ѧ
JOURNAL OF ASTRONAUTICS
1999 20 3 vol.20 No.3 1999



Ӧ˲ڵ
ƵеӦ*
ܡݶĽ馡
ժҪ˲Kalman˲Gauss-Newton㷨ɣһԶ̬ģͺͷԲģɵķϵͳĽ˲ʱͳƹֵϣõһӦ˲ڲͳ޷ǰ֪£˲Ȼá󣬱ĽӦ˲ӦڵıƵ⣬ַ֤㷨Чԡ
ʡ˲Ӧ˲Ƶ
ADAPTIVE TWO-STEP FILTER 
WITH APPLICATIONS TO MISSILE 
GUIDANCE WITH PASSIVE MEASUREMENTS
Zhou Di Mu Chundi
(Department of Automation,Tsinghua University.Beijing.100084)
Hu Hengzhang
(Department of Control Engineeringm,Harbin Institute of Technology.Harbin.150001)
AbstractThe two-step filter suits a class of nonlinear systems which include a linear dynamic model and a nonlinear measurement model.This filter consists of a Kalman filter and a Gauss-Newton iterative algorithm.We integrate the two-step filter with a time-varying measurement noise statistical estimator to obtain an adaptive two-step filter,which still performs well in the case where the statistical properties of the measurement noise are unknown apriori.The adaptive two-step filter is applied to missile guidance with passive measurements and the numerical results show that this filter is really effective.
Key wordsTwo-step filter Adaptive filter Missile guidance with passive measurements
1
ϵͳ״̬һҪо⡣õĵͷ˲ƹKalman˲(EKF)Ӧô˲ҪԷϵͳԻЩ£EKFнϴĹᷢɢĽһһضϵͳ˲12ݡһضķϵͳָһԵĶ̬ģͺһԵĲģɵϵͳν˲ָȲһµ״̬(ϵͳʵ״̬ķԺ)ʹϵͳ״̬Ժڵһ˲УӦKalman˲õ״̬Źֵڵڶ˲У״̬ĹֵΪӦGauss-Newton㷨ϵͳʵ״̬Ž⡣ʵ£ͳʱģ޷ǰ֪ʱӦͳƹֵߵȷľֵͷʱͳƹֵ˲ϣԵõһӦ˲
2Ӧ˲
Ƿϵͳ
(1)
(2)
F(..)һwkvkΪ˹Ewk=0,Ewk wTk=Qk,Evk=rk,Evk vTk=Rk,Ewk vTk=0
ָ꺯Ϊ
(3)
һµ״̬ykzkΪykԺ
(4)
yk=fk(xk,tk)һȡyk=xTkFT(xk,tk)Tȡһ˲ʱָ꺯Ϊ
(5)
ΪʹJyﵽСֵKalman˲ykĹֵеһ˲
(6)
(7)
,
ʽ(6)еrkRk޷ǰ֪ʱʱһ״̬⣬Կʱͳƹֵ3ݶrkRkʵʱƣ
(8)
(9)
dk=(1-b)/(1-bk+1),0b1bӣ,kϢKk=PykHTkR-1k㷨˵һ˲Ĳܡ
棬ǰѵһ˲ĹֵykΪڶ˲ĲֵΪ
(10)
вekЭΪPykȡڶ˲ʱָ꺯Ϊ
(11)
ȻxkʹJkﵽСֵһһԹ⣬Gauss-NewtonŽ⣬
(12)
ifĶ׵͸߽׵Ա֤Lk,iԣ
(13)
Ӧ(13)ʽ(12)ʽΪ
(14)
Сʱ̽
״̬xkĹЭΪ֤1
(15)
(14)ʽɵڶ˲Ĳܡ
Ȼϵͳ״̬xkһԤֵԤЭ㷨ã
(16)
(17)
(16)ʽ(17)ʽ˵ڶ˲ʱܡ
Ƶ״̬ykһԤֵԤЭļ㹫ʽ

(18)
ɽ
(19)
Э
(20)
Ϊ,öͣ

(21)
(19)ʽ(21)ʽ˵һ˲ʱܡ
ϸƵܽᣬɵ½ۣ(6)(9)ʽ(14)ʽ(16)ʽ(17)ʽ(19)ʽ(21)ʽ˳ִ㣬ԹӦ˲
3Ӧ˲ڵƵеӦ
ĿٶȿһGauss-MarkovôֱϵĿ͵˶ϵɢʽĶ̬ף6ݡƵ£ԵĲҲף6ǰӦ˲Ӧһ⣬֤ͨ˲Чԡ
ʵĩƵʼʱ̣Ŀꡪ֮λΪrx0=15km,ry0=8km,rz0=6kmٶΪvx0=-1100m/s,vy0=-150m/s,vz0=-50m/s,ĿٶΪaTy0=aTz0=10m/s2;˲ĳֵΪ=14km,=7.5km,=5.5km,=-950m/s,=-100m/s,=-100m/s,=0m/s2,ȡyk=xTkFT(xk,tk)TMy0=diag105I3104I3102I5ݣ=10-8I2,b=0.98ף6зͬƵָԶ͵ɣ4ȷϵͳ״̬ʵֵʵ֡
100Monte-Carlo棬õӦ˲(ATSF)Ծ롢ٶȼĿٶȵĹֱͼ1ͼ3ͼ1ͼ3˵ͬķ£һ˲ƹKalman˲(EKF)Ҳף6ӦƹKalman˲(AEKF)

ͼ1Ծ

ͼ2ٶȹ

ͼ3Ŀٶȹ

4
һӦ˲һԶ̬ģͺͷԲģɵķϵͳϵͳͳԿǰδ֪ġĽӦ˲ӦڵƵ⣬ַ˲ܡ
*йʿĿ
ߵλܡݶĽ馣廪ѧԶϵ..100084
£ҵѧƹϵ..150001
ο
1Haupt GT,Kasdin NJ,Keiser GM,and Parkinson BW.Optimal Recursive Iterative Algorithm for Discrete Nonlinear Least-Squares Estimation.Journal of Guidance,Control,and Dynamics,1996,19(3):643-649
2Kasdin NJ,and Haupt GT.Second Order Correction and Numerical Considerations for the Two-Step Optimal Estimator.Journal of Guidance,Control,and Dynamics,1997,20(2):362-369
3һ.ִʱзģ˲ȥԤͿ.֪ʶ磬19896
4Hull DG,Speyer JL,and Burris DB.Linear Quadratic Guidance Law for Dual Control of Homing Missiles.Journal of Guidance,Control,and Dynamics,1990,13(1):137-144
5Song TL,Speyer JL.A Stochastic Analysis of Modified Gain Extended Kalman Filter with Applications to Estimation with Bearings Only Measurements.IEEE Transactions on Automatic Control,1985,AC-30(10):940-949
6ݶ.ӦƹKALMAN˲ӦڵıƵ.ѧ199718(4)31-36
ո:19971021,޻:1998715
