Professional Documents
Culture Documents
The Higher Order Unscented Filter
The Higher Order Unscented Filter
The Higher Order Unscented Filter
2n
2 The Unscented Filter
i=O
Consider a nonlinear transformation of the random
variable X with mean 2 and variance P, where the weights remain the same as for matching the
input moments, equations 7 and 8.
where we would like to approximate the statistics of the 3 The Higher Order Unscented Filter
transformed random variable Y , for example the mean
and variance. Note, that this paper improves on the It is a well known fact that the Unscented Filter de-
approximation of the statistics of the aformentioned scribed by equations 12, 13 exhibits an accuracy in es-
transformation, which can be seen as the prediction timating the statistical characteristics up to the second
algorithm of the state estimation. moment. This has been achieved by selecting a a-set
resembling the input statistics. By utilizing the Taylor
+
The unscented filter selects a o-set consisting of 2n 1
series expansion it can be shown that the transformed
points, which are perturbations from the mean by a
statistics match the covariance up to the third order,
scaled deviation. The deviations oi are defined as the
i.e. errors are introduced at the fourth and higher or-
columns of the matrix square root of P, [2].
ders [2].
o= (2) Since the higher order terms of the Taylor series con-
sist of a combination of higher order moments, it is es-
The 0-set is defined as: sential to correctly estimate these moments to achieve
an increase in the performance of the estimator. Ap-
pendix A shows the derivation of the Taylor series ex-
pansion. The following section introduces an algorithm
matching higher order moments of the input.The per-
such that the a-set exhibits the same probabilistic char- formance increase is illustrated on a simple example.
acteristics as the random variable X and K. is a free
variable of the unscented filter. The weighting scheme The higher order unscented filter (HOUF) remains the
same structure as the original unscented filter. The
HOUF consists of an augmented o-set with separate
weights. The new o-set can be constructed as follows:
CO = x (14)
<i = CO + oi (15)
for oi = &GI,3 ~ 2 , . .. ,*om ,
where m = n f q is the augmented dimension and q is
the number of additional U points. The statistical prop-
erties of the a-set are defined by the weighted sums:
2m
has been selected to match the first four central mo-
ments of X, i.e. f = Cwi<i (16)
i=O
- 2m
Px=Pc
2=(
p.”,= p: = 0 p: = pt
(9)
(10)
p; = CW&
i=O
-f)j (17)
which is true for all Gaussian random variables if K. is where and p i represent the mean and the jth central
+
selected to satisfy the constraint, n K = 3. moment.
The o points are transformed by the nonlinear relation- 3.1 2-0 HOUF
ship The statistics of the o-set are selected to match the
vi = g [ < i , t ] for i = 1 . .. 2 n (11) statistics of the input X.This procedure is exemplified
Proceedings of the American Control Conference
2442 Denver, Colorado June 4-6, 2003
for the HOUF with two cr's, (TI and 172. Expanding where X is a zero mean Gaussian random variable with
equations 16 and 17 up to the eighth central moment, variance P,. It is a well known fact that the trans-
yields the following relationships: formed random variable Y is non-Gaussian and its odd-
order moments are non-zero. A Monte Carlo (MC)
1 = WO + 2Wl+ 2w2 simulation with 25 million samples has been performed
P, = 2(" + IE,(WlU1" + w 2 a 3 and the resulting moments are shown in Table 1 along
CL: = 0 with the estimates of the Higher Order Unscented Fil-
CL2 = 2(m + IE)2(w,cril + 21124)
ter and the standard Unscented Filter (UF). The pa-
rameters for the HOUF are derived from equation 21
CL: = 0 and listed in Table 2. We arrive at two solutions, which
PI6 = 2(m + +
I E ) ~ ( w ~ (w
T2~4 ) in turn are identical since the indices are swapped, i.e.
7 (TI becomes n2 and vice versa, etc. Observing Table 2,
PI = 0
Pz
8
= 2(m + I E ) ~ ( w ~ o+: w 2 4 ) ,
where pk is the ith central moment of X. With the
C1 g2 WO w1 w2 K.
knowledge of the input moments we are able to solve
1.6495 0.7827 0.5333 0.0113 0.2221 1
for the cr points oi,the weights wi and IE. The general
0.7827 1.6495 0.5333 0.2221 0.0113 1
solution becomes rather difficult to present and one has
to decide on a case by case basis about possible simpli-
fications. The following transformation of the weights the even order moments of the transformed random
and the requirement of the standard unscented filter variable 2 remain the same as the input X , whereas
m + IE = 3 lead t o an amenable solution in the partic- the odd order moments change significantly. The UF
ular case of Gaussian moments. matches the MC's even order moments up to the fourth
order, while the odd order moments differ. The HOUF
w; = 2(m + IE)Wl w; = 2(m + K)W2 (19) closely approximates the even and odd order moments
of the MC simulation up to the eights order moment.
The equations 18 can be rewritten as:
For greater clarity, we expand the above example to a
problem with known analytical solution. Let the ran-
dom variable Y describe the distance of a particle above
a barrier (121. The particle at the initial position zois
traveling with a velocity of W O , then the distance from
the barrier is given as:
3.2 3-U H O U F
The performance of the HOUF depends on the num-
ber of auxiliary a’s, where a larger a-set should better
approximate the true solution. The expansion to 3 CT’S
enables the HOUF to match up to the twelfth order
moment. Solving an expanded set of equations simi-
lar to equations 20, yields the parameters, which are
shown in Appendix B. Figure 3 compares the true so-
lution with the HOUF for 2 and 3 a’s for the particle
above a barrier example. It can be seen that 3-a HOUF
I
2 1 0
“0‘
1 2 3 approaches the true solution.
and the weights wi. The optimal solution is obtained The unscented filter by Uhlman and Julier [2] matches
by constructing a cost function consisting of the square up to the fifth order moment of the input variable (as-
error of the mean and variance estimate. suming a Gaussian distribution), whereas the trans-
formed covariance introduces errors at the fourth and
) (g - f j ) 2
J(ai,~i= + (P, - PV)’ (27) higher orders of a Taylor series expansion. The higher
order unscented filter is capable of matching any desir-
The cost function J ( a i ,wi) can be minimized for the able input moment by selecting a modified a-set. This
unscented filter (equations 12, 13) with respect to the results in a higher order approximation of the trans-
design variables 01 and w1, where W O = --&and formed covariance. This has been shown on an example
+
n K = 3 has been assumed. The numerically de- with variable degree of nonlinearity, i.e. the distance
termined optimal al and w1 are shown in Figure l for of a particle above a barrier.
a time varying mean vot within the interval [-3 31.
The “jumps”, for example at vot = 1.3, result from the The improved approximation of the transformed mean
selected nonlinear function, the absolute value, which and covariance can be utilized in dynamic estimations
is part of the weighted sequence in equations 12 and such as Kalman filtering. With the assumption of
13. Furthermore, the suggested a-set and weights by Gaussian distribution of the state throughout its prop-
Uhlman and Julier [2] are illustrated by the dashed agation the HOUF can approximate the mean and co-
lines. The optimal solution approaches Uhlman’s a-set variance with adjustable accuracy. On the contrary,
as the input mean increases and the influence of the a larger a-set increases the computational load, which
nonlinearity vanishes. leads to a trade-off depending on the nonlinear trans-
formation. As we have seen from the aforementioned
Figure 2 shows the analytical mean and deviation, i.e. example, after one transformation the Gaussian as-
the square root of the variance, for a time varying mean sumption does not hold and odd order central mo-
of the input X . The optimal solution (Figure 1) closely ments are generated. Based on the characteristics of
matches the analytical solution. Both unscented filters the nonlinear function it might be desirable to track
reasonably approximate the mean of the particles dis- higher order moments, which can be approximated by
tance, whereas the HOUF better describes its devia- the HOUF after slight modification of the matching
Proceedings of the American Control Conference
2444 Denver, Colorado June 4-6, 2003
1
c
m
09
f 08
5
$07
06
-3 -2 1 0 I 2 3
'0'
(a) Mean distance above the barrier [3] N.J. Gordon, D.J. Salmond, and A.F.M. Smith.
Novel approach to nonlinearlnon-gaussian bayesian
state estimation. IEE Proc. on Radar & Signal Pro-
cessing, 140(2):107-113, Apr. 1993.
(b) Variance of the distance above the barrier [7] A. Doucet, N. de F'reitas, and N. Gordon. Se-
quential Monte Carlo Methods in Practice. New York:
Figure 2: Deviation of the distance of a particle above a Springer-Verlag, Jan. 2001.
barrier
[8] Rudolph van der Merwe, Arnaud Doucet, Nando
equations. de Freitas, and Eric Wan. The unscented particle filter.
Technical Report CUED/F-INFENG/TR 380, Cam-
The HOUF derived in this paper currently applies to bridge University Engineering Department, Aug. 2000.
one dimensional systems, but its algorithm is amenable
to address multi-dimensional extensions. [9] Sanjeev Arulampalam,.Simon Maskell, Neil Gor-
don, and Tim Clapp. A tutorial on particle filters for
online nonlinearlnon-gaussian bayesian tracking. IEEE
Trans. on Signal Processing, 50(2):174-188, Feb. 2002.
Acknowledgments
[lo] S.Challa and Y . Bar-Shalom. Nonlinear filter de-
This research has been supported by the Center sign using Fokker-Planck-Kolmogorov probability den-
for Multisource Information h s i o n (CMIF) and the sity evolutions. IEEE Trans. on Aerospace and Electr.
Calspan/UB Research Center (CUBRC). Systems, 36(1):309-315, Jan. 2000.
+ ... (28)
The mean and variance of the transformed random
variable can be determined by taking expectations of
the Taylor series
E{Yl
1 d2g 1 d3g
g(X) + -2!dx2
-P, + -3!dx
7 E { A X 3 } + ...
(29)
E { (Y - V,”}
[(g)2--
(
i
”
,
’
.
] 2 ! dx2 Px+
d3
3! dx2 dx3 (30)
39”. +
- 6 3 P , ~ ; 105P:~:- 35P: = 0 (31)
3 ( ~ ~ + ( 3 ~ ~ - 2 1 P , ) ( ~ ~ - 2 1 P ~ ( ~=~0+(32)
3~~+35P~
cl; + +
U; 03” - 7Px = 0 (33)
The closed form solution of the weights is:
w1 =
(126P: +1 8 ~ : 72a$PX- 5 4 ~ , ” P 2 ) -
- ( ~567P:
: + + +
9SP:u,Z 6 6 6 P : ~ i 2 7 ~ -
: 270(~,6P,
54(T3$(oT - a;)(a; - a;)((T;- a?)
(34)
5P2 - 3 P Z 4 - 3Pxu,”+ 3 u : 4
w2 = (35)
18a,”(aT- (T,”)((T,”- u;)