摘要 :
In this paper, we propose a novel design of evolving fuzzy classifiers (EFCs) to handle online multiclass classification problems in a data-streaming context. Therefore, we exploit the concept of all-pairs (AP), a.k.a. all-versus-...
展开
In this paper, we propose a novel design of evolving fuzzy classifiers (EFCs) to handle online multiclass classification problems in a data-streaming context. Therefore, we exploit the concept of all-pairs (AP), a.k.a. all-versus-all, classification using binary classifiers for each pair of classes. This benefits from less complex decision boundaries to be learned, as opposed to a direct multiclass approach, and achieves a higher efficiency in terms of incremental training time than one-versus-rest classification techniques. For the binary classifiers, we apply fuzzy classifiers with singleton class labels in the consequences, as well as Takagi–Sugeno (T–S) fuzzy models to conduct regression on [0, 1] for each class pair. Both are evolved and incrementally trained in a data-streaming context, yielding a permanent update of the whole AP collection of classifiers, thus being able to properly react to dynamic changes in the streams. The classification phase considers a novel strategy by using the preference levels of each pair of classes that are collected in a preference relation matrix and performing a weighted voting scheme on this matrix. This is done by investigating the reliability of the classifiers in their predictions: 1) integrating the degree of ignorance on samples to be classified as weights for the preference levels and 2) new conflict models used in the single binary classifiers and when calculating the final class response based on the preference relation matrix. The advantage of the new EFC concept over the single model (using a direct multiclass classification concept) and multimodel architectures (using a one-versus-rest classification concept) will be underlined by empirical evaluations and comparisons at the end of this paper based on high-dimensional real-world multiclass classification problems. The results also show that integrating conflict and ignorance concepts into the preference relations can boost classifier accuracies.
收起
摘要 :
This paper presents a 16-element phased-array receiver for 76–84-GHz applications with built-in self-test (BIST) capabilities. The chip contains an in-phase/quadrature (I/Q) mixer suitable for automotive frequency-modulation cont...
展开
This paper presents a 16-element phased-array receiver for 76–84-GHz applications with built-in self-test (BIST) capabilities. The chip contains an in-phase/quadrature (I/Q) mixer suitable for automotive frequency-modulation continuous-wave radar applications, which is also used as part of the BIST system. The chip achieves 4-bit RF amplitude and phase control, an RF to IF gain of 30–35 dB at 77–84 GHz, I/Q balance of $< {hbox {1 dB}}$ and $< {hbox{10}}^{circ }$ at 76–84 GHz, and a system noise figure of 18 dB. The on-chip BIST covers the 76–84-GHz range and determines, without any calibration, the amplitude and phase of each channel, a normalized frequency response, and can measure the gain control using RF gain control. System-level considerations are discussed together with extensive results showing the effectiveness of the on-chip BIST as compared with standard $S$-parameter measurements.
收起
摘要 :
This paper presents a W-band wafer-scale phased- array transmitter with high-efficiency on-chip antennas. The 4$,times,$4 array is based on an RF beamforming architecture with an equiphase distribution network and phased shifters...
展开
This paper presents a W-band wafer-scale phased- array transmitter with high-efficiency on-chip antennas. The 4$,times,$4 array is based on an RF beamforming architecture with an equiphase distribution network and phased shifters placed on every element. The differential on-chip antennas are implemented using a 100 $mu{hbox {m}}$ thick quartz superstrate and with a simulated efficiency of $sim {hbox {45}}%$ at 110 GHz. The phased array is designed with low mutual coupling between the elements and results in a stable active antenna impedance versus scan angle. The phased array is built in the Jazz SBC18H3 SiGe BiCMOS process, and is 6.5$,times ,$6.0 ${hbox {mm}}^{2}$. Measurements show two-dimensional pattern scanning capabilities with a directivity of 17.0 dB, an array gain of $sim {hbox {26.5}}~{hbox {dB}}$ at 110 GHz, and an EIRP of 23–25 dBm at 108–114 GHz. The power consumption is 3.4 W from a 1.9 V supply. To our knowledge, this work represents the first W-band wafer-scale phased array to-date. The application areas are in point-to-point communication systems in the 100–120 GHz range.
收起
摘要 :
A highly sensitive CMOS-based sensing system is proposed for permittivity detection and mixture characterization of organic chemicals at microwave frequencies. The system determines permittivity by measuring the frequency differen...
展开
A highly sensitive CMOS-based sensing system is proposed for permittivity detection and mixture characterization of organic chemicals at microwave frequencies. The system determines permittivity by measuring the frequency difference between two voltage-controlled oscillators (VCOs); a sensor oscillator with an operating frequency that shifts with the change in tank capacitance due to exposure to the material under test (MUT) and a reference oscillator insensitive to the MUT. This relative measurement approach improves sensor accuracy by tracking frequency drifts due to environmental variations. Embedding the sensor and reference VCOs in a fractional-$N$ phase-locked loop (PLL) frequency synthesizer enables material characterization at a precise frequency and provides an efficient material-induced frequency shift read-out mechanism with a low-complexity bang–bang control loop that adjusts a fractional frequency divider. The majority of the PLL-based sensor system, except for an external fractional frequency divider, is implemented with a 90-nm CMOS prototype that consumes 22 mW when characterizing material near 10 GHz. Material-induced frequency shifts are detected at an accuracy level of 15 ${hbox{ppm}}_{rm rms}$ and binary mixture characterization of organic chemicals yield maximum errors in permittivity of ${<}$1.5%.
收起
摘要 :
The use of multiphase motor drives is an increasingly important strategy nowadays. These multiphase machines are usually modeled by a reference frame transformation to avoid the cross-coupling of variables. This transformation dec...
展开
The use of multiphase motor drives is an increasingly important strategy nowadays. These multiphase machines are usually modeled by a reference frame transformation to avoid the cross-coupling of variables. This transformation decomposes the original $n$-dimensional vector space into orthogonal subspaces. Mapping the voltage and current harmonics into the subspaces in distributed machines is important because it allows to identify which components are related to the torque and which ones just increase the machine losses. The sequence identification of each harmonic is also important in closed-loop current harmonic compensation to set the controllers. In addition, the harmonic mapping is interesting in multimotor systems to know how harmonics from one machine can affect the other machines in the system. In this paper, a simple graphical method for time harmonic subspace and sequence identification is proposed. This method is valid for symmetrical machines of any phase number $n$, it provides full subspace and sequence identification and it can be used in multimotor systems. Experimental results using a five- and a six-phase motor in single-drive configuration and a series-connected two-motor six-phase drive validate the proposed method.
收起
摘要 :
A general solution of the magnetic field in the airgap of conventional and alternate field-excited switched-flux (FE-SF) machines is proposed in this paper. The analytical model is based on the subdomain method. It involves the so...
展开
A general solution of the magnetic field in the airgap of conventional and alternate field-excited switched-flux (FE-SF) machines is proposed in this paper. The analytical model is based on the subdomain method. It involves the solution of governing field equations in a doubly slotted airgap using the variable separation method. The complete model is derived and described in a general manner so that it can be easily extended to unconventional FE-SF topologies. By means of example, analytical predictions of airgap field are extensively compared and validated using 2D FE results. FE simulations were performed on a 24-10 classical FE-SF structure and also on a novel 18-11 FE-SF machine with additional spacer teeth.
收起
摘要 :
Battery emulation with a controllable high-power dc supply enables repeatable hardware-in-the-loop testing of powertrains for hybrid and electric vehicles. For this purpose, not only the power flow but also the nonlinear character...
展开
Battery emulation with a controllable high-power dc supply enables repeatable hardware-in-the-loop testing of powertrains for hybrid and electric vehicles. For this purpose, not only the power flow but also the nonlinear characteristic and dynamic impedance of batteries need to be emulated. In this paper, nonlinear local model networks are used to obtain dynamic battery models with high fidelity that can be computed in real time. This approach also allows the extraction of local linear impedance models for high-bandwidth impedance emulation, leading to a tighter coupling between the test bed and simulation model with predictable closed-loop dynamics. A model predictive controller that achieves optimal control with adherence to system constraints is extended to impedance control and robustness against constant power loads. This results not only in superior dynamic performance but also in stable dc-bus voltage control even for testing of tightly controlled electric motor inverters with negative differential input resistance. Since the controller design is based on a model of the test bed setup including the virtual battery model, emulator hardware, and input characteristics of the powertrain under test, it is possible to systematically analyze stability.
收起
摘要 :
The deployment of coastal observatories motivates the development of acoustic inversion schemes able to characterize rapidly time-varying range-dependent environments. This paper develops feature models as parameterization schemes...
展开
The deployment of coastal observatories motivates the development of acoustic inversion schemes able to characterize rapidly time-varying range-dependent environments. This paper develops feature models as parameterization schemes for the range-dependent temperature field, when the latter is mainly influenced by an identified oceanic feature, here thermal fronts. The feasibility of feature-oriented acoustic tomography (FOAT) is demonstrated in two cases of coastal thermal front known to occur regularly: the Ushant tidal front, France (48.5$^{circ}$ N, 5$^{circ}$ E), and the Cabo Frio coastal upwelling, Brazil (23$^{circ}$ S, 42$^{circ}$ W). Realistic scenarios simulated with regional circulation models provide typical environmental variations for testing the validity of the FOAT approach, with both global optimization and sequential filtering of the (synthetic) full-field acoustic data. Matched-field processing at multiple frequencies is used to reduce ambiguities between parameters and to achieve a good tradeoff between robustness and sensitivity. The proposed feature-model parameterization is shown to provide robust estimates of the 2-D temperature field even when the simulated environment presents smaller scale inhomogeneities. Moreover, the sequential filtering based on a random walk model of the thermal front parameters enables a stable tracking of typical temperature field variations along several days. This sequential approach is particularly convenient for continuous, long-term monitoring operated with bottom-moored ocean observatories.
收起
摘要 :
Image registration (IR) is a challenging topic in both the computer vision and pattern recognition fields; its main aim is to find the optimal transformation to provide the best overlay or fitting between two or more images. Usual...
展开
Image registration (IR) is a challenging topic in both the computer vision and pattern recognition fields; its main aim is to find the optimal transformation to provide the best overlay or fitting between two or more images. Usually, the success of well-known algorithms, such as iterative closest point, highly depends on several assumptions, e.g., the user should provide an initial near-optimal pose of the images to be registered. In the last decade, a new family of registration algorithms based on evolutionary principles has been contributed in order to overcome the latter drawbacks. However, their performance highly depends on carefully tuning (usually by hand) the control parameters of the algorithm, which is an error-prone and a time-consuming task. In this paper, we propose a new self-adaptive evolution model to deal with IR problems. To our knowledge, this is the first time a self-adaptive approach has been used for tuning the control parameters of evolutionary algorithms tackling computer vision tasks. Specifically, we introduce a novel design of the proposed self-adaptive approach facing pair-wise range IR problem instances, which is a challenging real-world optimization problem. In addition, several classical approaches, as well as state-of-the-art evolutionary IR methods, have been considered for numerical comparison.
收起
摘要 :
Background modeling is challenging due to background dynamism. Most background modeling methods fail in the presence of intensity changes, because the model cannot handle sudden changes. A solution to this problem is to use intens...
展开
Background modeling is challenging due to background dynamism. Most background modeling methods fail in the presence of intensity changes, because the model cannot handle sudden changes. A solution to this problem is to use intensity-robust features. Despite the changes of an edge's shape and position among frames, edges are less sensitive than a pixel's intensity to illumination changes. Furthermore, background models in the presence of moving objects produce ghosts in the detected output, because high quality models require ideal backgrounds. In this paper, we propose a robust statistical edge-segment-based method for background modeling of non-ideal sequences. The proposed method learns the structure of the scene using the edges' behaviors through the use of kernel-density distributions. Moreover, it uses segment features to overcome the shape and position variations of the edges. Hence, the use of segments gives us local information of the scene, and that helps us to predict the objects and background precisely. Furthermore, we accumulate segments to build edge distributions, which allow us to perform unconstrained training and to overcome the ghost effect. In addition, the proposed method uses adaptive thresholding (in the segments) to detect the moving objects. Therefore, this approach increases the accuracy over previous methods, which use fixed thresholds.
收起
