Learning Center
Plans & pricing Sign in
Sign Out
Get this document free

Atlantic Inertial Systems


									Atlantic Inertial Systems
MEMS Sensor and Integrated Navigation Technology for Precision Guidance
MC2139 Rev 1 March 2009

ABSTRACT Atlantic Inertial Systems (AIS), based at Plymouth UK, has developed a range of Inertial Measurement Units (IMUs) based on vibrating structure MEMS sensors. The first generation IMU is in military service and in volume production; a second generation IMU, using digital control and providing enhanced performance and shock tolerance, is presently completing development and is in production. To provide a full navigation capability, the MEMS IMU has also been ultra tightly coupled with a suitable GPS receiver, using a highly efficient modular Kalman Filter, in a package designed for missile, munition and projectile guidance. This paper describes the capabilities of MEMS inertial sensors applied to military guidance applications. The technologies associated with vibrating structure sensors are explained with consideration to how the key objectives of miniaturisation, survivability, affordability and performance are achieved in a gun hard IMU. A programme of analysis and testing, aimed at establishing design margins, has been conducted. This has included a number of firing trials to confirm reliable operation following exposure to over 20,000g launch shock. Results from these tests are presented, including integrated navigation performance for a unit following a 155mm artillery firing. For further information please do not hesitate to contact:

Derrick Cox BS (Hons) Business Development Executive Clittaford Road, Southway Plymouth, Devon PL6 6DE United Kingdom Telephone +44 (0)1752 722103 Fax +44 (0)1752 695927 Mobile +44 (0)7801 712406 Email

Kevan Flintoff Business Development Manager Clittaford Road, Southway Plymouth, Devon PL6 6DE United Kingdom Telephone +44 (0)1752 723386 Fax +44 (0)1752 695927 Mobile +44 (0)7801 716424 Email

Atlantic Inertial Systems

MEMS Sensor and Integrated Navigation Technology for Precision Guidance

upper pole silicon

package lid magnet

BACKGROUND Micro Electro-Mechanical Systems (MEMS) are the result of the evolution of microfabrication techniques, originally developed for integrated circuit manufacture, for the integration of micromechanical structures with electronic components. The technology has produced devices which are inherently small, robust and amenable to batch fabrication, creating unprecedented levels of functionality and reliability, at a relatively low cost.
1.2 MEMS VIBRATING STRUCTURE GYROSCOPE Coriolis vibrating structure rate sensing technology has been an alternative to rotating mass and optical gyroscopes for many years. However, the emergence of MEMS techniques has allowed this technology to deliver a class of instrument with a wide range of applications at a very low price. Atlantic Inertial Systems has applied this technology to develop a rate sensor based on a micromachined silicon ring, manufactured from a silicon wafer, illustrated at Fig 1. First produced in 1999, over 15,000,000 of these sensors have been delivered to date.

support glass pedestal glass package base

lower pole

Fig. 2 Resonator head construction


Resultant vibration

Fc = Coriolis force

Fc Fc
Fig. 3 Coupling of Coriolis forces in resonating ring sensor

Fig. 1 MEMS ring resonator batch fabrication

The rate sensor is formed by mounting the silicon ring within a magnetic field. A drive current is applied through a conductive path deposited on the ring to establish a vibrating resonance through electromagnetic induction. In the presence of angular motion the vibration nodes precess around the circumference of the ring, and are detected by a pick-off circuit. The device operates in closed loop mode, with a secondary drive circuit configured to oppose the precession. The resonator head is illustrated at Fig. 2 and Fig. 4, and the effect of rotation on the vibrating mode at Fig. 3.

Fig. 4 Exposed resonator head assembly

SiIMU02® (MEMS IMU) SiIMU02® is an all-digital, second generation MEMS IMU and is an evolution of Atlantic Inertial Systems' first generation SiIMU01®, which is in volume production at the facility in Plymouth, UK and is in service with a number of armed forces globally. SiIMU02® utilises mature and proven MEMS sensor technology in a self-contained 4 cubic inch package, delivering angular and linear measurement in 3 axes.

The core pod weighs less than 250g and consists of a main processor board that provides overall signal processing and interfacing functions, three digital gyro boards, each housing an Atlantic Inertial Systems MEMS SiVSG® rate sensor, and three MEMS accelerometer boards all mounted orthogonally. The product implements a variable rate range gyro, giving low noise performance at low rate, but a rate range capability of 18000°/s. Due to the modular nature of the product, sensor parameters such as accelerometer operating range can be tailored during manufacture for specific applications, with no effect on the form or fit of the product.

Signal Pre-processing and Sampling

Correlators and Tracking Loop #1 Velocity / Acceleration Clock Drift and Bias Estimates

GPS Receiver r1
Accelerometers Gyros Inertial Sensor Measurement Pre-processing


Navigation and Filtering Algorithms

Blended Navigation Solution


Integration Software

Fig. 6 SiNAV02® ultra tightly coupled architecture

The unit has the spare processing capacity and a software architecture to allow hosting of guidance and control algorithms for complete mission and flight management of the host platform.


SiNAV02® is a Atlantic Inertial Systems-designed MEMS IMU/GPS integrated navigation system aimed at meeting the demanding requirements of guided munition applications. SiNAV02® incorporates the SiIMU02® MEMS IMU for inertial sensing, a C/A code GPS receiver and processor card hosting the Modular Integrated Navigation Kalman (MINK) filter. The capability to incorporate a SAASM based Precise Positioning Service receiver will be available in the future. The MINK filter has been specifically designed around a MEMS IMU performance and has been structured to allow fast and simple inclusion of extra aiding sources. The GPS and IMU are ultra tightly coupled through the MINK Filter. The filter uses satellite line of sight (LOS) data allowing the GPS to aid the navigation solution when tracking fewer than 4 satellites, much improved from looselycoupled, where the GPS needs to form its own solution to aid the system navigation. The filter also aids the GPS, allowing it to track the satellite signal through dynamics and maintain lock with a much reduced bandwidth. This reduction in bandwidth improves the signal to noise ratio (SNR) for each satellite and gives approximately 15-20 dB anti-jam improvement.

Fig. 7 SiNAV02® MEMS IMU/GPS 2.


2.1 EFFECTS OF SHOCK MEMS sensors have an inherent tolerance of extreme environments and have the potential, through appropriate development, to survive the high shock and vibration conditions of artillery projectile launch [1]. The silicon ring structure employed by Atlantic Inertial Systems has multiple support limbs and is highly resistant to failure under extreme acceleration. Finite Element Modelling has simulated gunfire acceleration and has shown that, at 20,000g, the forces experienced are less than 10% of the material fracture limit. Fig. 8 illustrates a sector of the ring and its support structure. Fig. 9 indicates (with exaggerated displacement) the effects of in-plane and across-plane acceleration. This analysis has been confirmed by centrifuge and gunfire testing.

Fig. 8 MEMS silicon ring support structure

Fig. 9 Analysis of resonator structural stresses under high acceleration 2.2


Fig. 10 Aerobutt facility TABLE 1

AEROBUTT An Aerobutt test facility, located at the BAE Systems Ridsdale Range, in Northumberland UK, has been used for initial gun hardness evaluation of the integrated SiIMU02® unit. This system is a Mortar Soft Recovery System, comprising an 81mm mortar pressure barrel mounted horizontally with a small gap between the end of the barrel and the recovery tube. The recovery tube is 160 meters long and has a restriction at the far end. The mortar round is decelerated by an increase of forward air pressure as it travels down the tube. Shock levels are adjusted by the varying the propellant mass and conditioning temperature. The barrel pressure is measured by the deformation of copper balls placed at the rear of the launch tube and this is used to determine the peak shock level. During the development of SiIMU02®, thousands of sensors and hundreds of IMUs have been Aerobutt tested. Two recent tests are described below.
2.2.3 SURVIVABILITY TEST To establish design margins and to confirm readiness to proceed to an artillery firing trial, SiIMU02® and SiNAV02® units were tested at levels in excess of the normal 20,000g threshold. Further to this, providing units survived their first firing, they would be fired again. This trial demonstrated 100% success at the first firing and the ability for units to survive and perform following 3 successive firings at levels between 12,000g and 22,000g.

Pre Firing Gyro Turn On Bias (1σ) Gyro Scale Factor (1σ) Gyro Misalignment (max) Acc Turn On Bias (1σ) Acc Scale Factor (1σ) Acc Misalignment (max) 85 °/hr 690 ppm 4.0 mrad 3.7 mg 465 ppm 2.2 mrad Post Firing 110 °/hr 905 ppm 4.0 mrad 8.8 mg 400 ppm 3.8 mrad

These Aerobutt test results demonstrate that the SiIMU02® performance is substantially insensitive to the effects of gun launch shock.
2.2.5 155mm ARTILLERY

A fully representative artillery firing trial was carried out with SiNAV02® units at the Alkantpan (Vastrap) firing range in the Northern Cape of South Africa (27°50’S21°38’E). Vastrap is located in a large region of soft sand, allowing test rounds to be recovered intact by excavation; impact shock levels being around 5,000g. The launch platform was a 155mm, 45 Calibre G5 towed Howitzer. Test equipment available at the test range included a muzzle velocity radar, piezo and copper transducers for barrel pressure measurement and tracking radars.

IMU PERFORMANCE FOLLOWING AEROBUTT TEST To quantify the sensor performance changes following gun shock, a further Aerobutt test was conducted to fire a population of IMUs at shock levels greater than 12,000 g. Their performance was measured both before and after the firing and the changes quantified. The results are summarised at Table 1.

Fig. 11 G5 Towed Howitzer used for firing trial

dynamics. Results from an IMU having experienced 20,000g exposure are at Fig. 14 to Fig.17. Fig. 14 and Fig.15 cover the bias estimates during the initial period with GPS enabled. The test procedure includes a series of manoeuvres to progressively enable state observability. The results give an indication of the run-to-run repeatability and in-run stability. The in-run stability for the gyros is noted as better than 10°/hr and for the accelerometers as better than 2 mg. The periodic shift in gyro bias is an observation of Earth rate, due to the unit being only nominally aligned in heading.


Bias Degrees/Hr









Fig. 12 SiNAV02® fitted to the slug projectile prior to firing

X Gyro Case A X Gyro Case B Y Gyro Case A Y Gyro Case B Z Gyro Case A Z Gyro Case B


Units were fired at three intended setback accelerations of 12,000g 16,000g and 20,000g. The control procedures applied during the trial ensured the actual forces were achieved to within 2.5%. The ability to rapidly recover and dismantle the test round allowed the units to be checked between firings and to modify subsequent tests should a failure occur. The units were to be non-operational throughout the flight and powered upon recovery. All projectiles were successfully fired, recovered and separated.


-200 Acquisition & Levelling

X&Y Gyro Bias Estimated

Rotation Z Gyro Bias Estimated




Fig. 14 Gyro bias estimates


0 20
Acc Bias milli g






X Acc Case A X Acc Case B Y Acc Case A Y Acc Case B Z Acc Case A Z Acc Case B





Fig. 15 Acc bias estimates

Fig. 13 Dismantling the round post firing

The on-site facilities allowed for basic functionality tests to be conducted on units immediately after the firing. IMU sensors were individually checked for correct measurement of rate and acceleration, and it was confirmed that the GPS receiver data retention and clock functions had been preserved through launch. After the gun fire trial and subsequent soft recovery, the SiNAV02® units were returned to Plymouth for detailed assessment of any performance shifts. All units successfully completed the factory acceptance test, confirming that the inertial sensors and GPS receiver oscillator were still performing within specification. They were then subjected to laboratory and road trials to assess the performance of the filter, GPS and IMU. The laboratory tests consisted of a period of GPS-aided navigation, followed by free inertial navigation for 60 seconds: case A representing low dynamics and case B high

Fig. 16 and Fig. 17 indicate the free inertial navigation performance: following alignment, GPS aiding is removed (at time =0) and the subsequent navigation accuracy assessed for up to 60s. At the end of this period the position error growth is 2m vertical and 12m horizontal. The extremely low error growth in the vertical axis confirms accelerometer stability better than 300µg; the horizontal drift is consistent with gyro performance of better than 10°/hr.


Velocity m/s Radial Vertical


0 -30 -20 -10 0 10 20 30 40 50 60


Fig. 16 Free inertial velocity performance


std dev attitude error (deg)

1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0 5 10 15 GPS outage tim e sec 20 Bank Elevation Heading



Position m


Radial Vertical


Fig. 20 Attitude performance
0 -30 -20 -10 0 10 20 30 40 50 60


Fig. 17 Free inertial position performance

Further to the live fire trials results and subsequent laboratory testing, simulation work has been carried out to assess integrated performance in representative dynamic conditions. Simulation results from guided munition trajectories are presented below, and particular attention is paid to the performance of the system following the denial of GPS. The route used was a low dynamic flight, with GPS denied at progressively earlier times in the flight. The route was deliberately low dynamic to limit the potential for the Kalman Filter to observe and correct IMU errors states. The performance of the navigation solution is presented with respect to position, velocity and attitude at the target with varying durations of GPS denial. The x axis represents the GPS outage period before reaching the target.

The dynamics of the flight represent a typical target approach, with the manoeuvres close to the target allowing observability of error states while GPS is present, but generating errors if GPS is denied. This is seen most strikingly with the high accuracy maintained by the system, even with 10 seconds of GPS denial, once error states are observed.
3. APPLICATIONS The military environment is now dominated by terms such as “Network Centric Warfare” and “Asymmetric Warfare”, and this is driving developments in guided missile technology. The general requirements generated by today’s armed forces include:

• • •

Precision guidance and controlled effect for single-shot kill probability with minimal collateral damage. All weather capability and rapid reaction between initial target detection and destruction. Reduced logistic resource supporting rapid reaction deployment

70 std dev position error (m) 60 50 40 30 20 10 0 0 5 10 15 GPS outage tim e sec 20 North East Vertical

Fig. 18 Position performance
6 5 4 3 2 1 0 0 5 10 GPS outage tim e sec 15 20 North East Vertical

std dev velocity error (m/s)

Inertial integrated navigation solutions are fast becoming the technologies of choice to provide this precision guidance in today’s theatres of conflict. This has never been highlighted more than during the recent campaigns in Afghanistan and Iraq, which saw mainly US technology in the form of GPS guided munitions, dominate the air campaigns. The importance of precision guided bombs and munitions has continued to increase as they continue to replace non-guided “dumb bombs”. The transition to guided weapons is now becoming a global requirement with inertial and GPS based guidance solutions being evaluated and used in a variety of applications: • • • • Air Launched Munitions Artillery launched shells and mortars Land and air launched rockets Variety of platform launched missile

Fig. 19 Velocity performance

Driving the insertion of MEMS based inertial technologies into precision guided weapons is the need to utilise new technologies that meet the requirement of maintaining the correct price per kill ratio, whilst leveraging other benefits that MEMS technology has to offer over conventional inertial systems: reduced size, power; cost; fast start up time, whilst meeting the environmental performance required for new gun launched applications.

Figs. 21 and 22 indicate pictorially, and Fig. 23 graphically, the sequence of events comprising a guided munition attack profile. The air launched profile is generally more benign from the perspective of the navigation system, but the key events are similar; following power up the objectives are to derive the platform attitude and acquire GPS. At this point the navigation Kalman filter is initialised and can rapidly progress to a GPS/inertial navigation configuration. The IMU then supports continuous navigation through the terminal phase, during which GPS may be denied by jamming.
Fig. 21 Air launched precision guided munitions.

Fig 22 Gun launched guided shells and mortars

Time from launch Distance to target




40s 15Km 10Km 5Km

IMU data available GPS acquired Attitude initialisation Alignment Aided navigation Rate control regime Navigation control regime Free inertial navigation

Fig.22 Typical timeline for guided munition

SUMMARY MEMS based inertial systems designed and manufactured at Atlantic Inertial Systems in Plymouth, UK are meeting the performance and environmental needs of the growing precision guided market for new applications whilst replacing conventional inertial systems in a range of existing application types. AIS MEMS based technologies, such as SiIMU02®, are in service with armed forces and are being used in a large number of development programmes globally. The integrated SiIMU02® and GPS system SiNAV02® is undergoing extensive customer air launched and further gun launched firing trials.

Performance levels that, hereto, are not met by conventional inertial technology: • • High spin rate applications. Ability to withstand and operate up to spin rates of 18,000º/s. Rate sensor control to maintain high performance over a variable rolling mission profile Ability to survive and operate within the required performance limits following a 20,000g launch shock.


Customer selection of these products is based on the decision that these lower cost MEMS based sensors can meet the performance required and are allowing the design of new applications such as guided projectiles and small calibre guided rockets. Key to the development of the new suite of guided systems is the ability for the navigation system to meet the performance and environmental tolerances required.

The ability to provide an integrated MEMS based IMU/GPS system, such as Atlantic Inertial Systems' SiNAV02®, is providing access to the growing demand for all-weather GPS guided munitions. In designing an Ultra Tightly Coupled system, the aforementioned benefits of a MEMS based IMU system have been coupled with the enhanced benefits of GPS integration and guidance; providing MEMS based solutions for today’s precision guided warfare.
5 [1]


J-M Stauffer, Current Capabilities of MEMS Capacitive Accelerometers in Harsh Environments, PLANS 2006.

MC2139 Rev 1, March 2009

To top