of UnmannedVehicle strapdown inertial navigation system transfer alignment technology has greatpractical significance and broad application prospects.This paper researched the error model of Unmanned Vehicle strapdown inertialnavigation system transfer alignment and inertial device, compared different matchingmodel performance thus selected velocity matching mode to research transfer alignmenttechnology. Considering Unmanned Vehicle use different navigation system for differentneeds, respectively make study on pure strapdown inertial navigation system transferalignment and GNSS/MIMU integrated navigation system transfer alignment method.Through carrying on the analysis of the characteristics of Unmanned Vehicle strapdowninertial navigation system and advantages and disadvantages of traditional Kalman filtermethod, put forward that apply the least squares support vector machine method to thesystem. Use the traditional Kalman filter input and output variables as the samples totraining the LS-SVM. Under the same conditions, compare with the conventional Kalmanfilter simulation, for the introduction of LS-SVM method can shorten the transferalignment time, and increase the system alignment accuracy appropriately. Finally thealgorithm is transplanted into GNSS and low cost of MIMU integrated navigation system, to carry on the actual car test, the results show that the alignment accuracy basic achievedin practical application and accuracy requirement.