A scatterometer neural network sensor model with input noise

The ERS-1 satellite carries a scatterometer which measures the amount of radiation scattered back toward the satellite by the ocean's surface. These measurements can be used to infer wind vectors. The implementation of a neural network based forward model which maps wind vectors to radar backscatter is addressed. Input noise cannot be neglected. To account for this noise, a Bayesian framework is adopted. However, Markov Chain Monte Carlo sampling is too computationally expensive. Instead, gradient information is used with a non-linear optimisation algorithm to find the maximum em a posteriori probability values of the unknown variables. The resulting models are shown to compare well with the current operational model when visualised in the target space.

A scatterometer neural network sensor model with input noise

The ERS-1 satellite carries a scatterometer which measures the amount of radiation scattered back toward the satellite by the ocean's surface. These measurements can be used to infer wind vectors. The implementation of a neural network based forward model which maps wind vectors to radar backscatter is addressed. Input noise cannot be neglected. To account for this noise, a Bayesian framework is adopted. However, Markov Chain Monte Carlo sampling is too computationally expensive. Instead, gradient information is used with a non-linear optimisation algorithm to find the maximum em a posteriori probability values of the unknown variables. The resulting models are shown to compare well with the current operational model when visualised in the target space.

Ultrasonic field and temperature sensor based on short in-fibre Bragg gratings

The authors demonstrate that in-fibre Bragg gratings may be successfully used to measure megahertz acoustic fields if the grating length is sufficiently short and the optical fibre is appropriately desensitised. A noise-limited pressure resolution of 4.5 × 10 –3 atm vHz was found. The capability to simultaneously act as a temperature sensor is also demonstrated.

Combined ultrasound and temperature sensor using a fibre Bragg grating

We demonstrate the feasibility of using in-fibre Bragg gratings to measure MHz acoustic fields and temperature simultaneously. We achieved a noise-limited pressure resolution of ˜4.5×10-4 Atm/vHz and a temperature resolution of 0.2°C.

Resonant cavity Fibre Bragg grating sensor interrogation

This thesis presents a novel high-performance approach to time-division-multiplexing (TDM) fibre Bragg grating (FBG) optical sensors, known as the resonant cavity architecture. A background theory of FBG optical sensing includes several techniques for multiplexing sensors. The limitations of current wavelength-division-multiplexing (WDM) schemes are contrasted against the technological and commercial advantage of TDM. The author’s hypothesis that ‘it should be possible to achieve TDM FBG sensor interrogation using an electrically switched semiconductor optical amplifier (SOA)’ is then explained. Research and development of a commercially viable optical sensor interrogator based on the resonant cavity architecture forms the remainder of this thesis. A fully programmable SOA drive system allows interrogation of sensor arrays 10km long with a spatial resolution of 8cm and a variable gain system provides dynamic compensation for fluctuating system losses. Ratiometric filter- and diffractive-element spectrometer-based wavelength measurement systems are developed and analysed for different commercial applications. The ratiometric design provides a low-cost solution that has picometre resolution and low noise using 4% reflective sensors, but is less tolerant to variation in system loss. The spectrometer design is more expensive, but delivers exceptional performance with picometre resolution, low noise and tolerance to 13dB system loss variation. Finally, this thesis details the interrogator’s peripheral components, its compliance for operation in harsh industrial environments and several examples of commercial applications where it has been deployed. Applications include laboratory instruments, temperature monitoring systems for oil production, dynamic control for wind-energy and battery powered, self-contained sub-sea strain monitoring.

Fibre Bragg grating sensor interrogation using an acousto-optic tunable filter and low-coherence interferometry

We describe how an acousto-optic tunable filter can be used to both demultiplex the signals from multiple fibre Bragg grating sensors and simultaneously provide wide bandwidth signal demodulation in a system using interferometric wavelength shift detection. In an experimental demonstration, the approach provided a noise limited strain resolution of 24.9 n epsilon Hz(-1/ 2) at 15 Hz.

Ultrasonic field and temperature sensor based on short in-fibre Bragg gratings

The authors demonstrate that in-fibre Bragg gratings may be successfully used to measure megahertz acoustic fields if the grating length is sufficiently short and the optical fibre is appropriately desensitised. A noise-limited pressure resolution of 4.5 × 10 –3 atm vHz was found. The capability to simultaneously act as a temperature sensor is also demonstrated.

Combined ultrasound and temperature sensor using a fibre Bragg grating

The feasibility of in-fiber Bragg gratings for simultaneous acoustic field and temperature sensing was demonstrated. A noise-limited pressure resolution of about 4.5×10-4 Atm/√Hz and a temperature resolution of 0.2 °C was achieved.

Fibre Bragg grating sensor interrogation using an acousto-optic tunable filter and low-coherence interferometry

We describe how an acousto-optic tunable filter can be used to both demultiplex the signals from multiple fibre Bragg grating sensors and simultaneously provide wide bandwidth signal demodulation in a system using interferometric wavelength shift detection. In an experimental demonstration, the approach provided a noise limited strain resolution of 24.9 nε Hz -1/2 at 15 Hz. © 2007 IOP Publishing Ltd.

An adaptive code position modulation scheme for wireless sensor networks

In this letter, an energy-efficient adaptive code position modulation scheme is proposed for wireless sensor networks to provide the relatively stable bit error ratio (BER) performance expected by the upper layers. The system is designed with focus on the adaptive control of transmission power, which is adjusted based on the measured power density of background noise. Interfaces among the modulation module, packet scheduling module and upper layer are provided for flexible adjustments to adapt to the background noise and deliver expected application quality. Simulations with Signal Processing Worksystem (SPW) validate the effectiveness of the scheme. © 2005 IEEE.

Optical sensor interrogation with a blazed fiber Bragg grating and a charge-coupled device linear array

We present what is to our knowledge the first comprehensive investigation of the use of blazed fiber Bragg gratings (BFBGs) to interrogate wavelength division multiplexed (WDM) in-fiber optical sensor arrays. We show that the light outcoupled from the core of these BFBGs is radiated with sufficient optical power that it may be detected with a low-cost charge-coupled device (CCD) array. We present thorough system performance analysis that shows sufficient spectral-spatial resolution to decode sensors with a WDM separation of 75 ρm, signal-to-noise ratio greater than 45-dB bandwidth of 70 nm, and drift of only 0.1 ρm. We show the system to be polarization-state insensitive, making the BFBG-CCD spectral analysis technique a practical, extremely low-cost, alternative to traditional tunable filter approaches.

High accuracy interrogation of a WDM FBG sensor array using radiation modes from a B-FBG

We report the development of a WDM optical sensor array interrogation system using the radiation modes from a BFBG. We present results indicating 70nm bandwidth, with 0.2um RMS noise and a minimum WDM spacing of 30um. We further show the system to be polarization independent.

Resonant cavity time-division-multiplexed fiber Bragg grating sensor interrogator

The first resonant-cavity time-division-multiplexed (TDM) fiber Bragg grating sensor interrogation system is reported. This novel design uses a pulsed semiconductor optical amplifier in a cyclic manner to function as the optical source, amplifier, and modulator. Compatible with a range of standard wavelength detection techniques, this optically gated TDM system allows interrogation of low reflectivity "commodity" sensors spaced just 2 m apart, using a single active component. Results demonstrate an exceptional optical signal-to-noise ratio of 36 dB, a peak signal power of over +7 dBm, and no measurable crosstalk between sensors. Temperature tuning shows that the system is fully stable with a highly linear response. © 2004 IEEE.

Projected sequential Gaussian processes:a C++ tool for interpolation of large datasets with heterogeneous noise

Heterogeneous datasets arise naturally in most applications due to the use of a variety of sensors and measuring platforms. Such datasets can be heterogeneous in terms of the error characteristics and sensor models. Treating such data is most naturally accomplished using a Bayesian or model-based geostatistical approach; however, such methods generally scale rather badly with the size of dataset, and require computationally expensive Monte Carlo based inference. Recently within the machine learning and spatial statistics communities many papers have explored the potential of reduced rank representations of the covariance matrix, often referred to as projected or fixed rank approaches. In such methods the covariance function of the posterior process is represented by a reduced rank approximation which is chosen such that there is minimal information loss. In this paper a sequential Bayesian framework for inference in such projected processes is presented. The observations are considered one at a time which avoids the need for high dimensional integrals typically required in a Bayesian approach. A C++ library, gptk, which is part of the INTAMAP web service, is introduced which implements projected, sequential estimation and adds several novel features. In particular the library includes the ability to use a generic observation operator, or sensor model, to permit data fusion. It is also possible to cope with a range of observation error characteristics, including non-Gaussian observation errors. Inference for the covariance parameters is explored, including the impact of the projected process approximation on likelihood profiles. We illustrate the projected sequential method in application to synthetic and real datasets. Limitations and extensions are discussed. © 2010 Elsevier Ltd.