No abstract

Over the past decades, extensive research has been carried out on various possible implementations of automatic speech recognition (ASR) systems. The most renowned algorithms in the field of ASR are the mel-frequency cepstral coefficients and the hidden Markov models. However, there are also other methods, such as wavelet-based transforms, artificial neural networks and support vector machines, which are becoming more popular. This review article presents a comparative study on different approaches that were proposed for the task of ASR, and which are widely used nowadays. † training time increases linearly with increase in vocabulary size [42] † quantisation error in the discrete representation of speech signals [42] † temporal information is ignored [42] PCA † reduction in the feature vector's size, while retaining much of the significant information [131] † robust [59, 60] † computationally expensive for high-dimensional data [8] LDA † maximises the distance between classes, but minimises the within class distance [132] † robust [133] † sample distribution is assumed a priori to be Gaussian [63] † class samples are assumed to have equal variance [63] Classification technique Advantages Disadvantages HMM † able to model time distribution of speech signals [103] † simple to adapt [68] † capable to model a sequence of discrete or continuous symbols [13] † inputs can be of variable length [40] † based on the assumption that the probability of being in a particular state is dependent only on its preceding state, ignoring any long-term dependencies [82] † emission probabilities are arbitrarily chosen; hence, these might not even represent properly the output probabilities of the corresponding state [82] ANN (in general) † good classifiers [16, 45] † highly adequate for pattern recognition applications [16, 45] † self-organising [16, 45] † self-learning [16, 45] † self-adaptive in new environments [16, 45] † robust [7] † based on ERM; hence, prone to over training a local minima problems [45, 103] MLP † good discriminating ability [2] † unable to model time distribution of speech signals [2] † inputs have to be of fixed length [2] † able to deal with small vocabularies only [2] SOM † no a priori information is required for training a SOM [134] † can easily adapt if a new sample is presented to it [134] † capable of parallel computation [134] † SOM algorithm is not well defined mathematically; hence, values for the network parameters need to be found by trial-and-error [134] † ordered mapping obtained after the training phase may be lost when applied in real environments due to frequent adaptations [134] RBF † simple to implement [135] † Good discriminating ability [135] † robust [135] † online learning ability [135] † shift invariant in time [91] RNN † able to model time distribution of speech signals thanks to the feedback connections [95, 103] † complex training algorithm [94] † training algorithm is highly sensitive to any changes [94] FNN † does not need large amount of samples during the learning process [99] † ...

The disordering of In0.53Ga0.47As/InP single quantum wells has been studied using an error function distribution to model the compositional profile after interdiffusion. When considering interdiffusion on the group-III sublattice only, a large strain buildup results during the early stages of disordering. Details are presented showing how this interdiffusion and the effects of strain lead to an interesting carrier confinement profile which differs from that of disordered AlGaAs/GaAs and InGaAs/GaAs quantum-well structures. An abrupt confinement profile is maintained even after significant interdiffusion, with a well width equal to that of the as-grown quantum well. The combined effects of strain with the unstrained band-gap profile results in a potential buildup in the barrier near the interface, while it gives rise to two ‘‘miniwells’’ inside the potential wells. The sub-band-edge structure shows that the potential buildup can result in quasibound subband states, while the heavy-hole well can support the ground state within the miniwells. In contrast, when identical interdiffusion on both group-III and group-V sublattices is considered, the structure remains lattice matched, the confinement profile changes to that of a graded profile, and the ground-state transition energy shifts to shorter wavelengths.

This review paper presents a comparative study of published integrated submicron CMOS quadrature voltage-controlled oscillator designs, based on LC resonator tanks operating at gigahertz frequencies. Although special reference to phase noise reduction is made, the comparison also concerns issues such as power consumption, tuning range and the phase accuracy of the quadrature signals. The effect of supply voltage reduction on the choice of the oscillator topology is also included in the discussion.

A complete analysis of the spur characteristics of edge-combining delay-locked loop (DLL)-based frequency multipliers is presented in this brief. The novelty of this analysis is the fact that it can be used to estimate the effect of both the in-lock error and the delay-stage mismatch on the spurious level of the frequency multiplier with low computational complexity. In addition, a way to reduce the mismatch between the delay cells in the delay line is discussed via an analytic model and verified by the implementation of a delay cell in a 65-nm CMOS process.Index Terms-Delay mismatch, delay-locked loop (DLL), frequency synthesis, in-lock error, spurious level.