Institute of

Cognitive Integrated Sensor Systems

Prof. Dr.-Ing. Andreas König

In the context of the German Federal research ministry BMBF, in the microsystems programme, sub programme autonomous, networked sensor systems, mst-avs, the programme PAC4PT (Process Integrated, Autonomous Measurement Units for Process Technology, German title: Prozessintegriertes autonomes Überwachungssystem für die Verfahrenstechnik auf Basis vernetzter, multifunktionaler MST-Funksensoren) with six sub projects was started in 2009 with the overall goal to provide a novel system for in-situ distributed process measurements and logging of relevant process data and location/time context. The sensor modules of the swarm was envisioned in a proprietary microsystem integration technology (RMPD 3D-CSP of microTEC). From several options, the fermentation tank monitoring was chosen from the field of beer brewing industry (Warsteiner). The sub project ROSIG of TUKL/ISE contributes to sensor node and system architecture design and implementation, related signal processing, electronics design, and functional test of components and system as well as multivariate distributed data visualization. In particular, the providing of location context as well as self-x properties in the nodes are pursued. As already anticipated in the ROSIG-proposal, state-of-the-art RF-based sensor node localization concepts/practice in wireless sensor networks cannot be adequately transferred to the regarded application in stainless steel containers, due to communication problems by refelctions from container hull and attenuation due to liquid media. Thus, from existing tracking technology, an approach for localization based on magnetic field generation and sensing by MEMS AMR-sensors was adapted, validated and is currently refined (Patent application pending, filing date:18.05.2012). The following picture gives a rough sketch of the overall PAC4PT-project (pdf with higher resolution):

The task of ISE, TU Kaiserslautern, was the development of a hardware/software architecture of the swarm's sensor nodes, which included the choice of components, circuits, and prototyping of the system (wire-wrap PCB prototype) before aspired MEMS integration in 3D-CSP technology of microTec. Additionally, the magnetic localization scheme, based on Sensitec's AMR sensor AFF755, and low-power conditioning electronics with analog sleep-mode and a coil system with basic actuation electronics were developed and combined with state-of-the-art and proprietary localization algorithms. Details of this research activity can be found in the following slide set from the 2011 public status seminar of the mst-avs programme. The magnetic localization concept also suits well for general indoor-localization problems and since numerous similar activities can be observed by now, with publication dates as our work, but after the patent filing date.

Another relevant contextual feature in wireless sensor networks, in addition to location context, is temporal context, which requires synchronous clocks of the swarm elements and the base station. Limited accuracy/stability of microcontrollers time-base require periodical synchronization activities. Commonly applied RF-based synchronization faces the same issues as localization itself. So, the magnetic localization method was extended to sensor node time-base synchronization. The invented method was included in the patent application after first-cut experimental validation (see slide set above and publications enlisted below.).

For flexible testing of hardware and software of the sensor swarm operation, a modular PCB has been developed from March 2012, following up the first wire-wrap prototype:

The ISE-Sensor-board for PAC4PT project allows the modular inclusion of every sensor component, e.g., atmel Xmega 256A3 microcontroller with EEPROM extension, temperature sensor (UST), pressure sensor (microTec), RF-module (IMST), and localization modules (position sensor, flip module) from ISE, can be added as both standard PCB and MEMS 3D-CSP realizations and tested/applied in-situ. A lot of five boards has been developped and is seeing currently complete assembly to be tested in the ISE laboratory demonstrator (see also slide given above). The following picture shows the ISE-board equipped with 3D-CSP prototypes of the microcontroller/EEPROM, position sensor, and flip module:

Currently, since March 20120, the integration of this node into the ISE laboratory setup for distributed measurement, localization, and synchronization, is advanced, replacing the NI DAQ-based centralized prototype hardware. The localization algorithms are advanced and implemented in Matlab as well for research purposes in Python. The node software is evolved in cooperation with the industrial partners and testing is going on based a swarm size of 3-5 modules. Further, calibration activities, e.g., sensor sensitivity, sensor conditining electronics, and microcontroller ADC are pursued to increase measurement and localization accuracy. After validation, in the frame project PAC4PT the overall MEMS integration of the swarm's sensor node is aspired:

The project, due to delays in integration of sensor nodes had to be prolonged until 31.12.2012 and ISE had focused in the additional nine months on improvement of localisation techniques and results, finalisation and demonstration of node hardware and software, synchronization issues, inclusion of calibration and self-x capabilities, e.g., by MEMS-switch inclusion for reconfiguration, and completion of a laboratory ''nutshell'' ISE-demonstrator as well as the (still pending) final project demonstrator with all industrial partners. The sensor node software had been achieved and delivered based on PCB-development systems described above (a complete 3D-CSP node was unfortunately unavailable until the end of the project) The aspired PAC4PT ROSIG demonstrator has been completed in January 2013 based on our PCB-development systems:

The PAC4PT ROSIG demonstrator implements a complete front-to-back realization of a measurement campaign based on the PCB-development systems equipped with UST temperature sensor module, IMST communication module, a discrete pressure sensor module, and our 3D-AMR sensor module. The IMST gateway was employed to link to PC- based host application in Matlab for sensor configuration, data acquisition and mesurement, localisation computation, and visualization of sensor swarm data with T,p,..., time, and location context. A dedicated real-time pattern generator, based on the same microcontroller unit employed on our development boards, has been implemented for real-time coil driving at ISE. Thus, autonomous magnetic localization as well as synchronization (start-up) has been implemented in January 2013 and demonstrated for a small-scale measurement campaign at the end of the last project meeting on 16th January at ISE, TU Kaiserslautern (Complementary slide set). In the central container volume, localization errors in the order of 10 cm were already obtainable for the microcontroller-based small-scale measurements. Also, first self-x features for the sensors have been explored and implemented. After a significant overhead of development effort in the project, the actual research on numerous interesting topics becomes amenable at its end and will be pursued in self-funded activities from 2013, such as:

• Improvement of localisation and synchronization range and accuracy
• Speed-up of localisation rate by suitable power electronics
• Self-x features inclusion for accurate as well as dependable sensory systems, a prerequisite demanded, e.g., for Cyber-Physical-Systems (CPS) etc.
• Investigation of alternative/complementing 3D integration/packaging (System-in-Package, SiP) technologies for seamless integration & merging of every day life devices with sensors & electronics
• Complete integration of low-power sensory node with magnetic localization and synchronization, as well as additional sensing modalities, e.g., Integrated-Impedance-Spectroscopy and energy harvesting/management
• Transfer to other application domains and ISE activities, e.g., Smart-Enviroments, e.g., Smart-Kitchen (Lab-on-Spoon, ISE), Ambient-Intelligence/Assistance Systems, Medical Applications etc.

Our magnetic localization and synchronization system, embodied by the ROSIG-demonstrator, is presented on Hannover fair (see recent press release, presenters: Dennis Groben and Kittikhun Thongpull of ISE) .

The research progress on the first five research items given above will be reported on a new web-page opening in 2014 entitled:

''3D-Integration of 3D-Sensing, Self-x Sensor Swarms''

 

Acknowledgment: The contributions of R. Stemler, H. Brass, and H. Heintz of TU Kaiserslautern in PCB design and assembly are gratefully acknowledged.

The funding by BMBF of this research and the support by VDE/VDI-IT are gratefully acknowledged.