Markus Pichler

Fault detection in multi-sensor networks based on multivariate time-series models and orthogonal transformations

Abstract We introduce the usage of multivariate orthogonal space transformations and vectorized time-series models in combination with data-driven system identification models to achieve an enhanced performance of residual-based fault detection in condition monitoring systems equipped with multi-sensor networks. Neither time-consuming annotated samples nor fault patterns/models need to be available, as our approach is solely based on on-line recorded data streams. The system identification step acts as a fusion operation by searching for relations and dependencies between sensor channels measuring the state of system variables. We therefore apply three different vectorized time-series variants: (i) non-linear finite impulse response models (NFIR) relying only on the lagged input variables, (ii) non-linear output error models (NOE), also including the lags of the own predictions and (iii) non-linear Box–Jenkins models (NBJ) which include the lags of the predictions errors as well. The use of multivariate orthogonal space transformations allows to produce more compact and accurate models due to an integrated dimensionality (noise) reduction step. Fault detection is conducted based on finding anomalies (untypical occurrences) in the temporal residual signal in incremental manner. Our experimental results achieved on four real-world condition monitoring scenarios employing multi-sensor network systems demonstrate that the Receiver Operating Characteristic (ROC) curves are improved over those ones achieved with native static models (w/o lags, w/o transformations) by about 20–30%.

Precise distance measurement with IEEE 802.15.4 (ZigBee) devices

In modern wireless communications products it is required to incorporate more and more different functions to comply with current market trends. A very attractive function with steadily growing market penetration is local positioning. To add this feature to low-cost mass-market devices without additional power consumption, it is desirable to use commercial communication chips and standards for localization of the wireless units. In this paper we present a concept to measure the distance between two IEEE 802.15.4 (ZigBee) compliant devices. The presented prototype hardware consists of a low- cost 2.45 GHz ZigBee chipset. For localization we use standard communication packets as transmit signals. Thus simultaneous data transmission and transponder localization is feasible. To achieve high positioning accuracy even in multipath environments, a coherent synthesis of measurements in multiple channels and a special signal phase evaluation concept is applied. With this technique the full available ISM bandwidth of 80 MHz is utilized. In first measurements with two different frequency references-a low-cost oscillator and a temperatur-compensated crystal oscillator-a positioning bias error of below 16 cm and 9 cm was obtained. The standard deviation was less than 3 cm and 1 cm, respectively. It is demonstrated that compared to signal correlation in time, the phase processing technique yields an accuracy improvement of roughly an order of magnitude.

Indoor Localization of Passive UHF RFID Tags Based on Phase-of-Arrival Evaluation

This paper introduces a two-dimensional localization system for passive UHF RFID transponders based on phase-of-arrival evaluation of the backscattered tag signal. A multiple-channel system, where alternately one path is configured as a transmitter and the remaining paths work as receivers, is used to permit the position estimation. To improve localization accuracy and reduce disturbances caused by multipath propagation, a frequency-stepped continuous-waveform approach is employed. To achieve proof of concept, a local position measurement system demonstrator comprising an RFID reader, passive EPC Gen 2 RFID tags, several transceivers, baseband hardware, and signal processing was used for measurements. The results show excellent localization accuracy, even in an indoor environment.

Influence of systematic frequency-sweep non-linearity on object distance estimation in FMCW/FSCW radar systems

The performance of frequency estimation procedures for sinusoidal signals in the presence of phase perturbations plays an important role in continuous-wave radar systems for distance measurement, as it directly determines the ranging accuracy. For the most widely used frequency estimator, the fast Fourier transformation, the influence of deterministic deviations of the transmit signal from the ideal curve in FMCW/FSCW radars is investigated, quantified and verified with measured radar data.

Phase-Error Measurement and Compensation in PLL Frequency Synthesizers for FMCW Sensors—I: Context and Application

The synthesis of linear frequency sweeps or chirps is required, among others, in frequency-modulated continuous-wave radar systems for object position estimation. Low phase and frequency errors in sweeps with high bandwidth are a prerequisite for good accuracy and resolution, but, in certain applications where high measurement rates are desired, the additional demand for short sweep cycles has to be met. Transient phenomena in dynamic synthesizers as well as nonlinear system behavior usually cause unknown phase errors in the system output. For the class of phase-locked-loop (PLL)-based frequency synthesizers, a novel output phase-measurement method and dedicated circuitry are proposed that allow significant reduction of phase errors by adaptive input predistortion. The measurement procedure is implemented within the PLL control circuitry and does not require external equipment. The application of this method to PLL system identification and linearization of extremely short frequency sweeps is shown

Identification of SAW ID-tags using an FSCW interrogation unit and model-based evaluation

Surface acoustic wave (SAW) devices are well suited for the application as wireless and completely passive identification (ID)-tag. They operate even under harsh environmental conditions, especially under high temperatures. Interrogation units operate similar to conventional radar systems and therefore, have the same resolution restrictions when an inverse fast Fourier transformation (IFFT) is applied. The application of model-based algorithms to the evaluation of identification tags is presented and enhancements in either improved resolution or reduced system bandwidth are shown. Furthermore, optimizations in the coding scheme for reduced tag sizes are discussed, and measured results obtained by a built interrogation unit and fabricated ID-tags for an identification system are shown.

Multi-Channel Distance Measurement With IEEE 802.15.4 (ZigBee) Devices

The addition of positioning capabilities to low-cost communications such as IEEE 802.15.4 compliant (ZigBee) networks can open up interesting markets, in particular if existing hardware and standard communications packets can be used for this purpose. High-precision localization requires large transmission bandwidths and thus the use of multiple frequency-channels. In the presence of oscillator frequency uncertainty the coherent synthesis of individual measurements poses a challenge. This paper proposes a particular frequency-hopping and signal processing scheme by which most transmitter and receiver frequency-errors can be eliminated in signal processing. Measurements show that high precision distance estimation with errors down to a few cm can be achieved.

UHF RFID Localization Based on Phase Evaluation of Passive Tag Arrays

This paper presents a 2-D localization system based on phase-of-arrival evaluation of passive ultrahigh frequency (UHF) radio frequency identification (RFID) transponders. To handle the ambiguity caused by phase evaluation, several tags are arranged to form a uniform linear array. A multiple input multiple output system, where sequentially each frontend is configured to work as transmitter, while the remaining frontends serve as receivers, is used to enable position estimation. For proof of concept, a local position measurement system demonstrator was built comprising conventional passive EPCglobal Class-1 Gen-2 UHF RFID tags, one commercial off-the-shelf RFID reader, several transceiver frontends, baseband hardware, and signal processing. Measurements were carried out in an indoor office environment, where the

Improved fault detection employing hybrid memetic fuzzy modeling and adaptive filters

3.5~{\rm m} \times 2.5

Application of state-space frequency estimation to a 24-GHz FMCW tank level gauging system

m measurement zone was surrounded by drywalls and concrete floor and ceiling, and the experimental results showed robust and accurate localization with a root-mean-square deviation of 0.011 m and a maximum error of 0.032 m. These experimental results were confirmed by simulations, which were performed to determine the limits of the system accuracy.