Session B1, Paper #1

Successful Flight Test of a GPS and Radar Aided Inertial Navigation System
J. Anders, C. Johnson, A. Luckau, T.A. Moore, R.S. Ornedo, Raytheon Electronic System

On January 25, 2001, a Standard Missile 3 (SM-3) flight test was successfully conducted off the coast of Kauai to evaluate the SM-3's airframe stability and control through kinetic warhead separation. This is the third in a planned series of nine test flights to demonstrate an exo-atmospheric hit-to-intercept of a ballistic missile target. The SM-3 is multi-stage missile that is launched from an AEGIS cruiser. Its third stage is equipped with an inertial navigation system that is aided by both GPS and uplinked radar data. This paper will describe the operational performance of this inertial navigation system based upon preliminary analysis of the January flight test data. The paper also will provide a brief top level design description of the GPS and radar aided inertial navigation system as well as its navigation accuracy requirements. Data comparing the differences between the GPS, radar, and inertial navigation solution will be presented. The navigation system performed superbly during the entire guided duration of the mission as well as during the unguided portion of the flight until the last telemetry signal. This aspect of the navigation system functional performance will be also addressed.

[Return to Program]


Session B1, Paper #2

Integrated GNSS/INS System Guaranteed Achievable Accuracy Exploration by Robust Minimax Estimate Approach
A. Bobrov, IMT s.r.l, Italy; V. Bobrov, Alenia Spazio S.p.A, Italy; C. Neusipin, M. Perelli

An inertial navigation solution exibits relatively low noise from second to second but tends to drift over time. Tipical aircraft inertial navigation errors grow at rates between one and ten nautical miles per hour (1.8 to 18 km/hr) of operation. In contrast GNSS errors are relatively noisy from second to second but exibit no long term drift. Using both of these sensors is superior to using either alone. Integrating of the information from of each sensor results in a navigation system, which operates like a drift-free INS.

The paper discusses the operational advantages coming from the integration of the navigation data derived from GNSS receivers on board commercial aeroplanes and ballistic missile with the data derived from inertial navigation systems. While the inertial systems allow for an operational autonomy of up to one hour between consecutive position fixes, the GNSS allows to determine with high accuracy the position fixes, overcoming the problems due to the long-term drifts typical of the inertial systems and which limit they dependability when used alone. On its turn the inertial systems eliminate a few drawbacks connected to the use of data derived from the GNSS: therefore the symbiosis between the two systems is a winning one from the functional and operational viewpoint.

An integrated GNSS/INS architecture uses a "polynomial" Kalman filter to estimate position, velocity, and acceleration states of the vehicle, and clock bias and clock drift rate. Kalman filtering is an estimation method, which minimises "average" estimation error. More precisely, the Kalman filter minimises the variance of the estimation error. But there are a couple of serious limitations to the Kalman filter.
The Kalman filter assumes that the noise properties are known.
The Kalman filter minimises the "average" estimation error.
Tacking into account that as a rule we don't know anything about the system noise statistics, the Minimax Estimate Approach is used for GNSS/INS guaranteed achievable accuracy estimation.
Minimax filtering is based on a bounded arbitrary correlated error model and minimises the "worst-case" estimation error. More precisely, the minimax filter minimises the maximum singular value of the transfer function from the noise to the estimation error.

The simulation have shown:
When GNSS is available its measurements dominate navigation performance. The steady state navigation error will be reduced by inertial aiding, which simply concidered allows GNSS measurement noise to be "averaged out". Imrovement of steady state error with improving inertial quality is not as dramatic.
Tightly coupled GNSS/INS system is superior to loosely coupled one for maintaining lock in a jamming enviroroment, but the gain is hard to quantify exept by enviroment in signal-to-noise (jammer) ratio.

[Return to Program]


Session B1, Paper #3

Robust Rapid Transfer Alignment with an INS/GPS Reference
P.D. Groves, G.G. Wilson, C.J. Mather, QinetiQ ltd (formerly DERA), UK

Transfer alignment is the process of aligning and calibrating a guided weapon INS using data from the host aircraft INS. An all-purpose rapid transfer alignment algorithm set has been developed in-house by QinetiQ (formerly DERA). Rapid transfer alignment algorithms differ from their conventional counterparts in that attitude data from the host aircraft is used for the main alignment process in addition to velocity data. Previous work [1] has shown how enhancements to a basic rapid transfer alignment algorithm can improve accuracy, particularly where lower grade inertial sensors are used. This paper presents the results of further studies, showing how transfer alignment robustness may be maximised.

To optimise conventional transfer alignment performance, it is necessary to align for a certain duration and for the host aircraft to perform a manoeuvre between weapon initialisation and launch. Simulation studies show that this also applies to rapid transfer alignment, but that the manoeuvre requirements are less stringent than for conventional techniques.

A problem with basic rapid alignment, where the weapon is mounted on a wing pylon of a fixed-wing aircraft, is that roll manoeuvres, such as wing rocks, can destabilise the Kalman filter, disrupting the estimation of attitude errors [2]. This is verified by simulation and two different solutions are demonstrated.

Weapon navigation accuracy may be significantly improved by using the aircraft INS/GPS navigation solution as the reference for transfer alignment as opposed to a pure inertial reference. However, where GPS is re-acquired after an outage of several minutes, a transient occurs in the INS/GPS solution which can disrupt transfer alignment [3]. Transient detection and handling techniques have therefore been added to the QinetiQ transfer alignment algorithm set. Where the transfer alignment algorithm is provided with a pure INS reference in addition to the INS/GPS reference or with aircraft INS corrections, transients may be handled directly. Otherwise, Kalman filter based integrity monitoring techniques must be used. Both approaches are verified by simulation.

Acknowledgement
This work was funded by the Applied Research Programme of the United Kingdom Ministry of Defence.

1. P D Groves & J C Haddock, An All-purpose Rapid Transfer Alignment Algorithm Set, ION NTM January 2001.
2. R M Rogers, Applied Mathematics in Integrated Navigation Systems, AIAA Education Series 2000.
3. P D Groves, Transfer Alignment using an Integrated INS/GPS as the Reference, ION Annual Meeting June 1999.

[Return to Program]


Session B1, Paper #4

Integration of Satellite and Cellular Positioning Technologies
R. Chen, The Finnish Geodetic Institute, Finland

A vehicle can be located with the Global Positioning System (GPS), Inertial Navigation System (INS) or cellular positioning techniques such as the Enhanced Observed Time Difference (E-TOD). However, none of these techniques work uniquely under all circumstances. The GPS has a global coverage and higher positioning accuracy, however, it requires the sight-of-lines to at least three satellites, thus it is not suitable for urban canyons and not applicable in indoor environments; the cellular positioning techniques do not need the sight-of-lines to the Base Transceiver Stations (BTS), however it has a limited coverage which is defined the cellular network. A portable INS system degrades rapidly with time, therefore it can not be used as a stand alone technique for positioning purpose. It is typically integrated with GPS.

The satellite and cellular positioning techniques are complemented each other. The manufacturers of mobile terminals are now starting to provide open architectures, e.g. EPOC and J2ME/MIDP(Java 2 Micro Edition/Mobile Information Device Profile), in their devices in order to support application developments from third-party. These open architectures provide a very good foundation for the integration of the satellite and cellular positioning technologies. The positioning accuracy with the current GSM technology is about 50 meters and it will be improved with the future 3G technologies. Although the positioning accuracy with cellular techniques is lower than that from GPS, the integration of these two technologies still makes senses especially in critical areas such as urban canyons, where the number of visible satellites is typically less than four. It is not always possible to locate a vehicle with GPS under such circumstances. Even occasionally more than four satellites are available, however the GDOPs defined by the visible satellites are typically poor because of the narrow streets and high buildings along both sides of the streets in city canyons.

Different integration scenarios, with different radio coverages and different cellular positioning accuracies, have been investigated based on a simulation of a cellular network in a city canyon. The study results show that the poor positioning solutions can be improved significantly by adding a couple of cellular measurements, especially in the situations when the satellite GDOPs are very poor. The cellular measurements improve the GDOPs significantly and are capable to provide location estimates for the situations when there are only two satellites available. The integration is still beneficial when the cellular positioning accuracy is about 50 meters.

[Return to Program]


Session B1, Paper #5

H-764GU EGI: Upgrade Path for CNS/ATM and NAVWAR
B.E. Fly, J. Elchynski, Honeywell Sensor and Guidance Products

The H-764G Embedded GPS/INS (EGI) has been an enormously successful product with over 70 different s/w adaptations (or missionizations) currently flying and more missionizations under development. When the H-764G was first introduced in 1994, one important design feature was the availability of spare card slots that could be populated with additional functionality as technology advanced or as new requirements were placed on the host platform. That forethought has now borne fruit with the impending changes to both the Military and the civil airspace environment. Now, many DoD and industry leaders are in the throes of determining how their current military systems will evolve to allow operation within the Communication Navigation Surveillance / Air Traffic Management (CNS/ATM) airspace mandates that are currently being levied. Key elements of the CNS/ATM requirements are performance and safety. CNS/ATM will eventually result in an environment the FAA has termed "free flight". Free flight requires a shift in Air Traffic Control (ATC) philosophy, from control to management. Free flight creates a homogenous airspace that will ultimately allow aircraft operators to select user preferred flight plans saving time, money, and improving airspace use efficiency. The reality is that the environment is dramatically changing for both the aircraft operators and the air traffic service providers. No operator is immune to these changes and those who desire to continue operating in civil airspace will be required to be compliant in terms of new procedures and supporting avionics . In addition to CNS/ATM mandates, Military Aircraft will need to consider the emerging NAVigation WARfare (NAVWAR) requirements along with new Military GPS service changes. NAVWAR is the localized denial of radio navigation service to adversaries while maintaining a superior navigation capability for ally forces in a theater of war.

In many platforms, these new requirements/upgrades mean additional avionics equipment in already over-crowded equipment bays and cockpits. The spare card slots in the H-764G unit provide an opportunity for these platforms to adhere to CNS/ATM and NAVWAR mandates without increasing LRUs and in some cases actually reducing overall space claims for these technologies. This paper will discuss Honeywell's technical upgrade path and the new capabilities of the H-764GU (U for Upgrade) that will satisfy these requirements. CNS/ATM Features to be incorporated in the H-764GU include the addition of Multi-Mode Receiver (MMR) functionality such as VOR, ILS, Marker Beacon, and LAAS capabilities, along with Navigation Management capabilities with associated Integrity Monitoring for RNP area navigation. Military features include a GPS receiver upgrade to address NAVWAR and over-the-air- keying, along with Joint Precision Approach & Landing System (JPALS) capabilities.

[Return to Program]


Session B1, Paper #6

Mitigation of Sculling and Commutation Errors by Frequency-Domain Development of INS Algorithmic Part
A. Soloviev, F. van Graas, Ohio University

A new approach for improving Inertial Navigation System (INS) accuracy is proposed. The approach utilizes block-processing development of the INS algorithmic part in the frequency-domain. A generic frequency-domain INS computation starts with the reconstruction of continuous-time signals from large enough blocks of input discrete samples. A discrete Fourier transform, compensated for boundary discontinuities, is applied for the reconstruction. Reconstructed signals are represented in the frequency domain as sets of Fourier frequencies and corresponding spectrum amplitudes. Frequency-domain represented signals are analytically transformed into navigation outputs, contrary to conventional methods, which rely on numerical solutions. The approach proposed therefore serves as an efficient tool for the mitigation of conventional drawbacks associated with numerical solutions utilized.

The paper implements frequency-domain techniques to mitigate the influence of sculling effect and commutation error on INS accuracy. The sculling effect originates from the fact that conventional inertial algorithms operate with velocity increments that are obtained by pre-integration of accelerometer specific forces. As a result of sculling, a motion is sensed in the direction perpendicular to the true motion even though no net position change in that direction occurs. Frequency-domain INS algorithms reconstruct and subsequently process continuous-time specific forces instead of velocity increments thus eliminating the sculling origin. The frequency-domain approach is also applied for suppressing the effect of commutation error on INS accuracy. Commutativity of attitude angles is only valid when the angles are infinitely small. If the angles are not infinitely small, commutation error can only be avoided if coning motion can be reconstructed mathematically. In this paper, a discrete Fourier transform, compensated for boundary discontinuities, is used to represent the continuous-time solution of attitude differential equations as an alternative to conventional discrete-time solution methods. As a result, a significant mitigation of commutation error is achieved. Contrary to existing methods of coning reconstruction, the method proposed in this paper does not assume particular coning environments and mitigates commutation error for an arbitrary type of angular motion. Test results are presented that confirm the efficacy of the proposed approach.

[Return to Program]


Session B1, Alternate #1

Integrated Accelerometer/GPS Heading Estimator
L.S. Wang, F.R. Chiang, F.T. Huang, National Taiwan University, Taiwan

Based on the idea of Kalman filtering, the heading estimator of a baseline vector integrating accelerometer and Global Positioning System]GPS^is designed in this paper. The accelerometers not only can provide the acceleration of a vehicle, but also can be used to obtain attitude information of the vehicle by suitable arrangement and integration. For the problem of heading estimation, we may place four sets of one axis accelerometers at both ends of the baseline. By measuring the tangential acceleration and the centripetal acceleration, an extended Kalman filter can be used to compute the heading angle of the baseline vector. On the other hand, using the GPS Carrier-Smoothed-Code]CSC^ algorithm, the heading angle estimate can be obtain in the rate of 1Hz. Although the accelerometers can provide the attitude information at a higher sampling rate]e.g. 100Hz^, the accumulation errors due to integration make it necessary to perform adjustment periodically. On the contrary, although GPS estimation has lower sampling rate, it can be used to correct the drifting errors of the accelerometers. To integrate the two systems, a multi-rate filter is designed to obtain the optimal and real-time heading angle estimation. From the simulation results and the experimental results described in this paper, it is observed that the 2nd order EKF can solve the singularity problem from the nonlinear measurement equation when the angular velocity is close to zero. Moreover, the proposed accelerometer/GPS integrated heading estimator can indeed eliminate the errors from the drift phenomenon, and provide the accurate heading angle estimate of the baseline vector in real time.

[Return to Program]


Session B1, Alternate #2

Modernization of a High Precision, Heavy Payload Centrifuge
H.A. Stoffel, L.P. Fallon, B.A. Blanchard, C.V. Trainor, C.S. Draper Laboratory, Inc.

With the anticipated reduction of missile flight-testing and the continued need for dynamic G-load testing of inertial guidance systems, ground based replacements for flight test dynamics become extremely attractive from cost, performance and repeatability standpoints. Draper Laboratory of Cambridge Massachusetts has recently completed significant technological enhancements to its 28-ton, 57-foot long centrifuge. The Centrifuge is located in a DSS approved secure facility adjunct to Hanscom Air Force Base in Concord, Massachusetts.

The original system, designed in the early 1950's, was a 35 G hydraulic drive centrifuge. The main 65-inch spindle mounted bullgear drove the 28-ton arm. In addition to the hydraulic main axis, a hydraulically driven counter rotating platform (CRP) was also available to "unwind" the main axis rotation for angular rate sensitive test articles. The system was modified for limited special testing in 1962 to provide 100 G's of acceleration by adding additional motors and pumps.

By the early 90's, the hydraulics were showing their aging and reliability limited usage to very short tests. In 1996, Draper decided to upgrade the system to an all-electric servo motor drive system. Carco Electronics (Pittsburgh, PA) was selected to redesign and upgrade the system from hydraulic to electric. The redesign system drive specified a 25 G minimum capability (firmware limited), under acceleration controlled ramps of 1 G/ min from 0-25 G. Servo position readout was specified to 1.2 arc sec, rate accuracy of 25 ppm (per rev basis) with a rate stability of 0.1% peak.

In addition to Draper revamping the drive system several other upgrades were incorporated in the overall plan. These upgrades included an improved CRP, a laser interferometer, 389 slip rings, and a dynamic monitoring system with state-of-the-art sensors for real-time measurement of arm stretch, shaft/spindle deflection, and main axis wobble.

The original intent for the centrifuge was to test inertial guidance systems for ballistic missiles. With the recent enhancements and versatility, the centrifuge can easily be adapted to handle numerous other military and commercial applications. Payloads of 5000 pounds are reasonable. If required, modifications to the counterweight side of the arm will allow for heavier load balancing.

This paper details the technological refurbishment effort for the C. S. Draper Laboratory high precision, heavy payload centrifuge and describes the dynamic environments and capabilities of this unique test facility.

[Return to Program]