Coning algorithm based on
singular perturbation
Li Fu and Lingling Wang
Automation Science and Electrical Engineering Department, Beihang University, Beijing, People’s Republic of China, and
Jianghai Hu
Electrical and Computer Engineering Department, Purdue University, West Lafayette, Indiana, USA
Abstract
Purpose – The aim of this paper is to propose a new coning correction algorithm, based on the singular perturbation technique, for the attitude update
computation with non-ideal angular rate information.
Design/methodology/approach – Unlike conventional coning correction algorithms, the new method uses angular rate two-time scale model to
construct the coning correction term of attitude update. In order to achieve balanced real/pseudo coning correction performance, the selection
guidelines of the new algorithm parameters are established.
Findings – Performance of the new algorithm is evaluated by comparison with the conventional algorithm in no ideal sensors undergoing stochastic
coning environments. The accuracy of attitude update can be improved effectively with reduced computational workload by using this new coning
algorithm as compared with conventional ones.
Practical implications – The proposed coning correction algorithm can be implemented with angular rate sensors in UAV (unmanned aerial vehicle)
and other aircrafts attitude estimation for navigation and control applications.
Originality/value – Singular perturbation is an effective method for structuring coning correction algorithm with filtered angular rate outputs in
stochastic coning environments. The improved coning correction algorithm based on singular perturbations reduces the real and pseudo coning effects
effectively as compared with conventional ones. It is proved to be valid for improvement of accuracy with reduced computations of the attitude update.
Keywords Coning algorithm, Singular perturbation, Two-time scale model, Attitude update, Pseudo coning correction, Aircraft, Navigation
Paper type Research paper
Introduction
The primary function in strapdown inertial navigation system
is the attitude update computation. In order to improve the
accuracy of attitude update calculations, modern strapdown
inertial navigation attitude update algorithms are generally
composed of two parts: the calculation of the rotation vector
to determine the vehicle attitude change over the iteration
interval, and the quaternion update to determine the vehicle
attitude (Kim and Tae, 2001). The calculation of the rotation
vector is used to account accurately for the noncommutativity
error or coning error which is one of the major error sources
in solution of attitude equation. Therefore, the core of
attitude update algorithms is to minimize the coning error
under the unpredictable motion of the vehicle.
Bortz (1971) pioneered the rotation vector concept to
compensate the noncommutativity error within the attitude
update interval. By extending his research works, many
sophisticated coning correction algorithms have been
developed to reduce the coning error in a pure coning
environment, two representatives of which are shown in Miller
(1983) and Park et al. (1999). A rigorous comprehensive
description of the high-speed high-order coning correction
algorithm design process is provided by Savage (1998). In order
to match the frequency response characteristics of the
gyroscopes which induce pseudo-coning error in attitude
update, Mark and Tazartes (2001) discloses a method of
tuning high-order coning correction algorithms.
Modern-day general design procedure for coning correction
algorithms is based on the assumption of coning correction
model composed by the multiple sample gyro outputs to update
the rotation vector in the quaternion update interval (Roscoe,
2001). Usually the greater the model orders of the coning
correction algorithm, the lower the coning errors (Miller, 1983;
Park et al., 1999). A key point of these coning correction
algorithms is in optimizing the model’s coefficients to achieve
optimum coning correction performance in a pure coning
environment. According to Savage (2010), two different optimal
approaches were derived to tune the model’s coefficients. The
first approach, as time-series coning correction algorithm,
was based on truncated Taylor time-series expansion
approximations for sensed angular rate. The second approach
was frequency-series coning correction algorithm introduced by
Miller (1983) and Ignagni (1996), which was achieved by using
a truncated Taylor series expansion formulation in powers of
coning frequency. Rather than the above optimal approaches,
Savage (2010) introduced a new explicit frequency shaping
concept for coning correction algorithm design that used
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1748-8842.htm
Aircraft Engineering and Aerospace Technology: An International Journal
85/3 (2013) 178 –185
q Emerald Group Publishing Limited [ISSN 1748-8842]
[DOI 10.1108/00022661311313614]
This work was supported in part by a grant from the National Natural
Fund of China, Grant No. 61071014.
178