第42卷增刊1 中南大学学报(自然科学版) Vol.42 Suppl.1
2011年9月 JournalofCentralSouthUniversity(ScienceandTechnology) Sep.2011
SquarerootunscentedKalmanfilterforMEMSgyroscope random
driftcompensation
ZHOUJianjun(周建军),SONGChunlei(宋春雷),ZHUANGHuihui(庄会慧)
(SchoolofAutomation,BeijingInstituteofTechnology,Beijing100081,China)
Abstract:Themicroelectromechanicalsystem(MEMS)gyroscopesarewidelyusedinmanyapplicationsforitssmallsizeandlow
cost.Theyusuallyhavedeterministicsystemerrorandrandom errorwhichcanonlybedescribedbystatisticalmodels.Firstofall,its
random drifterrormodeliscreatedbyusing thetimeserialanalysis,andthentheprocessofdecreasingthis randomdrifterrorby
makinguseofthesquarerootunscentedKalmanfilter (SRUKF)basedontheaboveerrormodelisexpounded.Thecompensating
results for the practical testing data of a MEMS gyroscope show thattherandom drift error can be controlled effectively by the
filteringmethodpresented,anditsapplicationprecisioninpracticalsystemcanbefurtherproved.
Key words: MEMSgyroscope;SRUKF;randomdrifterror;timeseriesanalysis
CLCnumber: V241.5 Documentcode:A ArticleID: 1672−7207(2011)S1−0469−04
1Introduction
Recently, with the development of the micro
electronicstechnology,MEMSinertialsensorshavebeen
improved greatly. Compared with conventional
gyroscopes, MEMS gyroscope has a lot of good
characteristics such as low cost, small size, light mass,
mass production, long life, high reliability, low power
consumptionandso on,whichmakesensorsarewidely
used in such fields as navigation, defense, medical
instruments and consumer applications. The MEMS
gyroscopes have a big potential to be used in future
micro inertial system. However, the low precision
property has seriously prevented its application in high
accurateinertialsystems.Howtoimproveitsapplication
precision is not only a hot point which is attended by
many experts, but also an application problem of
practicality.Wecaneffectivelyimprovetheaccuracyof
theMEMSgyro by analyzing thegyroerror signal,and
then modeling and compensation. They usually have
deterministicsystemerrorwhichcanbecompensatedby
testing and demarcating and random error which can
onlybedescribedbystatisticalmodels.
Inthispaper,theMEMSgyrorandomerrormodel
was created by using the timeserialanalysis. However,
most time series of practical relevance are complicated
thatmakesconventionalpredictionmethodsinapplicable.
In 1927, the mathematician Yule proposed
autoregressive(AR) modelto forecast time series.After
several years, in 1931, another mathematician Walker
invented moving average (MA) model and
autoregressive moving average (ARMA) based on
autoregressive(AR)model
[1]
.Andinthisstudy,ARMA
(2,1)modelwasusedforparameterestimation.
Manymethodshavebeendevelopedtodecreasethis
MEMS gyro randomdrift error.These methods include
Kalmanfilter,robustfiltering,nonlinearfilteringandthe
enhancements and improvements. Of course, these
methods can effectively solve the gyro’s random drift
error compensation; especially have good work on the
drifterrormodel.Kalmanfiltercanbeusedtodealwith
the aboverandom errormodel, but it just isa filter for
linear systems with no consideration of environmental
noise. So in the case of serious nonlinear, its model
precision will be affected by divergence of filtering. In
1997, Julier and Uhlman proposed a new type of
nonlinearfilter,namedunscentedKalmanfilter(UKF).It
is based on the nonlinear Unscented Transformation
(UT). Thereby, UKF does not use the linearization like
kalman filter. UKF produces a set of selecting point
(called sigmapoint) in such a way that they together
capture the full mean and covariance of the state and
maketheestimationbasedonthesepoints.Itprovidesan
excellent solution for nonlinear system, but it is
susceptibletopossibledivergenceofcovariancematrices
due to accumulative runoff errors in the transition
process. Sothesquareroot UKF(SRUKF)isproposed,
it’s the square root filtering form of UKF. Square root
UKF can not only guarantee the positive definite of
covariance matrix, but also increase the accuracy of
numericalcalculation
[2]
.
First, this article expounds the square root
unscentedtransformation;thesquarerootUKFprinciple
and recursion steps. Based on these works, the square
root UKF was designed. According to the practical
testingdata,thecompensatingresultscanbeobtainedby
makinguseofthesquarerootUKF,therandomdrift
Receiveddate:2011−04−15;Accepteddate:2011−06−15
Correspondingauthor:SONGChunlei,PhD,AssociateProfessor;Tel:+8613520186893;Email:songchunlei@bit.edu.cn