Models and algorithms for tracking of maneuvering objects using variable rate particle filters

Standard algorithms in tracking and other state-space models assume identical and synchronous sampling rates for the state and measurement processes. However, real trajectories of objects are typically characterized by prolonged smooth sections, with sharp, but infrequent, changes. Thus, a more parsimonious representation of a target trajectory may be obtained by direct modeling of maneuver times in the state process, independently from the observation times. This is achieved by assuming the state arrival times to follow a random process, typically specified as Markovian, so that state points may be allocated along the trajectory according to the degree of variation observed. The resulting variable dimension state inference problem is solved by developing an efficient variable rate particle filtering algorithm to recursively update the posterior distribution of the state sequence as new data becomes available. The methodology is quite general and can be applied across many models where dynamic model uncertainty occurs on-line. Specific models are proposed for the dynamics of a moving object under internal forcing, expressed in terms of the intrinsic dynamics of the object. The performance of the algorithms with these dynamical models is demonstrated on several challenging maneuvering target tracking problems in clutter. © 2006 IEEE.

Enhanced 10-Gbit/s link performance for directly Modulated complex-coupled DFB lasers via resonance frequency, damping rate, and chirp

This paper demonstrates the respective roles that combined index- and gain-coupling play in the overall link performance of distributed feedback (DFB) lasers. Their impacts on both static and dynamic properties such as slope efficiency, resonance frequency, damping rate, and chirp are investigated. Simulation results are compared with experimental data with good agreement. Transmission-oriented optimization is then demonstrated based on a targeted specification. The design tradeoffs are revealed, and it is shown that a modest combination of index- and gain-coupling enables optimum transmission at 10 Gbit/s.

Passively harmonically mode-locked vertical-external-cavity surface-emitting laser emitting 1.1 ps pulses at 147 GHz repetition rate

Coupled-cavity passive harmonic mode-locking of a quantum well based vertical-external-cavity surface-emitting laser has been demonstrated, yielding an output pulse train of 1.5 ps pulses at a repetition rate of 80 GHz and with an average power of 80 mW. Harmonic mode-locking results from coupling between the main laser cavity and a cavity formed within the substrate of the saturable absorber structure. Mode-locking on the second harmonic of the substrate cavity allows a train of 1.1 ps pulses to be generated at a repetition rate of 147 GHz with 40 mW average power. © 2010 American Institute of Physics.

Growth rate of YBCO single grains containing Y-2411(M)

Y-Ba-Cu-O (YBCO) single grains have the potential to generate large trapped magnetic fields for a variety of engineering applications, and research on the processing and properties of this material has attracted world-wide interest. In particular, the introduction of flux pinning centres to the large grain microstructure to improve its current density, Jc, and hence trapped field, has been investigated extensively over the past decade. Y 2Ba4CuMOx [Y-2411(M)], where M = Nb, Ta, Mo, W, Ru, Zr, Bi and Ag, has been reported to form particularly effective flux pinning centres in YBCO due primarily to its ability to exist as nano-size inclusions in the superconducting phase matrix. However, the addition of the Y-2411(M) phase to the precursor composition complicates the melt-processing of single grains. We report an investigation of the growth rate of single YBCO grains containing Y-2411(Bi) phase inclusions and Y2O3. The superconducting properties of these large single grains have been measured specifically to investigate the effect of Y2O3 on broadening the growth window of these materials. © 2010 IOP Publishing Ltd.

Discriminative language model adaptation for Mandarin broadcast speech transcription and translation

This paper investigates unsupervised test-time adaptation of language models (LM) using discriminative methods for a Mandarin broadcast speech transcription and translation task. A standard approach to adapt interpolated language models to is to optimize the component weights by minimizing the perplexity on supervision data. This is a widely made approximation for language modeling in automatic speech recognition (ASR) systems. For speech translation tasks, it is unclear whether a strong correlation still exists between perplexity and various forms of error cost functions in recognition and translation stages. The proposed minimum Bayes risk (MBR) based approach provides a flexible framework for unsupervised LM adaptation. It generalizes to a variety of forms of recognition and translation error metrics. LM adaptation is performed at the audio document level using either the character error rate (CER), or translation edit rate (TER) as the cost function. An efficient parameter estimation scheme using the extended Baum-Welch (EBW) algorithm is proposed. Experimental results on a state-of-the-art speech recognition and translation system are presented. The MBR adapted language models gave the best recognition and translation performance and reduced the TER score by up to 0.54% absolute. © 2007 IEEE.

Exploiting Chinese character models to improve speech recognition performance

The Chinese language is based on characters which are syllabic in nature. Since languages have syllabotactic rules which govern the construction of syllables and their allowed sequences, Chinese character sequence models can be used as a first level approximation of allowed syllable sequences. N-gram character sequence models were trained on 4.3 billion characters. Characters are used as a first level recognition unit with multiple pronunciations per character. For comparison the CU-HTK Mandarin word based system was used to recognize words which were then converted to character sequences. The character only system error rates for one best recognition were slightly worse than word based character recognition. However combining the two systems using log-linear combination gives better results than either system separately. An equally weighted combination gave consistent CER gains of 0.1-0.2% absolute over the word based standard system. Copyright © 2009 ISCA.

Language model combination and adaptation using weighted finite state transducers

In speech recognition systems language model (LMs) are often constructed by training and combining multiple n-gram models. They can be either used to represent different genres or tasks found in diverse text sources, or capture stochastic properties of different linguistic symbol sequences, for example, syllables and words. Unsupervised LM adaptation may also be used to further improve robustness to varying styles or tasks. When using these techniques, extensive software changes are often required. In this paper an alternative and more general approach based on weighted finite state transducers (WFSTs) is investigated for LM combination and adaptation. As it is entirely based on well-defined WFST operations, minimum change to decoding tools is needed. A wide range of LM combination configurations can be flexibly supported. An efficient on-the-fly WFST decoding algorithm is also proposed. Significant error rate gains of 7.3% relative were obtained on a state-of-the-art broadcast audio recognition task using a history dependently adapted multi-level LM modelling both syllable and word sequences. ©2010 IEEE.

Heat release rate measurement in turbulent flames

The dominant industrial approach for the reduction of NO x emissions in industrial gas turbines is the lean pre-mixed prevaporized concept. The main advantage of this concept is the lean operation of the combustion process; this decreases the heat release rate from the flame and results in a reduction in operating temperature. The direct measurement of heat release rates via simultaneous laser induced fluorescence of OH and CH 2O radicals using planar laser induced fluorescence. The product of the two images correlated with the forward production rate of the HCO radical, which in turn has correlated well with heat release rates from premixed hydrocarbon flames. The experimental methodology of the measurement of heat release rate and applications in different turbulent premixed flames were presented. This is an abstract of a paper presented at the 7th World Congress of Chemical Engineering (Glasgow, Scotland 7/10-14/2005).

High rate deposition of ta-C:H using an electron cyclotron wave resonance plasma source

A compact electron cyclotron wave resonance (ECWR) source has been developed for the high rate deposition of hydrogenated tetrahedral amorphous carbon (ta-C:H). The ECWR provides growth rates of up to 1.5 nm/s over a 4-inch diameter and an independent control of the deposition rate and ion energy. The ta-C:H was deposited using acetylene as the source gas and was characterized as having an sp3 content of up to 77%, plasmon energy of 27 eV, refractive index of 2.45, hydrogen content of about 30%, optical gap of up to 2.1 eV and RMS surface roughness of 0.04 nm. © 1999 Elsevier Science S.A. All rights reserved.

High rate deposition of Ta-C:H using an electron cyclotron wave resonance plasma source

A compact electron cyclotron wave resonance (ECWR) source has been developed for the high rate deposition of hydrogenated tetrahedral amorphous carbon (ta-C:H). The ECWR provides growth rates of up to 900 angstrom/min and an independent control of the deposition rate and ion energy. The ta-C:H was deposited using acetylene as the source gas and was characterized in terms of its bonding, stress and friction coefficient. The results indicated that the ta-C:H produced using this source fulfills the necessary requirements for applications requiring enhanced tribological performance.

169 GHz repetition rate passively harmonically mode-locked VECSEL emitting 265 fs pulses

High repetition rate passively mode-locked sources are of significant interest due to their potential for applications including optical clocking, optical sampling, communications and others. Due to their short excited state lifetimes mode-locked VECSELs are ideally suited to high repetition rate operation, however fundamentally mode-locked quantum well-based VECSELs have not achieved repetition rates above 10 GHz due to the limitations placed on the cavity geometry by the requirement that the saturable absorber saturates more quickly than the gain. This issue has been overcome by the use of quantum dot-based saturable absorbers with lower saturation fluences leading to repetition rates up to 50 GHz, but sub-picosecond pulses have not been achieved at these repetition rates. We present a passively harmonically mode-locked VECSEL emitting pulses of 265 fs duration at a repetition rate of 169 GHz with an output power of 20 mW. The laser is based around an antiresonant 6 quantum well gain sample and is mode-locked using a semiconductor saturable absorber mirror. Harmonic modelocking is achieved by using an intracavity sapphire etalon. The sapphire then acts as a coupled cavity, setting the repetition rate of the laser while still allowing a tight focus on the saturable absorber. RF spectra of the laser output show no peaks at harmonics of the fundamental repetition rate up to 26 GHz, indicating stable harmonic modelocking. Autocorrelations reveal groups of pulses circulating in the cavity as a result of an increased tendency towards Q-switched modelocking due to the low pulse energies.

Language model cross adaptation for LVCSR system combination

State-of-the-art large vocabulary continuous speech recognition (LVCSR) systems often combine outputs from multiple subsystems developed at different sites. Cross system adaptation can be used as an alternative to direct hypothesis level combination schemes such as ROVER. In normal cross adaptation it is assumed that useful diversity among systems exists only at acoustic level. However, complimentary features among complex LVCSR systems also manifest themselves in other layers of modelling hierarchy, e.g., subword and word level. It is thus interesting to also cross adapt language models (LM) to capture them. In this paper cross adaptation of multi-level LMs modelling both syllable and word sequences was investigated to improve LVCSR system combination. Significant error rate gains up to 6.7% rel. were obtained over ROVER and acoustic model only cross adaptation when combining 13 Chinese LVCSR subsystems used in the 2010 DARPA GALE evaluation. © 2010 ISCA.

Symmetrical 2-D Hermite-Gaussian square launch for high bit rate transmission in multimode fiber links

A 2-D Hermite-Gaussian square launch is demonstrated to show improved systems capacity over multimode fiber links. It shows a bandwidth improvement over both center and offset launches and exhibits ±5 μm misalignment tolerance. © 2011 Optical Society of America.