Tuning of Coning Algorithms to Gyro Data Frequency Response Characteristics

Abstract

Substantial efforts have gone into the development of sophisticated algorithms that reduce system drift errors in the presence of coning motion. Present-day algorithms form high-order coning correction terms using multiple incremental angle outputs from the gyros. These algorithms assume a at transfer function for the processing of the incremental angle outputs and are structured to yield very high-order responses. However, these algorithms do not address the issue of nonideal gyro frequency response or of ltered gyro data. Many gyros exhibit complex frequency responses and violate the assumptions used in deriving the previously developed coning algorithms. The mismatch between the assumed and actual frequency response of the gyro data leads to degradation of performance in a coning environment as well as ampli cation of pseudoconing errors. A method of deriving algorithms that are tailored to the frequency response of the particular type of gyrosused is presented. These algorithmscan be designed to arbitrarily high order and can also supply an extremely sharp high-frequency cutoff to minimize pseudo-coning errors. Additionally, this method can be used to design coning algorithms that are tuned to deliberately ltered gyro data. The technique developed equally applies to mechanical, ber-optic, and other types of gyros.

Cone Algorithm of Spinning Vehicles under Dynamic Coning Environment

“…Due to the sampling number of a triaxial gyro, single-sample algorithm, two-sample algorithm, three-sample algorithm, and so on have been developed, and the validity of algorithm is described by drifting error. One stage is updating the attitude in an attitude updating cycle, and it is referred to in (8) to (12). The other stage is calculating the rotation vector including an integrated gyro sensed angular rate and a coning correction calculated from gyroscopes data, and it is referred to in (13) to (16).…”

“…Meanwhile, the gyro frequency response is a primary cause for pseudo-coning error that is another form of coning error, and the traditional coning correction algorithms cannot work to highorder accuracy. Mark and Tazartes proposed a tuning highorder coning algorithm to match the frequency response characteristics of the gyro and avoided overcompensation for pseudo-coning correction [12]. Kim and Lee added selfvibration of gyro to the causes of pseudo-coning.…”

Strapdown at the Crossroads

“…Ironically, some became widely available only after faster processing capabilities made them less critically urgent. Also, after decades of algorithmic efforts spawned by Miller [5], recent work [4] has shown how the coefficients need modification to account for gyro response (in fact, this work showed how the intended corrections, formed without accounting for gyro frequency response, can result in amplification -not reduction -of error). Fortunately, as the ensuing discussion illustrates, all this complexity can be sidestepped in a direct formation of position/velocity/attitude time histories from raw data inputs.…”

“…Savage [3] developed and propagated myriad techniques, providing beneficial contributions throughout the entire 40-plus-year period. Mark [4], and others too numerous to mention, perfected their algorithmic approaches until little or no further improvement in computational efficiency was necessary (or perhaps even possible).…”

Error Modeling, Calibration, and Compensation of Inertial Measurement Unit (IMU)

“…However, the filter will change amplitude-frequency characteristic of gyro output, which will further cause inaccurate conical error compensation. In [49], a method to correct the rotating vector attitude algorithm by combining frequency characteristics of filter is introduced, and the system therein adopts digital filter able to improve the gyro resolution. It is pointed out that when amplitude-frequency characteristic F ( ) of digital filter satisfies…”