Skip to main content

MEMS IMU for long-term position stability and high precision rate monitoring

Periodic Report Summary - MILEPOST (MEMS IMU for long-term position stability and high precision rate monitoring)

Project context and objectives:

The consortium of three small and medium-sized entreprises (SME)and two RTD partners (SensorDynamics Graz, Mems Foundry Itzehoe, Xsens Enschede, Fraunhofer Institute for Silicon Technology Itzehoe and Consorzio Pisa Ricerche Pisa) develop a novel movement sensors (acceleration, angular rate) within the EU program 'Research for small and medium sized enterprises'. These Inertial Sensors will work significantly more precise than the sensors available up to now and will facilitate new fields of application. These movement sensors are manufactured by methods of silicon microsystem technology. Several of these sensors are combined to Inertial Measurement Units (IMUs). By means of these data it is possible to determine whereto and how fast an object is moved within space and, of course, at which place it is located. Nowadays, IMUs have already been applied in automobiles for electronic stabilization as well as in smartphones for positioning. Provided that the sensors work sufficiently precise navigation systems can be built up that will function also in the case they do not receive a GPS signal like in narrow street canyons, in tunnels as well within buildings. However, the quality of the sensors available today is not enough for these applications.

In particular, the long-term signal drift, the signal noise, the temperature dependence and the precision in data recording have to be improved considerably. The parameters mentioned above will be investigated within the EU project MILEPOST by the consortium partners and optimized for the intended application. The aim is the development of novel sensor designs, improved electrical connections to control the sensors, and the generation of innovative algorithms for data evaluation. Within the project time until June 2013 an IMU should be realized working in all directional axes with long-time position stability and high-precision rate recording on a silicon chip activated and read out by a single electronic circuit. The planned cost will be about 1.4 million EUR. The participating companies have excellent knowledge in system design and system integration of inertial sensors (SensorDynamics - SD / AT, Graz), a wide application background (Xsens / NL, Enschede), and deep experience in product realization (MEMS Foundry Itzehoe - MFI / DE, Itzehoe). The research partners in the project are Fraunhofer Institute for Silicon Technology ( ISIT / DE Itzehoe) for the field of sensor development by microsystem technology, and Consorzio Pisa Ricerche - (CPR, IT / Pisa) for the field of circuit development.

The project is structured in seven work packages:
WP1: Management
WP2: Device specification
WP3: Sensor development
WP4: ASIC development
WP5: Packaging, module characterisation and reliability
WP6: Demonstration (not applicable in this period)
WP7: Dissemination and exploitation

The report for project month 1-12 is about project phase 1: 1D gyrometer. The work within the WPs is focused on that item.

The project is supported by funds of the European Union's Seventh Framework Programme, organized by REA-Research Agency FP7/207-2013 under Grant Agreement Nr 262118.

Project Results:

The Device Specification started with the definition of an 'application matrix' summarizing the application specific performance requirements. The application matrix was then translated in device performance specifications. The device specification was divided into 2 different grades: Mid-grade (current generation) and High-grade (next generation). For several categories the respective requirements at the MEMS transducers as well as the signal conditioning chain were defined.

For the Sensor Development first of all preliminary analytical and numerical considerations are made. A set of dynamical equations for the primary motion has been developed and also the transferfunction for the secondary motion. In detail the parasitic effects impacting the sensing has been explored with particular attention to the sensor packaging. The bias stability was analysed with respect to the impact of design and the impact of different process parameters. An FEM model in Ansys has been established for sensor simulation. Based on this information new sensor designs could be realized which are expected to have by factor of 4 better bias stability. The same model is used to improve the wafer process with respect to the Quadbias effect and parasitic capacities. A further reduction could be achieved by using through silicon vias (TSV) for the connection path of the sensor element to the ASIC. Therefore a TSV fabrication process has been developed to be integrated into the sensor fabrication process. In addition a verification of the manufacturability of the MEMS process for sensor fabrication has been done.

In the work package 4 'ASIC Development' the Design and development of a signal processing unit for interfacing the sensors developed in the framework of WP3 has been done. The signal processing and conditioning will compensate and filter the relevant signals in order to improve the signal-to-noise ratio and the long-term stability and, after digital conversion, transmit them through a serial interface to a host unit. In order to successfully reach the desired objectives, doubtless a trade-off between performance optimization and system cost and complexity (area, power consumption, testing, design time) was needed; the roadmap to follow was to firstly perform a modeling activity to find the bottlenecks of current architecture for ad-hoc ASIC optimizations thus enhancing stability and low-noise performance with limited complexity/cost overheads. A deep analysis of the low-noise/high-stability MEMS gyro state-of-the-art was performed.

Starting from the previous considerations, four different ways were explored:
- RLS adaptive demodulation
- Mixed-signal readout interfaces with analog-CDMA
- Two stage simplified Kalman filter
- Chopper stabilized bandgap

In the work package 5 'Packaging, module characterization and reliability' consisted in the complete characterization of the bias stable one dimensional gyroscope introduced in work package 3. In addition to the complete characterization of the module, an in-depth numerical simulation on the utilized QFN-package was performed, in order to obtain a deeper in-sight into the different mechanical and thermo-mechanical properties of such a package and their impact on the module performance. In order to evaluate the behavior of the module over temperature a series of different testing were performed. The impact of the different noise sources and the maximum available bias stability are being determined by the root Allan Variance measurements. A lot of modules were examined. The best investigated modules showed a bias stability of less than 1.65 °/h.

In work package 7 'Dissemination and exploitation' the project website has been established and for presentation of the work in project phase 2 and 3 different conferences and journals has been identified. The project website is located at .

Potential Impact:

In agreement between all partners it has been decided to stop the project on 31.12.2011. The reason for that is the decision of project partner SensorDynamics (SD) to withdraw from the project to the above mentioned date. This is caused by the fact that SensorDynamics was acquired by Maxim Integrated Products Inc.. Due to business restructuring and group integration activities, SD will no longer be able to paricipate in consortium member activities of the MILEPOST project. Due to the fact that SDs participation is essential for the progress and success in the project and a comparable partner was not available the project work has to be stopped. The expected final results will not be achieved.

List of Websites: