Comparing Visual Feature Coding for Learning Disjoint Camera Dependencies

Determination of Relevant Neuron–Neuron Connections for Neural Prosthetics Using Time-Delayed Mutual Information: Tutorial and Preliminary Results

Performance of Automatic Speech Recognition (ASR) systems drops dramatically when transcribing conversations recorded in noisy conditions. Speech analytics suffer from this poor automatic transcription quality. To tackle this difficulty, a solution consists in mapping transcriptions into a space of hidden topics. This abstract representation allows to substantiate the drawbacks of the ASR process. The well-known and commonly used one is the topic-based representation from a Latent Dirichlet Allocation (LDA). Several studies demonstrate the effectiveness and reliability of this high-level representation. During the LDA learning process, distribution of words into each topic is estimated automatically. Nonetheless, in the context of a classification task, no consideration is made for the targeted classes. Thus, if the targeted application is to find out the main theme related to a dialogue, this information should be taken into consideration. In this paper, we propose to compare a classical topic-based representation of a dialogue, with a new one based not only on the dialogue content itself (words), but also on the theme related to the dialogue. This original representation is based on the author-topic (AT) model. The effectiveness of the proposed representation is evaluated on a classification task from automatic dialogue transcriptions between an agent and a customer of the Paris Transportation Company. Experiments confirmed that this author-topic model approach outperforms by far the classical topic representation, with a substantial gain of more than 7% in terms of correctly labeled conversations.

In order to solve the potential issues caused by network induced delays and dropouts which could arise the performance degradation and system instability, this paper studies the stochastic stability problem of Networked control systems (NCSs) with arbitrary time delays and packet dropouts by using an active time-varying sampling method. The random time delays and successive packet dropouts are driven by two separately Markov chains and NCSs are modelled as a discrete time Markovian jump linear systems. Based on Lyapunov approach, sufficient conditions for the stochastic stability of the networked control system are derived and stabilization controller is designed in terms of linear matrix inequalities (LMIs) correspondingly. Gridding approach is introduced to guarantee the solvability of the LMIs with finite jump modes. A numerical example is given to illustrate the effectiveness of the proposed method which stabilizes the NCS with random time delays and packet dropouts.

The shifted Jacobi delay operational matrix is introduced, and applied to the approximation of the solutions of linear systems with arbitrary time delay. The parameter-identification problem for time-delay systems is also dealt with. The new method is algebraic and computer-oriented. Two examples are given, and the results are very accurate and satisfactory.

SummaryWith the increasing industrial requirements such as bigger size object, stable operation, and complex task, multilateral teleoperation systems extended from traditional bilateral teleoperation are widely developed. In this paper, the integrated control design is developed for multilateral teleoperation systems, where n master manipulators are operated by human to remotely control n slave manipulators cooperatively handling a target object. For the first time, the control objectives of multilateral teleoperation including stability, synchronization, transparency, and internal force distribution are clarified systematically. A novel communication architecture is proposed to cope with communication delays, where the estimated environmental parameters are transmitted from the slave side to the master, to replace the traditional environmental force measurement in the communication channel. A kind of nonlinear adaptive robust control technique is used to deal with nonlinearities, unknown parameters, and modeling uncertainties existing in the master, slave, and environmental dynamics, so that the excellent tracking performance is achieved in both master and slave sides. The coordinated motion/force control is designed in the slave side by the optimal internal force distribution among n slave manipulators, and the impedance control is designed in the master side to realize the target transparency behavior. In summary, the proposed control algorithm can achieve the guaranteed robust stability, the excellent synchronization and transparency performance, and the optimal internal force distribution simultaneously for multilateral teleoperation systems under arbitrary time delays and various modeling uncertainties. The simulation is carried out on a 2‐master/2‐slave teleoperation system, and the results show the effectiveness of the proposed control design. Copyright © 2015 John Wiley & Sons, Ltd.

We present quantum-kinetic calculations for pump-and-probe signals of bulk semiconductors based on the second Born approximation for electron--LO-phonon interaction. The relaxation process of laser-excited electron-hole pairs is studied along the whole frequency domain including the exciton resonance. This allows a description of carrier cooling due to LO phonons at arbitrary long time delays, from initial hole burning at the pump frequency until exciton bleaching after accumulation near the band extrema. At low temperature, a bottleneck effect is predicted for small pump intensities. At room temperature, signatures of upward scattering become visible. The numerical results for GaAs are in good qualitative agreement with recent measurements [A. Leitenstorfer et al., Phys. Rev. Lett. 76, 1545 (1996); C. F\"urst et al., Phys. Rev. Lett. 78, 3733 (1997)].

One potential limitation of the LCLS is the maximum frequency—120 Hz— at which the individual pulses are produced. This maximum frequency is dictated by the fundamental design of the original linear accelerator and serves as a limit on the “frames per second” that can be illuminated by the beam. A potential solution for accessing ultrafast dynamics is by temporally splitting the X-ray pulses into two pulses with arbitrary relative intensity and time delay. In this manner, dynamics ranging from the X-ray pulse duration (between one and several hundred femtoseconds) and the maximum relative pulse delay can be explored by evaluating the X-ray response collected as a function of the relative pulse delay.

This chapter introduces and surveys bilateral (master-slave) control methods and system architectures for telerobotics known in the literature. The discussion takes a focus most relevant to subsequent chapters, see [1, 2, 3] for complementary and more complete surveys. Firstly, an introductory example of a control-oriented block diagram formulation for a typical two-channel force-velocity architecture is presented, where the human system interface (master) is force controlled and the teleoperator (slave) is velocity controlled. Then a network theoretic view using n-ports is taken and the basic concept of passivity is explained meaning that an interconnection of passive subsystems results in an overall passive system. Transparency is defined, meaning that the human operator should ideally feel as if directly acting in the remote environment (is not able to feel the technical systems/communication network at all). The well known and for ideal transparency required four-channel architecture is presented. The important discussion of time delay in the communication network, which impairs or might threaten stability properties of the closed loop teleoperation system, is then discussed. The most successful approach using the scattering (wave variable) transformation approach is presented, which guarantees that the communication two-port is lossless (passive) for arbitrary constant time delay. Finally, we mention the key challenges in the design of bilateral teleoperation systems and give a short preview of the subsequent chapters addressing some of these challenges.

In this study, we proposed a new method for the detection of abnormal event based on the social force model (SFM), combined with the local density information of the crowd. The method extracts the local density of the people based on the feature point clustering algorithm. The Latent Dirichlet Allocation (LDA) model is established based on the bag of words method combined with the temporal and spatial features of visual words, then identify the abnormal event using the maximum likelihood function.

The paper proposes a novel delay locked loop (DLL) for an ultra wideband-impulse radio (UWB-IR) system with time hopping/spread spectrum (TH/SS), and presents its characteristics. The DLL circuit adopts a discriminator composed of two correlators with arbitrary time delay. Jitter, lose lock time, and bit error rate (BER) performances are evaluated by computer simulations. The results show that these performances of the proposed method are superior to those of a conventional method.

For the practical multisensor systems, the communication channels are often unreliable, especially in large, wireless, multi hop sensor networks. Thus, the communication delays often appear and produce random and unknown influence upon the performance of fusion estimators. Recently, data fusion for "Out-of-sequence" measurements (OOSM) attracts a mass of attentions gradually and some filters based on single sensor and multiple sensors have been reported. However, the existing methods for OOSM problem have several shortcomings, such as limited application for single sensor, bad real-time performance for the smooth property, low fusion estimate for neglecting the delayed information, and so forth. To tackle these problems, a novel batch fusion estimator based on the relative measurements for multisensor tracking systems with OOSM is proposed. The estimator is optimal in linear minimum mean square error (LMMSE). The paper also presents the suboptimal recursive form of the proposed estimator. Its principle is to establish the relative measurement equations of delayed observations to the next latest systemic state. Compared with some existingfusion methods for OOSM, the proposed method with arbitrary time delays has the best total performance. At the same time, algorithm analysis and computer simulation validate the advantages of the proposed fusion estimator.

Two- and three-point (in time) statistics of fission chamber signals for multiplicity counting with thermal neutrons

Crowd stability analysis is one of research hotspots to alleviate the severe situation of stampede accidents worldwide. Different from the conventional analysis models for crowd stability based on pedestrian density, this study analyses the characteristics of external disturbances and internal obstacle disturbance based on Lyapunov's theory. The critical range of crowd acceleration in crowd evacuation is obtained, a crowd merging acceleration-critical density time delay model is established, and a stability criterion of acceleration vector based on Lyapunov is obtained based on Lyapunov stability analysis. This provides new information for ensuring the stability of crowd movement in public places, assessing the stability of the crowd in the area, and taking reasonable protection and guidance measures prior to instability of a crowd flow.

Person reidentification (Re-ID) matches images of the same person captured in disjoint camera views and at different times. To obtain a reliable similarity measurement between images, manually annotating a large amount of pairwise cross-camera-view person images is deemed necessary. However, this kind of annotation is both costly and impractical for efficiently deploying a Re-ID system to a completely new scenario, a new setting of nonoverlapping camera views between which person images are to be matched. To solve this problem, we consider utilizing other existing person images captured in other scenarios to help the Re-ID system in a target (new) scenario, provided that a few samples are captured under the new scenario. More specifically, we tackle this problem by jointly learning the similarity measurements for Re-ID in different scenarios in an asymmetric way. To model the joint learning, we consider that the Re-ID models share certain component across tasks. A distinct consideration in our multitask modeling is to extract the discriminant shared component that reduces the cross-task data overlap in the shared latent space during the joint learning, so as to enhance the target inter-class separation in the shared latent space. For this purpose, we propose to maximize the cross-task data discrepancy on the shared component during asymmetric multitask learning (MTL), along with maximizing the local inter-class variation and minimizing local intra-class variation on all tasks. We call our proposed method the constrained asymmetric multitask discriminant component analysis (cAMT-DCA). We show that cAMT-DCA can be solved by a simple eigen decomposition with a closed form, getting rid of any iterative learning used in most conventional MTL analyses. The experimental results show that the proposed transfer model gains a clear improvement against the related nontransfer and general multitask person Re-ID models.

For many multivariable systems, because of the strong cross coupling between controlled variables, it is important to design the decoupling controller. The paper presents a decoupling design of multivariable generalised predictive control by combining feedforward control with multivariable generalised predictive control, which can realise decoupling control for systems with arbitrary time delay structure.