1/30

���������Surveillance Automatisée de la Route pour l’Information

des conducteurs et des gestionnaires

Recherche des Attributs pour le Diagnostic

Avancé des Ruptures de la Route

CAHIER DES CHARGES POUR L’ESTIMATEUR DE

PROFIL DE SOL

Livrable n°1.6.1

01/08/2005

Prédit SARI – première tranche – Projet financé par la Direction de la Sécurité et de la Circulation Routières

Responsable : Gilles SCHAEFER Contributeur : Loic CEBALLOS

2/30

TABLE DES MATIERES

1. BUT DE L’ETUDE __________________________________________________________ 3

2. BUT DU DOCUMENT _______________________________________________________ 3

3. LES METHODES POSSIBLES DE RECONSTRUCTION DU SOL __________________ 4 3.1. Principe de la méthode du modèle embarqué _______________________________________ 5 3.2. L’approche du risque, approche embarquée RADARR ______________________________ 6 3.3. Rappel sur le modèle CALLAS___________________________________________________ 6 3.4. Méthode analytique ____________________________________________________________ 7 3.5. Méthode d’hybridation de capteurs _______________________________________________ 7 3.6. Méthode d’observateurs ________________________________________________________ 7 3.7. Conclusion sur les méthodes possibles _____________________________________________ 7

4. LES DIFFERENTS CAPTEURS POSSIBLES____________________________________ 8 4.1. Capteurs possibles _____________________________________________________________ 8 4.2. Avantages et Inconvénients des différents capteurs possibles __________________________ 9

5. CAHIER DES CHARGES DE L‘APPLICATION « RECONSTRUCTION DU SOL » ___ 10 5.1. Méthodologie de développement_________________________________________________ 10 5.2. Problèmes attendus ___________________________________________________________ 10

6. RECHERCHE DES METHODES ANALYTIQUES POSSIBLES ___________________ 11 6.1. Théorie simple du dévers_______________________________________________________ 11 6.2. Théorie simple de la pente______________________________________________________ 12 6.3. Algorithmes analytiques possibles _______________________________________________ 12

6.3.1. Terminologie________________________________________________________________ 12 6.3.2. Avantage et inconvénients a priori des méthodes __________________________________ 12 6.3.3. Les différentes méthodes de reconstruction du sol _________________________________ 13

7. CONCLUSION_____________________________________________________________ 13

8. ANNEXE : TENDANCES DANS L’INDUSTRIE DU CAPTEUR ___________________ 14 8.1. Sensor industry developments and trends _________________________________________ 14 8.2. BOSCH Semiconductors GYROSCOPE __________________________________________ 19 8.3. A Critical Review of MEMS Gyroscopes Technology _______________________________ 24 8.4. Capteur CROSSBOW _________________________________________________________ 30

3/30

1. BUT DE L’ETUDE Il y a deux intérêts à reconstruire le profil du sol :

a. dans le cadre des aides à la conduite, mieux séparer les facteurs et ainsi permettre une action plus pertinente des aides à la conduite :

• les mauvais ABS confondent les variations de vitesse des roues sur l’uni avec des amorces de blocages, réduisant les performances du freinage sur route mauvaise où le risque est plus fort

• les mauvais ESP confondent les « virages relevés » sur la route avec des amorces de dérapage et activent le freinage d’une roue

b. dans le cadre de SARI, permettre le diagnostic d’un itinéraire pour lequel le profil du sol au niveau global (pente et devers) et local (uni) peut augmenter le risque

2. BUT DU DOCUMENT Ce document est le livrable RADARR 1.6.1 intitulé « Cahier des charges pour l’estimateur de profil de sol ». Pour reconstruire le profil du sol, de nombreuses méthodes peuvent être envisagées, mais elles nécessitent toutes des capteurs. Or si l’algorithmie passionne de nombreuses personnes et coûte le même prix quelque soit le choix des capteurs, l’examen des capteurs possibles de manière réaliste (incluant l’angle économique) est souvent négligée. Or l’algorithmie n’est pas adaptable d’un choix de capteur à l’autre. Un choix un peu trop rapide des capteurs va donc limiter irrémédiablement l’applicabilité du système (dans une logique de duplication). Nous avons donc proposé de remettre rapidement un livrable « Cahier des charges pour l’estimateur de profil de sol » pour orienter les travaux ultérieurs des différents partenaires. Ce document a pour but de rendre compte des différentes méthodes de reconstruction du sol, d’en déduire un cahier des charges des capteurs à utiliser, puis enfin de proposer les algorithmes possibles. Le but privilégié est ici de donner l’ordre de grandeur du dévers et de la pente de la facette sous la voiture à cette échelle, et la méthode privilégiée est d’utiliser les capteurs déjà présents sur le véhicule. Différentes solutions de reconstitution sont possibles, selon les voies d’acquisition et de calcul de CALLAS, il faut donc déterminer quelles méthodes sont les plus performantes.

4/30

3. LES METHODES POSSIBLES DE RECONSTRUCTION DU SOL

Le LCPC, comme ses homologues mondiaux, effectue des mesures de route depuis des décennies :

• avec un homme et une règle sur la route • à partir d’un véhicule spécial, doté notamment d’un gyroscope qui donne une position

absolue de la caisse, permettant de redescendre de manière fiable à la route

Des publications ont proposé des méthodes optiques originales et intéressantes (Canudas de Wit, Carlos Laboratoire d'Automatique de Grenoble UMR CNRS 5528, ENSIEG-INPG , BP 46, 38402 Grenoble ) pour déterminer le roulis et le dévers.

Nous sommes ici dans un contexte de mesures embarquées que nous rappelons ci-après.

5/30

3.1. Principe de la méthode du modèle embarqué

Figure 1 : principe de la méthode embarquée

filtres numeriquesNEXYAD

Recalagede l'

Adherence

adherenceaffichée

commandes du conducteur

Capteurs + filtreshardware

filtres numeriquesNEXYAD

reconstructionVlin, TGL et AD

Vx, Vy,Psi', B

W,

erreur glissement = |Calcul|-|Acqui|entrée du recalage sur TGL AD Ax Ay

table floue :gradient

d'adhérence,contexte

intégration dugradient, adherence

modifiée

asservissementde mimétisme

Voiture réelle Voiture VirtuelleSimulation 3D embarquée

plancher etplafond

Fx, Fy ext

report

recalage del'adhérence

AcquisitionINRETS

Retardsnumériques SERA

Correction TGL et ADselon Fx et Fy ext.

Vx, Psip

RDev, TGL(si suivi vitesse sur

roues arrière)

MesureCapteursCalculMesureCapteursAcqui

TGL et AD

commandes

Vx, Psi'

AD, TGL AD, TGL

6/30

3.2. L’approche du risque, approche embarquée RADARR

Taux de sollicitation relatif Risque

Attributs géométriques locaux de la chaussée (dévers, uni …)

Détermination de sa vitesse d’approche par le conducteur

Adhérence mobilisable

Commandes volantdu conducteur en suivi de trajectoire

Lisibilité de la route

Attributs géométriques de la trajectoire de consigne

Dynamique du véhicule

Sollicitation absolue

Le véhicule

La chaussée Le conducteur

Figure 2 : le risque dans l’approche embarquée

3.3. Rappel sur le modèle CALLAS

Figure 3 : les Degrés de Liberté de CALLAS

6 DoF Body at CoG:3 rectilinear xyz

3 rotationsRoll Pitch Yaw

For every Wheel :2 mechanical DoF1 Rolling Radius

1 Rotation

For every Wheel, optional DoF :Brake pressure transient

3 Tire Force Transient

Steering, 2 DoF :Steering Wheel angle

Steering at rack

Engine DoF :1 rotation

Optional Temperature DoF for:Tire (2 DoF for each)

Brake, damper, engine

������������ ����������������������������

7/30

Ici, dans un contexte de mesures embarquées, nous développerons la reconstruction du sol à partir de capteurs proprioceptifs, en calculant d’abord la position absolue de la caisse puis en « redescendant au niveau du sol » à l’aide des capteurs de suspension.

3.4. Méthode analytique Personne n’a proposé de méthode analytique correcte ou efficace pour la reconstruction du sol. Les raisons sont que:

• l’utilisation des corrélations de capteurs (vitesse/accélération longitudinales) est dangereuse parce qu’elle est aussi sensible au glissement et/ou au dérapage

• les formules complètes comportent de nombreux termes transitoires qui ne sont pas souvent négligeables

On distingue aussi :

• les méthodes directes fournissant directement une valeur d’angle (dont l’avantage est d’être stables mais nécessitent d’être en régime stabilisé)

• les méthodes itératives qui passent par un calcul de vitesse angulaire et une intégration (dont l’inconvénient est d’avoir une forte tendance à dériver)

3.5. Méthode d’hybridation de capteurs

Il existe des accéléromètres asservis qui prétendent ne pas confondre verticale terrestre et accélération apparente. Les solutions à base d’hybridation accéléromètre-gyromètre-magnétomètres semblent prometteuses mais insuffisamment mûres (quid de l’utilisation au voisinage d’une voiture avec un champ EM perturbé ?)

3.6. Méthode d’observateurs La solution est possible car elle est très puissante, mais il faut faire attention à ne pas avoir une méthode trop lourde si l’on veut rester en Temps Réel sur un processeur avec plusieurs process en parallèle. Voir aussi la mise en oeuvre en réel.

3.7. Conclusion sur les méthodes possibles Plusieurs solutions sont intéressantes, donc une répartition des méthodes est effectuée au sein du groupe de travail sur la méthode embarquée. SERA a choisi une méthode analytique (très économe en puissance de calcul) dont on peut espérer une mise au point à court terme, en version directe et itérative. D’autres partenaires (Nexyad, UTC, Germe) vont s’orienter vers d’autres choix. Objectif Principes Acteur Pente, Dévers Reconstruction analytique directe SERA Pente, Dévers Reconstruction analytique itérative SERA Pente, Dévers Observateur GERME Uni Observateur HEUDIASYC

8/30

4. LES DIFFERENTS CAPTEURS POSSIBLES Les applications à grand succès comme l’ESP de BOSCH sont en fait le fruit de la rencontre d’un concept innovant avec l’aboutissement de la fabrication en très grande série du capteur de vitesse de lacet (voir plus loin l’article américain sur les coûts qui pourraient être de 3$ pièce). Bosch a fêté le 10ème anniversaire de l'ESP (ou ESC) apparu en 1995 sur la Mercedes Classe S et produit à 15 millions d'exemplaires dans le monde, dont 4 millions en 2004 !

4.1. Capteurs possibles Capteurs envisagés pour les différents principes :

• Accéléromètres X, Y, Z au Centre de Gravité (en série en x et y) • Gyromètres X, Y, Z en série en z et bientôt en x • Débattements de suspension • Mesure de vitesse (pour le passage du temps à la distance)

Capteurs ayant un coût acceptable, mais non prévus en série :

• Ultrasons de hauteur de caisse • Accéléromètres de roue : peu chers, proche de la route, mais la double intégration a

tendance à diverger Capteurs coûteux qu’il faut essayer de supprimer :

• Lasers de hauteur de caisse : trop exigeants sur l’acquisition • Centrale gyroscopique d’attitude (en français on distingue gyroscopique qui donne

l’angle et gyrométrique qui donne la vitesse, angular rate, l’anglais est ici moins précis)

• DGPS : la précision est insuffisante en Z et le coût est très élevé

9/30

4.2. Avantages et Inconvénients des différents capteurs possibles

capteur avantage inconvénient

Accéléromètre vertical fixé sur la masse non suspendue

Bonne dynamique mécanique de la masse non suspendue et du capteur, méthode absolue, capteur peu cher, en cours d'industrialisation massive dans l’automobile

La double intégration dérive beaucoup même avec la mise à zéro dès que l'on est à l'arrêt

Débattement de suspension

Capteur supposé existant pour la détermination du cas de charge : o potentiométrique linéaire ou angulaire o pression si suspension

hydropneumatique o ultrason (sans contact)

Méthode relative position caisse Dynamique moyenne

Accéléromètre vertical sur masse suspendue

Principe de position absolue Dynamique moyenne

Laser caisse/sol Bonne dynamique Variante ultrason à évaluer (performance/prix),

Reste cher, nécessite une acquisition haute fréquence

XYZ GPS Principe de position absolue Cher, fréquence faible, taux de couverture partiel

5ème roue Permet de résoudre l'indétermination pente adhérence longitudinale

Pas réaliste en série

Magnétomètre Permet de résoudre l'indétermination dévers adhérence latérale

Champs perturbés, taux de couverture?

Gyromètres de lacet et roulis

Capteur existant en lacet, possible en roulis Principe itératif qui dérive, même avec la mise à zéro dès que l'on est à l'arrêt

10/30

5. CAHIER DES CHARGES DE L‘APPLICATION « RECONSTRUCTION DU SOL »

1. Fonctionner avec des capteurs existants, d’un coût raisonnable et déjà sur le véhicule 307 de l’INRETS.

2. Pouvoir fonctionner largement en Temps Réel (une consommation de quelques % du CPU est idéale, mais quelques dizaines de % est toutefois acceptable).

3. Accepter une vitesse variable au niveau du véhicule, à cause des conditions réelles de mesure sur la route ; ceci implique de correctement modéliser les accélérations et les freinages (validité de l’angle de tangage) et de prendre en considération les effets anti-plongée et anti-cabrage.

4. Favoriser la robustesse (ne pas dériver, accepter les différents cas de charge,…) plutôt que la précision : 1% est un résultat correct pour nos objectifs.

5.1. Méthodologie de développement

• Test préliminaire avec méthode synthétique (simulation contre simulation)

• Rejeu de cas réels déjà enregistrés, par exemple des tests sur sol plan : la reconstruction doit trouver environ 0.

• Importer les fichiers de mesure PALLAS2 sur Callas

• Méthode INRETS

• Méthode SERA

• Faire rouler la 307 virtuelle sur le fichier PALLAS2 en suivi de trajectoire

1. Utilisation des données calculées

2. Utilisation des modèles de capteur pour la fabrication de mesures

• Faire rouler la 307 réelle sur la même route

5.2. Problèmes attendus

• Difficulté de synchronisation temps/distance à cause de la précision des mesures de timer et surtout de la vitesse

• Utilisation d’événements « remarquables » sur le parcours pour la resynchronisation

• Validation de la dynamique verticale du modèle CALLAS par rapport au paramétrage du constructeur.

11/30

6. RECHERCHE DES METHODES ANALYTIQUES POSSIBLES

6.1. Théorie simple du dévers Le roulis absolu (total absolute roll) est composé du dévers de la route (banking), du roulis de pneu (tires roll), et du roulis de suspension (suspension roll). Le roulis relatif a pour origine le débattement des suspensions (roulis suspension) et l’écrasement des pneus (roulis de pneu) Dévers = roulis absolu – roulis relatif Le dévers est positif lorsque le côté droit de la route est plus bas que le côté gauche de la route.

Figure 4 : représentation des différents roulis

L'accélération latérale est horizontale si le véhicule se déplace dans un plan horizontal, ce qui est le cas usuel même sur anneau très relevé. L'accéléromètre lit la projection de l'accélération apparente sur son axe, et c'est la somme des deux valeurs suivantes :

• l'accélération centrifuge (précisément la projection de l'accélération centrifuge sur l'horizontale caisse)

• la projection de la gravité sur la verticale caisse

A l’arrêt, un inclinomètre utilise la deuxième relation

Figure 5 : lecture du dévers par un inclinomètre

accélérationcentrifuge

grav

ité

accé

lérati

on ap

pare

nte

La vitesse de lacet est supposée autour d’un axe vertical, elle est mesurée par un gyromètre fixé à la caisse qui lit précisément : gyro_lacet = VitesseLacet*cos(Roulis_Abs)

Figure 6 : lecture de l’inclinaison de caisse (roulis) par l’accéléromètre

Roulisrelatif

g

Ayaxe de l'accéléromètre

12/30

6.2. Théorie simple de la pente

Le tangage absolu est la somme du tangage relatif et de la pente de la route. Le tangage relatif au sol a pour origine le débattement des suspensions et l’écrasement des pneus. Pente = Tangage relatif – Tangage absolu (inversion du signe par convention) La pente est positive lorsque la route monte et est négative lorsque la route descend. Par contre, le tangage relatif est positif lorsque la voiture pique (lors d’un freinage par exemple), et est négatif lorsque la voiture est cabrée (lors d’une accélération par exemple). De même, pour le tangage absolu, si la voiture pique par rapport au repère absolu, il est positif, et est négatif dans le cas contraire. L'accélération longitudinale est comme la vitesse, parallèle au sol. L'accéléromètre lit la projection de l'accélération apparente sur son axe, et c'est est la somme de :

• l'accélération longitudinale (précisément la projection de l'accélération longitudinale sur l'horizontale caisse)

• la projection de la gravité sur la verticale caisse

En l’absence de mouvement, un inclinomètre utilise la deuxième relation. (un inclinomètre est un accéléromètre utilisé à l'arrêt).

Tangageabsolu

Penteroute

axe de l'accéléromètre

g

Ax

TangageRelatif

Acco_x

Figure 7 : incidence du tangage sur l’accéléromètre

longitudinal

6.3. Algorithmes analytiques possibles

6.3.1. Terminologie On appelle :

• « Itératif » une méthode qui détermine une pente et dévers par (au moins) une intégration • « Directe » une méthode qui détermine directement les angles de pente et dévers sans

intégration. • Absolu : dans le repère terrestre • Relatif : relatif entre caisse et sol

6.3.2. Avantage et inconvénients a priori des méthodes Il est évident qu’il est intéressant de grouper les deux méthodes « directe » et « itérative » du fait de leur complémentarité :

• l’itératif marche toujours mais dérive • le taux de couverture de la méthode directe est incomplet (cette méthode n’est légitime que

lorsque certaines conditions sont respectées).

13/30

6.3.3. Les différentes méthodes de reconstruction du sol

Nom Principe de la méthode Relative caisse Itérative

Détermination absolue des angles caisse puis transport au sol ex : dévers = vitesse de roulis intégré – roulis relatif par lasers

Absolue Itérative

Calcul des z terrestres des centres de roue Z’’mns : double intégration avec washout rapide

Relative caisse Absolue

Calcul de la position absolue de la caisse puis transport au sol position GPS hybridée gyromètres

Absolue Directe

Un GPS dans chaque roue !

Relative caisse Absolue

Inclinomètres dans la caisse, en roulage il faut hybrider les accéléromètres avec des magnétomètres

7. CONCLUSION Le recensement des capteurs existant, leur ventilation du plus souhaitable au moins souhaitable, et le recensement des différentes approches possibles ont été effectués. Les différents partenaires vont développer des approches différentes : Objectif Principe Acteur Pente, Dévers Reconstruction analytique directe SERA Pente, Dévers Reconstruction analytique itérative SERA Pente, Dévers Observateur GERME Uni Observateur HEUDIASYC Pour sa part SERA a commencé de développer une méthode analytique simple basée sur des capteurs peu chers (cf ci après en ANNEXE les tendances actuelles dans l’industrie du capteur), qui donne des résultats intéressants. Cette méthode et les résultats seront présentés dans un livrable suivant.

14/30

8. ANNEXE : TENDANCES DANS L’INDUSTRIE DU CAPTEUR

8.1. Sensor industry developments and trends October 2003 : InvenSense's Technology Can Accelerate Gyro Market

Micromachined gyros are generating highly significant growth opportunities in diverse, high-volume applications, particularly automotive (e.g., vehicle stability control, rollover detection) and consumer (e.g., computer input devices, camcorder stabilization) applications. MEMS (microelectromechanical systems) technology is enabling the proliferation of lower-cost gyros into high-volume applications. Micromachined gyros leverage IC fabrication techniques to allow for making lower-cost, more compact, and more highly integrated gyros that are more cost-effective and convenient than traditional mechanical or optical gyros in large-scale, cost-sensitive applications.

However, largely due to design or packaging constraints, existing micromachined or MEMS gyros have not achieved the reduced cost point required to enable the gyro market to reach its optimal potential. The market for MEMS gyros would significantly expand, with the development and introduction of a high-performance gyro with integrated electronics that could profitably be sold for $10 or less per unit in high volumes.

InvenSense, Inc. (Saratoga, CA, 408-741-5443), incorporated in June 2003, is a pioneering MEMS start-up working on addressing the vital need in the marketplace for a very low-cost, readily manufacturable, yet high performing gyro. The company is spearheading the design and development of enhanced MEMS vibrating rate gyroscopes with integrated electronics that use an innovative electrostatic drive technique, bulk micromachining, and wafer-scale packaging to provide a significantly lower-cost, manufacturing-friendly, easily and cost-effectively packaged, yet high-performance MEMS gyro solution.

InvenSense's gyro technology-which leverages innovative MEMS actuation technology originally developed for MEMS optical mirrors, well-proven bulk silicon fabrication processes, as well as lower-cost packaging technology, and the benefits of integrated electronics-has key potential for enabling the realization of very low-cost gyros that will, in turn, drive and accelerate the wider use of gyros in large-volume, cost-sensitive applications.

InvenSense notes that micromachined gyroscopes are typically vibratory devices that use a vibrating mechanical elements (a proof mass) to sense rate of rotation. Vibratory gyros are based on the transfer of energy between the two vibration modes of a structure, due to Coriolis acceleration (which arises in a rotating reference frame and is proportional to the rate of rotation). Tuning fork gyros contain a pair of masses that are driven to oscillate with equal amplitude, but in opposite directions. When rotated, the Coriolis force creates an orthogonal vibration that can be sensed using various sensing techniques.

There are key challenges in achieving a low-cost, high-performance MEMS gyro that is not affected by key factors that can impact a MEMS gyro's performance, such as potential manufacturing variations, packaging (which typically accounts for a high portion of the overall cost for most types of MEMS sensors), linear acceleration, or temperature. Indeed, designing a MEMS gyro is inherently more challenging than designing a MEMS pressure sensor or accelerometer. The MEMS gyro is essentially two high-performance MEMS devices integrated into a single device: a self-tuned resonator in the drive axis; and a micro-g sensor in the sensing axis. The magnitude of the Coriolis force sensed in the MEMS gyro is orders of magnitude lower than that for a high-volume MEMS accelerometer. Capacitance technology is often used to measure the small Coriolis-induced forces or accelerations in the MEMS gyro.

15/30

InvenSense's MEMS capacitance, tuning fork gyro design includes: an integrated actuating mechanism driven electrostatically; an innovative dual-mass tuning fork configuration for the drive mode and a capacitance sensing mechanism for rate sensing; integration of the drive, sensing, and control electronics and a design allowing for low-cost packaging; the use of bulk silicon material, which allows for a larger proof mass and can be more cost-efficient to manufacture than a surface micromachined gyro; a robust out-of-phase resonating system; the ability to sense in the X and Y axes (which, compared to a Z-axis device, can facilitate placement of the gyro in the user's system and readily supports the design of an IMU (inertial measurement unit); insensitivity to linear acceleration/vibration; and built-in self-test and quadrature error tuning capabilities, as well as shock and over-travel protection for the pair of masses employed in the device.

InvenSense's gyro technology is designed to include all the cost-driven features at the wafer level to achieve very low-cost fabrication in high production volumes. InvenSense's MEMS vibrating gyro, moreover, incorporates sensitive analog electronic circuitry to enhance the sensor's performance, accommodates digital programmability and customization, has analog or digital output capability, on-chip self-diagnostics, an operating voltage of 8-16 V, an over-voltage capability of +36 V and reverse voltage protection, and is able to compensate for environmental influences that could cause sensor errors. Steven Nasiri, President, CEO, and founder of InvenSense, explained that the electrostatic actuation used in InvenSense's patent-pending gyro leverages technology that initially appeared in the patented Nasiri platform-a novel actuating mechanism and enabling technology for an advanced bulk silicon optical mirror array developed at Transparent Networks, where (prior to launching InvenSense) Nasiri was a co-founder and vice president of MEMS and packaging.

In the effort to create the optical switch product based on bulk silicon, Transparent Networks had to develop innovative technologies to create an integrated silicon chip with 1,200 mirrors. Wafer-scale integration facilitated the fabrication of the drive electronics using a standard semiconductor foundry and the integration of the drive electronics with bulk silicon MEMS wafers.

The Nasiri platform provided the actuation mechanism that enabled the fabrication of a 1,200 integrated mirror array with four actuators to be controlled with less than 120 connections and 2 watts of power for all 1,200 mirrors. The mirror array used vertically integrated electronics to control and drive all the mirrors. In contrast to MEMS mirrors that have a dual gimbal design, the Nasiri platform (which uses a virtual pivot and is able to perform angular amplification) allowed for steep drop in the drive voltage (from typically 400 V to less than 120 V), thereby enabling the integration of the sophisticated mirror array with drive electronics.

Nasiri is a 26 year veteran of the MEMS industry who has considerable expertise in MEMS design, fabrication, and packaging for manufacturability. He has spearheaded the design, development, and commercialization of over 50 MEMS product families; and served as managing director for the successful ISS-Nagano GmbH start-up in Germany, where he designed and developed innovative high-pressure sensor packages for high-volume automotive applications, such as MAP (manifold absolute pressure) sensing, common rail diesel fuel injection, gasoline direct injection, and vehicle dynamic control (measurement of pressure in the hydraulic braking system). (Integrated Sensor Solutions (ISS) was acquired by Texas Instruments in 1999.)

Nasiri (who has 10 patents issued and more than 12 pending) has been a co-founder and early-stage executive at several key Silicon Valley MEMS sensor companies, where he had served as vice president of operations and engineering, including SenSym (acquired by Honeywell in 2002), NovaSensor (acquired by GE from TRW in 2002), ISS, and IMST (acquired by Maxim Integrated Products). Nasiri began his career at National Semiconductor, where he developed the initial hybrid ceramic package for silicon pressure sensors.

InvenSense's gyro electrostatic actuation technique eliminates the need for external components (such as magnets), allowing for a streamlined design. The integration of the drive electronics and the capacitive sensing mechanism in a single unit makes the gyro invulnerable to any parasitic

16/30

effects. More importantly, InvenSense has achieved the ability to design an electrostatic drive that does not require high voltages yet can achieve high travel of the proof mass.

Gyros fabricated using bulk silicon can be more robust then surface micromachined gyros, since the former can accommodate a considerably larger mass. Bulk silicon micromachining technology is more mature than surface micromachining technology, and bulk micromachined devices, which can allow for more efficient use of the chip's real estate, can be less expensive to produce than surface micromachined devices. It can be inefficient to fabricate surface micromachined sensors with integrated electronics, resulting in lower yields. Moreover, there reportedly can be key cost issues with surface micromachined gyros, since polysilicon (used in surface micromachined gyros) can have yield issues and can be sensitive to, for example, the release and anti-stiction steps of the device fabrication process.

Nasiri explained that InvenSense's gyro technology is designed to enable applications for gyros that have been somewhat constrained because of the gyro's cost and size, to emerge, thrive, and proliferate. He noted that size is especially important for many consumer applications.

InvenSense's gyro technology has applicability for a wide range of markets/applications that would benefit from a lower-cost and small-size gyro. Initial target markets/applications for InvenSense's gyro technology include automotive applications (e.g., vehicle stability control/vehicle dynamic control (VSC/VDC), rollover detection/stability control, and navigation systems) and consumer applications (e.g., camcorder stabilization, computer input devices, handheld computing devices). Additional promising markets/applications over the near-term include camera or antenna stabilization/pointing (on, for example, a plane, helicopter, ship, train, or recreational vehicle), and providing precision guidance and control in land and off-road vehicles.

The VDC/VSC system is a safety system that controls the dynamic motion of the vehicle in emergency situations (e.g., when the car starts to skid). A VSC or VDC system essentially consists of a gyro (yaw rate) sensor, low-g accelerometer to determine whether the car is sliding laterally, and wheel speed sensors at each wheel. Wheel speed is measured, and the predicted yaw (or turn) rate of the vehicle is compared to that measured by the gyro. If the measured yaw rate differs from the computed yaw rate, or if lateral sliding is detected, single-wheel braking or torque reduction can be used to get the car back in line.

More advanced VSC systems use signals to derive the driver's intent (such as steering wheel angle, brake pressure, engine torque) and signals to derive the actual motion of the vehicle (e.g., angular velocity of the car around it's vertical axis (yaw rate) and lateral acceleration) to control the braking of individual wheels to keep the vehicle on its intended path.

Rollover detection or stability control systems have not to date been implemented on many vehicles, but have clear opportunities for use in vehicles with a higher center of gravity (such as SUVs). As noted in SENSORS magazine (February 2002), in a rollover detection system, the gyro is used to read the roll rate of the vehicle; and the roll rate is integrated to determine the vehicle's roll angle. The system uses information about roll angle and roll rate to determine if the vehicle is tipping over. If the car is tipping over, a side curtain airbag is fired to protect the occupants. The gyro used in rollover detection does not need to have the resolution of the gyro used in VSC/VDC, since the roll rates are nearly half an order of magnitude greater. However, the gyro for rollover detection (as with the gyro used in VSC/VDC) must be able handle high shock and vibration.

Rollover detection system can also require an accelerometer to measure vertical (Z axis) acceleration to safeguard against the rollover detection system interpreting a banked curve as a rollover event. Measuring lateral (Y axis) acceleration can also be desirable, since the probability of a rollover greatly increases if a car sliding sideways encounters an obstruction.

17/30

Moreover, rollover stability control systems, which use a gyro to detect vehicle body roll and can provide information used to apply braking to correct the vehicle's trajectory, are beginning to appear in production vehicles.

The penetration of VDC/VSC systems is currently higher than that of rollover detection/stability systems. VDC/VSC systems are expected to significantly increase over the next five years; and VDC/VSC systems will likely have somewhat greater growth opportunities than rollover systems, particularly over the near-term, as the VDC/VSC system is used to mitigate against rollover by adjusting the braking forces at each wheel to keep the vehicle on its intended course.

However, rollover detection/stability systems can play an important role in safety, since the VDC/VSC system can indicate the propensity for a rollover, but cannot directly predict a rollover. Rollover crashes typically occur after the vehicle leaves the road (e.g., trips over a curb or another object as it slides laterally).

There is an emerging trend toward using an inertial sensor cluster--consisting of at least one gyro (yaw rate sensor) and an accelerometer-for vehicle stability control. In the future, there are opportunities for using an inertial sensor cluster in the form of a six-degree-of-freedom IMU (inertial measurement unit) to serve multiple functions/systems (e.g., vehicle dynamic/stability control, rollover mitigation/stability control, navigation, etc.). In addition, inertial sensors (gyros and accelerometers) will have expanding opportunities for measuring the vehicle's dynamic behavior (such as the angular velocity of the X axis), as advanced, high-performance stabilizing systems or comfort/driver assistance systems are increasingly implemented in vehicles.

Vehicle dynamic control/vehicle stability control systems are expected to be increasingly used in vehicles, particularly as their cost decreases. According to CSM Worldwide (Farmington Hills, MI, 248-380-9000), in 2002, 730,000 light vehicles produced in North America were equipped with a stability control system. In 2003, the number of North American-produced light vehicles equipped with a stability control system is projected to reach 920,000, representing a growth rate of about 26% over 2002. In 2007, 2,400,000 light vehicles produced in North America are projected to have a stability control system, representing a compound annual growth rate during 2002-2007 of about 26.9%.

In 2002, 6,400,000 light vehicles produced in Europe were equipped with a stability control system. In 2003, 6,800,000 European-produced light vehicles are projected to be equipped with a stability control system. In 2007, 9,600,000 light vehicles produced in Europe are expected to have a stability control system, representing a compound annual growth rate during 2002-2007 of about 8.4%. InvenSense's low-cost gyro technology, moreover, could help further propel the adoption of gyro-equipped VDC/VSC systems in non-higher-end vehicles.

Key providers of VDC/VSC systems in North America include Robert Bosch (Stuttgart, Germany/Farmington Hills, MI)(the major provider of such systems); Continental Teves (Frankfurt, Germany/Auburn Hills, MI); TRW Automotive (Livonia, MI); and Delphi (headquartered in Troy, MI).

From 1995 through 2002, Bosch (which makes a silicon micromachined yaw rate and acceleration sensor cluster) produced 8 million ESP units. Bosch has developed a third-generation inertial measurement solution to address the latest ESP system requirements, as well as those of emerging or future high-performance vehicle stabilizing or comfort systems. The new inertial sensor cluster is designed to include a silicon micromachined angular velocity (yaw rate) sensor element, a silicon micromachined linear acceleration sensor element for measuring lateral and longitudinal acceleration, ASICs for the gyro and accelerometer, a microconcontroller, an SPI interface between the sensor elements and the microcontroller, and a CAN interface between the sensor cluster and the system's ECU.

18/30

Continental Teves produced about 2.5 million ESP systems in 2002, compared to about 2 million in 2001. In April 2003, Continental noted that, in '03, the output will be way above the 3 million mark.

Manufacturers of silicon micromachined gyro sensors include Bosch, Silicon Sensing Systems (a joint venture of Sumitomo Precision Products (Amagasaki, Japan) and BAE Systems (Portsmouth, England)); Analog Devices (whose Micromachined Products Division is located in Cambridge, MA), as well as SensoNor (Horten, Norway).

Moreover, BEI Technologies' Systron Donner Automotive Division (Concord, CA) is a major supplier of the GyroChip®, a micromachined piezoelectric quartz sensor, that is used in high volume as a yaw rate sensor in computer-controlled vehicle stability enhancement systems. BEI has reduced the thickness of the quartz MEMS sensor (which reduced etching time) and increased the number of forks per wafer, reducing the GyroChip's overall cost. Moreover, BEI supplies multi-axis inertial sensor clusters (consisting of a yaw sensor and an accelerometer, or two yaw rate sensors and two accelerometers). BEI Systron Donner Inertial Division (Concord, CA) provides inertial sensors (including rate of rotation and multi-axis sensor suites) for aviation, military, and aerospace applications.

InvenSense's mission is to develop the next-generation of lower-cost, high-performance gyros, which will use the company's MEMS integrated tuning vibrating gyro technology. The cost, performance, packaging, and manufacturability advantages offered by the company's technology can be leveraged to help expand the reach of gyro sensors into a significantly wider range of products and systems. InvenSense is dedicated to getting its gyro technology into production. To facilitate this objective, they are very amenable to partnering with a system company that will benefit from low-cost, high-performance MEMS gyro technology.

19/30

8.2. BOSCH Semiconductors GYROSCOPE

Figure 1. All MEMS gyroscopes take advantage of the Coriolis effect. In a reference frame rotating at angular velocity , a mass M moving with velocity v sees a force:

acor = 2V x �

Many types of MEMS gyroscopes have appeared in the literature, with most falling into the categories of tuning-fork gyros, oscillating wheels, Foucault pendulums, and wine glass resonators. Conventional (non-MEMS) spinning wheel gyros are common, but levitation and rotation of a MEMS device with no springs has not yet been commercialized.

Foucault Pendulum Gyroscopes. These devices are based on a vibrating rod that is typically oriented out of the plane of the chip [14,15]. They are therefore challenging to build with planar fabrication tools, but recent advances in MEMS technology allow very high aspect ratio MEMS that make it possible to fabricate the pendulum without hand assembly of the rod.

20/30

Tuning Fork Gyroscopes. Tuning fork gyros contain a pair of masses that are driven to oscillate with equal amplitude but in opposite directions. When rotated, the Coriolis force creates an orthogonal vibration that can be sensed by a variety of mechanisms. The Draper Lab gyro [7-9] shown in Figure 6 uses comb-type structures to drive the tuning fork into resonance.

Figure 6. The first working prototype of the Draper Lab comb drive tuning fork gyro is shown here in an SEM image. Due to the superior mechanical properties of single-crystal silicon, a much better performance was achieved using single-crystal silicon with the dissolved wafer process.

Rotation causes the proof masses to vibrate out of plane, and this motion is sensed capacitively with a custom CMOS ASIC. The technology has been licensed to Rockwell, Boeing, Honeywell, and others. The resonant modes of a MEMS inertial sensor are extremely important. In a gyro, there is typically a vibration mode that is driven and a second mode for output sensing. In some cases, the input and output modes are degenerate or nearly so. If the I/O modes are chosen such that they are separated by ~10%, the open-loop sensitivity will be increased due to the resonance effect. It is also critical that no other resonant modes be close to the I/O resonant frequencies. Samsung Corp. has put a large effort into inertial sensors for automotive and consumer electronics applications. Gyroscopic stabilization of camcorders has been shown to improve picture quality, and gyroscopes are enhancing vehicular safety in multiple ways [10]. With gyros costing as little as $10.00 per sensed axis, they should soon claim a sizeable market share. Vibrating-Wheel Gyroscopes. Many reports of vibrating-wheel gyros also have been published [4]. In this type of gyro, the wheel is driven to vibrate about its axis of symmetry, and rotation about either in-plane axis results in the wheel’s tilting, a change that can be detected with capacitive electrodes under the wheel (see Figure 7).

.

Figure 7. The vibrating wheel gyro made by Bosch Corp. [10], with capacitive sensing under the wheel, can be used to detect two in-plane rotational axes

21/30

It is possible to sense two axes of rotation with a single vibrating wheel. A surface micromachined polysilicon vibrating wheel gyro (see Figure 8) has been designed at the U.C. Berkeley Sensors and Actuators Center.

Figure 8. This polysilicon surface-micromachined vibrating wheel gyro was designed at the Berkeley Sensors and Actuators Center. The potential for combining the mechanical resonator and sense and drive electronics on a single chip permits extreme miniaturization.

Wine Glass Resonator Gyroscopes. A third type of gyro is the wine glass resonator. Fabricated from fused silica, this device is also known as a hemispherical resonant gyro. Researchers at the University of Michigan have fabricated resonant-ring gyros in planar form [11]. In a wine glass gyro, the resonant ring is driven to resonance and the positions of the nodal points indicate the rotation angle. The input and output modes are nominally degenerate, but due to imperfect machining some tuning is required. The output signal is monitored and nulled, yielding a rate gyro. Silicon Sensing Systems, a joint venture between Sumitomo and British Aerospace, has brought to market an electromagnetically driven and sensed MEMS gyro (see Figures 9 and 10). A permanent magnet sits above the MEMS device. Current passing through the conducting legs creates a force that resonates the ring. This Coriolis-induced ring motion is detected by induced voltages as the legs cut the magnetic field.

Figure 9. The Silicon Sensing Systems gyro is fabricated from single-crystal silicon with metal added for higher conductivity. This device measures 29 by 29 by 18 mm and is used to stabilize the Segway Human Transporter.

Analog Devices has been working on MEMS gyros for many years, and has patented [12,13] several concepts based on modified tuning forks.

22/30

Figure 10. The resonant ring at the heart of the Silicon Sensing Systems gyro is shown here as an SEM image. Both electromagnetic drive and sensing are accomplished with a permanent magnet in the center (not shown).

The company has recently introduced the ADXRS family of integrated angular rate-sensing gyros, in which the mass is tethered to a polysilicon frame that allows it to resonate in only one direction. Capacitive silicon sensing elements interdigitated with stationary silicon beams attached to the substrate measure the Coriolis-induced displacement of the resonating mass and its frame (see Figure 11).

Figure 11. The iMEMS ADXRS angular rate-sensing gyro from Analog Devices integrates an angular rate sensor and signal processing electronics onto a single piece of silicon. Based on the Coriolis effect, its very low noise output makes it a good choice for GPS receivers, where critical location information is required during temporary disruptions of GPS signals.

Summary MEMS inertial sensing is an established industry, with performance-to-cost rapidly improving each year. Gyroscopes and angular accelerometers are entering the marketplace and will soon make many non-MEMS components obsolete. They should also open up new applications due to their small size and weight, modest power consumption and cost, and high reliability.

23/30

References 1. N. Yazdi, F. Ayazi, and K. Najafi. Aug. 1998. “Micromachined Inertial Sensors,” Proc IEEE, Vol. 86, No. 8. 2. T.B. Gabrielson. May 1993. “Mechanical-thermal noise in micromachined acoustic and vibration sensors,” IEEE Trans, Electron. Devices, Vol. 40:903-909. 3. J. Bernstein et al. 8 June 1998. “Low Noise MEMS Vibration Sensor for Geophysical Applications,” Proc 1998 Solid State Sensor and Actuator Workshop, Hilton Head Island, SC:55-58. Extended version also in IEEE JMEMS Dec. 1999. 4. M. Lutz et al. June 1997. “A precision yaw rate sensor in sillicon micromachining,” Tech Dig 9th Intl. Conf Solid State Sensors and Actuators (Transducers ’97), Chicago, IL:847-850. 5. Y. Gianchandani and K. Najafi. June 1992. “A bulk silicon dissolved wafer process for microelectromechanical systems,” J Microelectromech Syst:77-85. 6. GS-14 geophone. Geospace Inc. 7334 N. Gessner Rd., Houston, TX 77040. 7. J. Bernstein et al. Feb. 1993. “A micromachined comb-drive tuning fork rate gyroscope,” Proc IEEE Micro Electro Mechanical Systems Workshop (MEMS ’93), Fort Lauderdale, FL:143-148. 8. J.J. Bernstein and M.S. Weinberg. 5 March 1996. U.S. Patent #5,496,436, “Comb-Drive Micromechanical Tuning Fork Gyro Fabrication Method.” 9. J.J. Bernstein and M.S. Weinberg. U.S. Patent #5,349,855, “Comb-Drive Micromechanical Tuning Fork Gyro.” 10. C. Song. June 1997. “Commercial vision of silicon based inertial sensors,” Tech Dig 9th Intl. Conf Solid State Sensors and Actuators. 11. M.W. Putty. March 1995. “A micromachined vibrating ring gyroscope,” Ph.D. dissertation, University of Michigan, Ann Arbor, MI. 12. J. Geen. 3 June 1997. US Patent #5,635,638, “Coupling for multiple masses in a micromachined device.” 13. J. Geen. 3 June 1997. U.S. Patent #5,635,640, “Micromachined device with rotationally vibrated masses.” 14. J.J. Bernstein. U.S. Patent #5,203,208, “Symmetrical Micromechanical Gyroscope.” 15. T.K. Tang et al. 1997. “A packaged silicon MEMS vibratory gyroscope for micro-spacecraft,” Proc IEEE Micro Electro Mechanical Systems Workshop (MEMS ’97), Japan:500-505. A Critical Review of MEMS Gyroscopes Technology Steven Nasiri, 408-988-7339 x108, snasiri@invensense.com 3150A Coronado Drive, Santa Clara, California 95054

24/30

8.3. A Critical Review of MEMS Gyroscopes Technology

Abstract Gyroscopes are expected to become the next “killer” application for the MEMS industry in the coming years. A multitude of applications already have been developed for consumer and automotive markets. Some of the more well known automotive applications such as vehicle stability control, navigation assist, and roll over detection are only used in high-end cars, where cost is not a major factor. Examples of consumer applications are 3D input devices, robotics, platform stability, camcorder stabilization, virtual reality, and more. Primarily due to cost and the size most of these applications have not reached any significant volume. This paper provides a top-level review of various vibrating mass gyroscopes and examines the technology and packaging methodology for top four commercially available gyroscopes companies. Advancement in MEMS technology, fueled by the optical bubble, such as, wafer-scale-integration, and wafer-scale-packaging will be reviewed. New opportunities for design and development of the next generation of low-cost and high-performance gyroscopes based on the latest MEMS technologies are discussed. Introduction Micromachined inertial sensors, consisting of accelerometers and gyroscopes, are one of the most important types of siliconbased sensors. Microaccelerometers alone have the second largest sales volume after pressure sensors. It is believed that gyroscopes will soon be mass-produced at similar volumes once manufacturers are able to meet a $10 price target. Applications for gyroscopes are very broad. Some example for these applications are; automotive; vehicle stability control, rollover detection, navigation, load leveling/suspension control, event recording, collision avoidance; consumers, computer input devices, handheld computing devices, game controllers, virtual reality gear, sports equipment, camcorders, robots; industrial., navigation of autonomous (robotic) guided vehicles, motion control of hydraulic equipment or robots, platform stabilization of heavy machinery, human transporters, yaw rate control of wind-power plants; aerospace/military; platform stabilization of avionics, stabilization of pointing systems for antennas, unmanned air vehicles, or land vehicles, inertial measurement units for inertial navigation, and many more. This paper presents a review of silicon MEMS gyroscopes (rate sensors), their production status, and challenges towards fabrication of the next generation of low-cost gyroscopes. Following a brief introduction to gyroscope operating principles and performance specifications, the present status in the commercialization of micromachined rate sensors are discussed. Inertial sensors have seen a steady improvement in their performance and their fabrication technology, and today, microaccelerometers are among the highest volume MEMS sensors for the automotive. While the performance of gyroscopes has improved by a factor of 10 every two years, their costs have not dropped as was originally predicted. The initial drive for lower cost, greater functionality, higher levels of integration, and higher volume had slowed down during the optical bubble, when the sensor market was over taken with high potential returns promised by the telecom market. Although the telecom boom had slowed the wide spread development in gyroscopes, it poured billions of dollars into development of next generation MEMS technologies, equipment, modeling tools, foundries, and micromachine experts. This paper will discuss some of these advancements in MEMS development and their potential use in the creation of the next generation of advanced, integrated MEMS gyroscopes that can meet the market cost expectations, and further their performance. Micromachined Gyroscope Technology Operating Principles and Specifications Almost all reported micromachined gyroscopes use vibrating mechanical elements (proof-mass) to sense rotation. They have no rotating parts that require bearings, and hence they can be easily

25/30

miniaturized and batch fabricated using micromachining techniques. All vibratory gyroscopes are based on the transfer of energy between two vibration modes of a structure caused by Coriolis acceleration. Coriolis acceleration, named after the French scientist and engineer G. G. de Coriolis (1792–1843), is an apparent acceleration that arises in a rotating reference frame and is proportional to the rate of rotation, Figure 1. Vibratory gyroscopes were demonstrated in the early 1980’s. An example of this type of devices is quartz tuning forks like the Quartz Rate Sensor by Systron Donner. Although quartz vibratory gyroscopes can yield very high quality factors at atmospheric pressure with improved level of performance, due to use of quartz as the primary material, their batch processing is not compatible with IC fabrication technology. In the late 1980’s, after successful demonstration of batchfabricated silicon accelerometers, some efforts were initiated to replace quartz with silicon in micromachined vibratory gyroscopes. Charles Stark Draper Laboratory demonstrated one of the first batch fabricated silicon micromachined rate gyroscopes in 1991. In general, gyroscopes can be classified into three different categories based on their performance: inertialgrade, tactical-grade, and rate-grade devices. Over the past decade, much of the effort in developing micromachined silicon gyroscopes has concentrated on “rate-grade” devices, primarily because of their use in automotive applications. Automotive applications generally requires a full-scale range of at least 50°-300° /s and a resolution of about 0.5°-0.05 °/s in a bandwidth of less than 100 Hz depending on the application. The operating temperature is in the range from -40 to 85° C. Tuning fork gyroscopes contain a pair of masses that are driven to oscillate with equal amplitude but in opposite directions. When rotated, the Coriolis force creates an orthogonal vibration that can be sensed by a variety of mechanisms. The Draper Lab gyro uses comb-type structures to drive the tuning fork into resonance, and rotation about either in-plane axis results in the moving masses to lift, a change that can be detected with capacitive electrodes under the mass. Vibrating-Wheel Gyroscopes have a wheel that is driven to vibrate about its axis of symmetry, and rotation about either in-plane axis results in the wheel’s tilting, a change that can be detected with capacitive electrodes under the wheel. It is possible to sense two axes of rotation with a single vibrating wheel. A surface micromachined polysilicon vibrating wheel gyro, has been designed at the U.C. Berkeley Sensors and Actuators Center that demonstrated this capability. Wine Glass Resonator Gyroscopes. A third type of gyro is the wine glass resonator. Fabricated from fused silica, this device is also known as a hemispherical resonant gyro. Researchers at the University of Michigan have fabricated resonant-ring gyros in planar form. In a wine glass gyro, the resonant ring is driven to resonance and the positions of the nodal points indicate the rotation angle. The input and output modes are nominally degenerate, but due to imperfect machining some tuning is required. Challenges with Designing a MEMS Gyroscope This section reviews some of the different choices that can be made for design of a vibrating gyroscope. Table 2 provides a summary of various design choices possible for a vibrating gyroscope, with over 2500 potential combinations. Most of research and development activities at the university level, with the University of California Berkeley being the most active, have been on surface micromachined gyros. One of the primary drivers for this has been DARPA and military interest for development of a single chip integrated six-axis inertial measurement units (IMU). Gyroscopes are much more challenging sensor products than acceleration or pressure sensors. Gyroscopes are basically two high performing MEMS devices integrated into one single device that have to work together to produce results. They are a selftuned resonator in the drive axis, and a micro-g sensor in the sensing axis. The absolute magnitude of the Coriolis force sensed is orders of magnitude lower than any high volume production MEMS accelerometer.

26/30

Capacitive sensors are generally used for measuring these small changes of capacitance. Gyroscope performance is very sensitive to all potential manufacturing variations, packaging, linear acceleration, temperature, etc. To achieve high and low-cost performance, lot of care must be taken during the initial design. Gyroscope designers must achieve a solution that can be insensitive to most of these potential variations.

Table 2: Shows the various options, >2500, that could be used for designing a gyroscope.

Gyroscope Packaging Challenges One of the most difficult decisions that can have the biggest effect on the cost is the choice for the final package. Generally, packaging is one of the highest components of the final cost for most types of MEMS sensors. In majority of cases, sensor designers and MEMS experts are not packaging experts. MEMS designers are primarily focused on the design and development of the sensor element, with the objective of demonstrating performance on the bench. The task of taking the MEMS sensor element and package is the packaging engineer’s problem. In order to have the lowest cost MEMS product packaging issues must be addressed up front in the initial phase of design cycle. MEMS by definition is a mechanical feature that can be manufactured in batch processing with little cost differential. It is very unlikely that a low-cost solution can be realized without addressing packaging issues properly on the outset. Path to High-Performance and Low-Cost Gyroscopes The main challenge for the MEMS gyroscope industry has been achieving high-performance and low-cost MEMS solutions. Although there are several high-performance MEMS gyroscopes in production already, they are still fairly costly for most applications. Today’s automotive gyroscopes cost more than $60 for more demanding applications like vehicle dynamic control, and navigation assist, and are slightly below than that for less demanding ones like roll-over detection. The automotive market has developed many safety and comfort related applications that relies on rate-sensors. Today they are only offered on the high-end cars, where cost is less of a concern. There are also many consumer applications that are waiting for smaller and more affordable gyroscopes, like input devices, camcorder stability control, and more. To benefit from all these high volume applications, a smaller size and lower cost gyroscopes are needed. A quick review of the price curve for MEMS accelerometers shows that high volume occurred once unit prices achieved $3 target prices. The initial manufacturers of these type of sensors, for the first few years were Lucas-Nova Sensors, and EG&G-IC-sensors. In spite of millions of dollars invested in automation, capacity build up, and quality control systems, they were only able to reduce their cost to just under $10, and could not meet market

27/30

demand for $3. Analog, Motorola, and few other companies recognized the market needs and focused their development from the beginning on fabrication of low cost accelerometers. They succeeded in the design and development of a new generation of accelerometers products that broke the price barrier. This allowed the market for these products to reach its full potential. All the original companies that failed to recognize the need were forced out of the market and are no longer producing accelerometers. There is a real market need for low-cost gyroscopes not unlike the accelerometers a decade ago. Once the cost targets are met this family of MEMS sensor will enjoy a drastic increase in volume and will become the next high-volume application for the MEMS industry. Vibrating Gyroscope Production Status This section provides some insight into the gyroscope designs, performance, and packaging of the top four vibrating gyroscope manufacturers worldwide, representing more than 95% of unit volumes shipped for this class of sensors. Robert Bosch has been the most active in design and fabrication of silicon vibratory gyroscopes. They have 50% of the gyro market for the automotive VDC, and related applications today. They are producing in high volume, with several million units shipped already. Bosch has developed both Z axis and X./Y axis rate sensors. Its Z-axis design, shown in figure 6, was introduced in 1998 and uses electromagnetic drive with capacitive sensing. The X/Y-axis gyro is a rotary vibrating mass, which was briefly discussed earlier. The MEMS sensor element along with its custom ASIC and all the discrete component are packaged in a hermetically sealed metal can, figure 7 and 8 which is then place inside its automotive style plastic housing, figure 9, with integral connectors and mounting brackets.

Figure 5; Shows automotive accelerometers price elasticity Figure 6; Bosch silicon dual mass tuning fork design, with Z-axis rate sense, in plan vibration and in plan sensing Figure 7; Bosch MEMS sense element along with all supporting electronics, in a metal header package Figure 8; Bosch metal header cross section, showing the permanent magnet suspended over the sensor chip. Figure 9; Bosch’s gyroscope package with integral connector and mounting brackets BEI Systron Donner is a major manufacturer of rate sensors for automotive. Their gyroscope is designed based on using one piece quartz inertial sensor. These micromachined inertial sensing elements measure angular rotational velocity, using tuning fork vibratory principals and piezoelectric actuation and sensing. These sensors generate a signal output proportional to the rate of rotation sensed. Each sensor element is packaged in a hermetically sealed metal headers, which are then combined with discrete electronic to produce the finished modules products.

28/30

Silicon Sensing Systems a joint venture between Sumitomo and British Aerospace has brought to market an electromagnetically driven and sensed MEMS gyro, with a permanent magnet sits above the MEMS device. Current passing through the conducting legs creates a force that resonates the ring. This Coriolis-induced ring motion is detected by induced voltages as the legs cut the magnetic field. Analog Devices has been working on MEMS gyros for many years, and has patented several concepts based on modified tuning forks. The company has recently introduced the ADXRS family of integrated angular rate-sensing gyros, in which the mass is tethered to a polysilicon frame that allows it to resonate in only one direction. Capacitive silicon sensing elements interdigitated with stationary silicon beams attached to the substrate measure the Coriolis-induced displacement of the resonating mass and its frame. Summary of State of Production During the telecom bubble, optical MEMS took the spot light away from all other MEMS sensor developments including gyroscopes. Most MEMS sensor companies had pushed aside most of their sensor related development in favor of designing the next best optical components. MEMS mirrors were in the forefront of all these activities. With Nortel paying over 3.0 Billion dollars for Xros mirrors, it was logical for all MEMS companies and MEMS experts to move into the optical MEMS. Although, the telecom market did not materialize and most all MEMS optical companies are things of the past, venture capital firms and corporate telecom companies had invested billions of dollars in the MEMS industry. These investment dollars resulted in a more mature industry, with better infrastructure, more experts, and more advanced technologies. MEMS telecom companies were all trying hard to outperform their competition and be the first in the market with the most advanced product. They were generally very protective of their IP and did not publish too many details on their design and development activities. Silicon bulk micromachining has benefited the most from these investment dollars. Many companies focused their development efforts on building bulk silicon mirrors. Transparent Networks (TNI) was one such company. Started in 2000, TNI raised $30M dollars to develop high port count, 1000x1000, optical switch product based on bulk silicon. TNI had to develop many new technologies for its integrated single silicon chip with 1200 mirrors. One of these technologies was wafer-scale-integration, where all the drive electronics were fabricated using a standard semiconductor foundry, and then they were integrated with the fabricated bulk silicon MEMS wafers. This enabled the fabrication of a 1200 integrated mirror arrays with four actuators each, to be controlled with less than 120 connections and 2 watts of power. The key enabling technology was the mirror actuator mechanism called the “Nasiri-Platform.” One of the main design features of Nasiri-platform is its virtual-pivot and its ability to do angular amplification. While almost all MEMS mirrors where designed using dual gimbals design, by using the Nasiri-Platform, TNI was able to drop the drive voltages from typically 400V to less than 120 volts, hence enabling electronic integration. The TNI mirror design has many features that are similar to vibrating gyroscopes such as, mechanical actuators, sensing capacitance, and the need for integrated electronics. Bulk silicon micromachine industry had never seen this level of sophistication in its 30+ years of history. Wafer-scale-packaging is another technology that has benefited from the optical boom. The most simple example of wafer-scale packaging is addition of the protective cap over the active MEMS component that can provide the required sealing for further packaging and handling. Number of companies is already using this technology in production.

29/30

Conclusion Gyroscope is the next killer application for the MEMS sensor industry. There are many mature applications already developed and produce in limited volumes in automotive, consumer, industrial, medical, and the military market. Many high volume applications, in excess of 100 million per year, are waiting for the availability of a low-cost gyroscope. Today’s gyroscope products are not designed to address the growing cost demand, creating an opportunity for the new generation of gyroscopes. The new generation designs are primed to take advantage of all the latest development in the MEMS industry. Bulk silicon can be a good low-cost solution for gyroscopes, due to its maturity and inherent higher mass, if the new design can eliminate the need for the use of magnets, metal can packaging, and allow for integrated electronics. Even more savings can be possible if standard off-the-shelf plastic packaging with lead frames can be utilized for the final assembly. About the Author Mr. Nasiri is a 26 years veteran in the MEMS industry with expertise in MEMS design, Fabrication and Packaging, presently the President and CEO of InvenSense Inc, is working on the design and development of next generation vibrating. He has 10 patents issued and over 12 pending. References 1. N. Yazdi, F. Ayazi, and K. Najafi. Aug. 1998. “Micromachined Inertial Sensors,” Proc IEEE, Vol. 86, No. 8. 2. Jonathan Bernstein, Corning-IntelliSense Corp , MEMS inertial sensing technology 3. S.Nasiri, Wafer Level Packaging of MOEMS Solves Manufacturability Challenges In Optical Cross Connect, 2003, 4. S.Nasiri patent, MEMS Mirrors and MEMS arrays, Pat# 6,480,320 5. T.B. Gabrielson. May 1993. “Mechanical-thermal noise in micromachined acoustic and vibration sensors,” IEEE Trans, Electron. Devices, Vol. 40:903-909. 6. J. Bernstein et al. 8 June 1998. “Low Noise MEMS Vibration Sensor for Geophysical Applications,” Proc 1998 Solid State Sensor and Actuator Workshop, Hilton Head Island, SC:55-58. Extended version also in IEEE JMEMS Dec. 1999. 7. M. Lutz et al. June 1997. “A precision yaw rate sensor in sillicon micromachining,” Tech Dig 9th Intl. Conf Solid State Sensors and Actuators (Transducers ’97), Chicago, IL:847-850. 8. Y. Gianchandani and K. Najafi. June 1992. “A bulk silicon dissolved wafer process for microelectromechanical systems,” J Microelectromech Syst:77-85. 9. J. Bernstein et al. Feb. 1993. “A micromachined comb-drive tuning fork rate gyroscope,” Proc IEEE Micro Electro Mechanical Systems Workshop (MEMS ’93), Fort Lauderdale, FL:143-148. 10. J.J. Bernstein and M.S. Weinberg. 5 March 1996. U.S. Patent #5,496,436, “Comb-Drive Micromechanical Tuning Fork Gyro Fabrication Method.”

30/30

8.4. Capteur CROSSBOW