A cognitive radio will have to sense and discover the spectral environments where it would not cause primary radios to interfere. Because the primary radios have the right to use the frequency, the cognitive radios as the secondary radios must detect radio signals before use. However, the secondary radios also need identifying the primary and other secondary radios where the primary radios are vulnerable to interference. In this paper, a method of simultaneously identifying signals of primary and secondary radios is proposed. The proposed bandwidth differentiation assumes the primary and secondary radios use orthogonal frequency division multiplexing (OFDM), and the secondary radios use at the lower number of subcarriers than the primary radios. The false alarm and detection probabilities are analytically evaluated using the characteristic function method. Numerical evaluations are also conducted on the assumption the primary radio is digital terrestrial television broadcasting. Result showed the proposed method could achieve the false alarm probability of 0.1 and the detection probability of 0.9 where the primary and secondary radio powers were 2.5 dB and 3.6 dB higher than the noise power. In the evaluation, the reception signals were averaged over the successive 32 snapshots, and the both the primary and secondary radios used QPSK. The power ratios were 4.7 dB and 8.4 dB where both the primary and secondary radios used 64QAM.

This paper aims to improve the domain robustness of language modeling for automatic speech recognition (ASR). To this end, we focus on applying the latent words language model (LWLM) to ASR. LWLMs are generative models whose structure is based on Bayesian soft class-based modeling with vast latent variable space. Their flexible attributes help us to efficiently realize the effects of smoothing and dimensionality reduction and so address the data sparseness problem; LWLMs constructed from limited domain data are expected to robustly cover unknown multiple domains in ASR. However, the attribute flexibility seriously increases computation complexity. If we rigorously compute the generative probability for an observed word sequence, we must consider the huge quantities of all possible latent word assignments. Since this is computationally impractical, some approximation is inevitable for ASR implementation. To solve the problem and apply this approach to ASR, this paper presents an n-gram approximation of LWLM. The n-gram approximation is a method that approximates LWLM as a simple back-off n-gram structure, and offers LWLM-based robust one-pass ASR decoding. Our experiments verify the effectiveness of our approach by evaluating perplexity and ASR performance in not only in-domain data sets but also out-of-domain data sets.

This paper proposes an efficient combination of state likelihood recycling and batch state likelihood calculation for accelerating acoustic likelihood calculation in an HMM-based speech recognizer. Recycling and batch calculation are each based on different technical approaches, i.e. the former is a purely algorithmic technique while the latter fully exploits computer architecture. To accelerate the recognition process further by combining them efficiently, we introduce conditional fast processing and acoustic backing-off. Conditional fast processing is based on two criteria. The first potential activity criterion is used to control not only the recycling of state likelihoods at the current frame but also the precalculation of state likelihoods for several succeeding frames. The second reliability criterion and acoustic backing-off are used to control the choice of recycled or batch calculated state likelihoods when they are contradictory in the combination and to prevent word accuracies from degrading. Large vocabulary spontaneous speech recognition experiments using four different CPU machines under two environmental conditions showed that, compared with the baseline recognizer, recycling and batch calculation, our combined acceleration technique further reduced both of the acoustic likelihood calculation time and the total recognition time. We also performed detailed analyses to reveal each technique's acceleration and environmental dependency mechanisms by classifying types of state likelihoods and counting each of them. The analysis results comfirmed the effectiveness of the combined acceleration technique.

This paper introduces robust optimization models for minimization of the network congestion ratio that can handle the fluctuation in traffic demands between nodes. The simplest and widely used model to minimize the congestion ratio, called the pipe model, is based on precisely specified traffic demands. However, in practice, network operators are often unable to estimate exact traffic demands as they can fluctuate due to unpredictable factors. To overcome this weakness, we apply robust optimization to the problem of minimizing the network congestion ratio. First, we review existing models as robust counterparts of certain uncertainty sets. Then we consider robust optimization assuming ellipsoidal uncertainty sets, and derive a tractable optimization problem in the form of second-order cone programming (SOCP). Furthermore, we take uncertainty sets to be the intersection of ellipsoid and polyhedral sets, and considering the mirror subproblems inherent in the models, obtain tractable optimization problems, again in SOCP form. Compared to the previous model that assumes an error interval on each coordinate, our models have the advantage of being able to cope with the total amount of errors by setting a parameter that determines the volume of the ellipsoid. We perform numerical experiments to compare our SOCP models with the existing models which are formulated as linear programming problems. The results demonstrate the relevance of our models in terms of congestion ratio and computation time.

Although Hidden Markov Modeling (HMM) is widely and successfully used in many speech recognition applications, duration control for HMMs is still an important issue in improving recognition accuracy since a HMM places no constraints on duration. For compensating this defect, some duration control algorithms that employ precise duration models have been proposed. However, they suffer from greatly increased computational complexity. This paper proposes a new state duration control algorithm for limiting both the maximum and the minimum state durations. The algorithm is for the HMM trellis likelihood calculation, not for the Viterbi calculation. The amount of computation required by this algorithm is only order one (O(1)) for the maximum state duration n; that is, the computation amount is independent of the maximum state duration while many conventional duration control algorithm require computation in the amount of order n or order n2. Thus, the algorithm can drastically reduce the computation needed for duration control. The algorithm uses the property that the trellis likelihood calculation is a summation of many path likelihoods. At each frame, the path likelihood that exceeds the maximum likelihood is subtracted, and the path likelihood that satisfies the minimum likelihood is added to the forward probability. By iterating this procedure, the algorithm calculates the trellis likelihood efficiently. The algorithm was evaluated using a large-vocabulary speaker-independent spontaneous speech recognition system for telephone directory assistance. The average reduction in error rate for sentence understanding was about 7% when using context-independent HMMs, and 3% when using context-dependent HMMs. We could confirm the improvement by using the proposed state duration control algorithm even though the maximum and the minimum state durations were not optimized for the task (speaker-independent duration settings obtained from a different task were used).

In human-machine systems, a user gets abstracted information of a machine via an interface and operates it referring to a manual. If a manual has an erroneous description leading to automation surprises, the user may be lost in his/her operations so that he/she may make a serious human error. In this paper, we propose an algorithm for generating a manual by which automation surprises never occur. We model the machine and the interface as a discrete event system and a mapping from machine's state to a display of the interface, respectively. First, we represent a manual as a finite language and model behavior of the system operated by the user with the manual as a tree called an operational tree. Next, we characterize three automation surprises using the tree. Finally, we propose an algorithm for generating an operational tree by which the machine reaches a target state.

In AMS/OFDM systems, base station is in control of the modulation level of each subcarrier, and then, adaptive modulated packet is transmitted from the base station to the mobile station. In this case, the mobile station is required the modulation level information (MLI) to demodulate the received packet. The MLI is generally transmitted as a data symbol, therefore, the throughput is degraded. In an OFDM, the channel response at a particular subcarrier frequency is not supposed to be totally different from its neighboring frequencies, and hence, they must have correlation which depends on the coherence bandwidth of the channel Bc. If we could assign the same modulation level for coherently faded subcarrier block, MLI is required only one time for each subcarrier block. Moreover, we can assign the data on the empty space of pilot signals for increasing the total transmission. In this paper, we propose an adaptive subcarrier block modulation with differentially modulated pilot symbol assistance for downlink OFDM using uplink delay spread.

This paper describes the effects of system LSI design with C language-based behavioral synthesis following several trials of design period reduction and quality improvement for a variety of circuit types. The results of these trials are analyzed from the viewpoints of description productivity, verification productivity, reusability and design flexibility as well as hardware and software co-verification. First the C-based design flow proposed by the authors is described, and the design productivity and verification productivity under this design flow is compared to RTL design. The reusability of the behavioral IP core and its efficiency with HW/SW co-verification are also shown using design examples. Next, using the example of an MPEG-4 video decoder design, a typical design process in a C-based design is shown with considerations regarding verification efficiency, reusability of the IP core and HW/SW co-verification. Finally, the authors' perspectives regarding future directions of system LSI design are discussed.

A model that combines free-space loss (proportional to the square of distance d) and excess loss has been known to assess the microwave line-of-sight (LOS) path loss in street microcell environments. The excess loss represents the effects of shadowing obstacles. We measure the path loss at the 3.35, 8.45, and 15.75 GHz frequencies in an urban environment, and analyze the distance characteristics of the pass loss for mobile antenna heights of 2.7, 1.6, and 0.5 m. Results show that using a new model that bases on a dα formula instead of d2 in the conventional model produced a better fit to the measured data. They also show that lowering the mobile antenna to a height of 0. 5 m made it possible to virtually ignore the excess loss factor and, instead, use the dα formula to assess the path loss characteristics.

This paper describes a new technique for isolated word recognition that uses both pitch information and spectral information. In conventional methods, words with similar phoneme features tend to be misrecognized even if their phonemes are accented differently because these methods use only spectral information. It is possible to improve recognition accuracy by considering pitch patterns of words. Many phonetically-similar Japanese words are classified by pitch patterns. In this technique, a pitch pattern template is produced by averaging pitch patterns obtained from a set of words which have the same accent pattern. A measure for word recognition is proposed. This measure based on a combination of the phoneme likelihood and the pitch pattern distance which is the distance between a pitch pattern of an input speech and pitch pattern templates. Speaker-dependent word recognition experiments were carried out using 216 Japanese words uttered by five male and five female speakers. The proposed technique reduces the recognition error rate by 40% compared with the conventional method using only phoneme likelihood.

MIMO (multiple-input multiple-output) transmission is a promising technology to improve the frequency usage efficiency in mobile radio communications. In this letter, MIMO transmission with eigenbeam space division multiplexing (E-SDM) is focused on and the site-dependent beamforming characteristics is examined to know the transmission characteristics. Site-dependent radiowave direction of arrival and impulse responses are obtained using ray tracing. Result shows that effect of radiowaves with longer delays is enhanced due to E-SDM beamforming, and rather more capability of treating a longer excess delay is necessary for time-domain multipath compensation.

Measurements of delay spread were performed at microwave frequencies of 3.35, 8.45 and 15.75 GHz along quasi line-of-sight streets in metropolitan Tokyo. It is found that the delay spreads increase with the measurement distance and reach around 600 ns up to 1 km. It is also confirmed that a cumulative probability of the delay spreads follows a log-normal distribution. The gradients of delay spreads against the distance are greater for a lower mobile antenna height hm = 1.6 m than for hm = 2.7 m in these measurements because of blocking effect by the traffic of vehicles and pedestrians on the road. When the mobile antenna height is 2.7 m, the delay spreads within the range before the break points are observed relatively small: 90 ns (3.35 GHz), 140 ns (8.45 GHz) and 150 ns (15.75 GHz) at the cumulative probability of 90%. The gradients of delay spreads against the distance are greater for wider streets in our measurements.

For mobile telecommunication systems, it is important to accurately predict the propagation-path loss in terms of the estimation of the radiowave coverage area. The propagation-path loss has been estimated in a median obtained spatially from many received amplitudes (envelopes) within a region of several tens times as long as the wavelength, rather than in the envelopes themselves. Although ray tracing can obtain the envelopes and their median that reflect the site-dependent characteristics, the estimated median sometimes does not agree with the measured one. Therefore, the accuracy improvement has been expected. In this paper, an accuracy improvement is achieved by substituting a median with random phases for the median obtained spatially from many envelopes. The characteristic function method is used to obtain the cumulative distribution function and the median analytically where the phases are randomized. In a multipath environment, the phase-estimation error accompanying the location error of the ray tracing input influences the spatially obtained median. The phase-randomizing operation reduces the effects of the phase-estimation error on the median prediction. According to our estimation, improvements in accuracy of 4. 9 dB for the maximum prediction error and 2. 9 dB for the RMS prediction error were achieved. In addition, a probability-based cell-design method that takes the radiowave arrival probability and the interference probability into consideration is possible by using the percentiles obtained by the characteristic function method and the cell-design examples are shown in this paper.

This paper discusses microwave path-loss characteristics as a function of mobile antenna height in an urban line-of-sight environment. Measurements were made in metropolitan Tokyo with high-density buildings, using base station antenna heights of 4 and 8 m. We describe the path-loss characteristics of vehicle-mounted mode (mobile antenna height is 2.7 m) and portable mode (mobile antenna heights are 1.6 and 0.5 m). Dependence of path loss on the distance between base and mobile stations was analyzed. This reveals that the break points shift to the near side in the vehicle-mounted mode. This phenomenon can be interpreted by the existence of an effective height h of the road. The typical value of h was found approximately 1.4 m. In the portable mode, on the other hand, break points were not observed. The mobile antenna heights (1.6 and 0.5 m) in this mode are close to or less than the average height (1-2 m) of pedestrians on the sidewalk; and the received waves at the mobile station are often disturbed by pedestrians. This explains the nonexistence of break points in portable mode. The average attenuation coefficients is observed 3.2 in this mode. The attenuation coefficients tend to be larger at lower base station antenna heights and narrower road widths.

Future high-resolution short-range automotive radar will have a higher false alarm probability than the conventional low-resolution radar has. In a high-resolution radar, the reception signal becomes sensitive to the difference between intended and unintended objects. However, automotive radars must distinguish targets from background objects that are the same order of size; it leads to an increase in the false alarm probability. In this paper, a CFAR circuit for obtaining the target mean power, as well as the background mean power, is proposed to reduce the false alarm probability for high-resolution radars working in automotive environments. The proposed method is analytically evaluated with use of the characteristic function method. Spatial correlation is also considered in the evaluation, because the sizes of the both target and background objects approach the dimension of several range cells. Result showed the proposed CFAR with use of two alongside range cells could reduce the ratio of 6.4 dB for an example of an automotive situation.

Users require speech recognition systems that offer rapid response and high accuracy concurrently. Speech recognition accuracy is degraded by additive noise, imposed by ambient noise, and convolutional noise, created by space transfer characteristics, especially in distant talking situations. Against each type of noise, existing model adaptation techniques achieve robustness by using HMM-composition and CMN (cepstral mean normalization). Since they need an additive noise sample as well as a user speech sample to generate the models required, they can not achieve rapid response, though it may be possible to catch just the additive noise in a previous step. In the previous step, the technique proposed herein uses just the additive noise to generate an adapted and normalized model against both types of noise. When the user's speech sample is captured, only online-CMN need be performed to start the recognition processing, so the technique offers rapid response. In addition, to cover the unpredictable S/N values possible in real applications, the technique creates several S/N HMMs. Simulations using artificial speech data show that the proposed technique increased the character correct rate by 11.62% compared to CMN.

Recognition errors caused by out-of-vocabulary (OOV) words lead critical problems when developing spoken language understanding systems based on automatic speech recognition technology. And automatic vocabulary adaptation is an essential technique to solve these problems. In this paper, we propose a novel and effective automatic vocabulary adaptation method. Our method selects OOV words from relevant documents using combined scores of semantic and acoustic similarities. Using this combined score that reflects both semantic and acoustic aspects, only necessary OOV words can be selected without registering redundant words. In addition, our method estimates probabilities of OOV words using semantic similarity and a class-based N-gram language model. These probabilities will be appropriate since they are estimated by considering both frequencies of OOV words in target speech data and the stable class N-gram probabilities. Experimental results show that our method improves OOV selection accuracy and recognition accuracy of newly registered words in comparison with conventional methods.

In an AMS/OFDM system, base station is in control of the modulation level of each subcarriers, and then, adaptive modulated packet is transmitted from the base station to the mobile station. In this case, the mobile station is required the modulation level information (MLI) to demodulate the received packet. The MLI is generally transmitted as a data symbol, therefore, the throughput is degraded. Moreover, it is necessary to have some transmission delay times and the processing time to make an adaptive modulation command (AMC) using feedback information (FBI). With the FBI delay and processing time, the system performance might be degraded. To reduce these problems, in this paper, we propose a differential modulated pilot symbol assisted adaptive OFDM for reducing the MLI with predicted FBI.

Signal path loss and propagation delay spread were measured at microwave frequencies of 3.35, 8.45, and 15.75 GHz along a straight quasi line-of-sight (LOS) street in an urban environment under different traffic conditions: daytime and nighttime. Comparison between daytime and nighttime measurements reveals that the break points shift toward the base station because of the increase in the effective heights of the road and sidewalk; break points were not seen during the daytime at a mobile antenna height (hm) of 1.6 m. According to the cumulative probabilities of the delay spreads during the nighttime, frequency dependence is not clearly observed and the delay spreads for hm = 1.6 m were clearly larger than those for hm = 2.7 m. This is because a lower hm results in stronger blocking of the LOS wave, as was also observed during the daytime. The plot of path losses versus delay spreads is confirmed to be represented by an exponential curve. The exponential coefficients during the daytime were observed to be greater than those during the nighttime. This indicates that a LOS wave is more likely to be blocked during the daytime.

This paper describes a method of predicting transmission performance to be obtained by applying RAKE reception and parallel transmission in realistic urban multipath environments. Delay profiles measured in metropolitan Tokyo at microwave frequencies were used to obtain the impulse responses of radio channels, and the closed-form equations corresponding to the performance of these anti-multipath techniques were derived, by means of the characteristic function method, under the assumption that the phases of the impulse responses are uniformly distributed. Results show that RAKE reception provides bit error rates 100 times lower than bare transmission does, whereas the improvement obtained by using parallel transmission should be especially valuable in broadband communication systems, such as those operating at data rates above 10 Mb/s.