We propose a new clustering algorithm based upon the maximin correlation analysis (MCA), a learning technique that can minimize the maximum misclassification risk. The proposed algorithm resembles conventional partition clustering algorithms such as k-means in that data objects are partitioned into k disjoint partitions. On the other hand, the proposed approach is unique in that an MCA-based approach is used to decide the location of the representative point for each partition. We test the proposed technique with typography data and show our approach outperforms the popular k-means and k-medoids clustering in terms of retrieving the inherent cluster membership.
In this paper, a new voltage-mode first-order all-pass filter is presented. The proposed filter provides both of non-inverting and inverting outputs at the same configuration. The circuit is composed of single universal voltage conveyor (UVC), operational transconductance amplifier (OTA), and capacitor. Considering the electronically tunability properties of the OTA phase responses of the proposed active-C circuit can be controlled by an external bias current. The circuit is suitable for wideband applications. All the active and passive sensitivities are low. The theoretical results are verified by PSPICE simulations using BJT implementations of UVC and OTA.
In mobile network design, the challenge is to efficiently determine the locations of base control stations (BSCs), mobile switching centers (MSCs), and their connecting links for given locations of base transceiver stations (BTSs) so that a predefined objective function is satisfied. In this paper, a particle swarm optimization- (PSO-) based optimization engine is used to effectively lay out the network components and their interconnections such that the overall deployment cost is kept as low as possible. The performance of the PSO-based engine is then compared with a genetic algorithm- (GA-) based one. The simulation results show that the PSO-based optimization engine is able to successfully optimize the network deployment cost and significantly outperforms the GA-based optimization engine.
A linear-phase FIR filter with more than two bands, designed by employing the Parks-McClellan algorithm, may exhibit an undesirable gain response in the transition band. In this paper, a practical FIR filter solution to such undesirable results is proposed, whereby a modified sampling kernel and the complementary filter concept are utilized. In particular, the proposed approach provides a closed-form design of type-1 linear-phase FIR filters satisfying the given filter specification.
This paper focuses on the analysis of a BJT based soil moisture sensor that outputs a voltage corresponding to the change in soil resistance. The measured sensor's sensitivity is 51.1mV/% with a deviation of 19.7mV from sensor to sensor. The deviation which is more than twice smaller than the sensitivity suggests that it is suitable to be used in a large array to detect gravimetric moisture in individual soil samples. Integration of read-out circuitry to the sensor shows that an array operation of the sensor is possible and can be used to aid in moisture detection in fertigation farms.
CASM is the abbreviation of Coherent Adaptive Subcarrier Modulation. CASM is one of the Combining Method of signals. It will be used in GALILEO satellite navigation system. In this paper, on the base of analyzing and studying CASM in theory, do some research in connection with the used case of CASM in E1 signals of GALILEO system. It is shown from the result that CASM is an effectual Combining Method of signals.
A novel versatile single FDCCII-based voltage-mode universal biquadratic filter with three inputs and four outputs is presented. The proposed circuit can act as a multifunction voltage-mode filter with two inputs and four outputs and can perform simultaneous realization of voltage-mode bandreject, highpass, bandpass and lowpass filter signals from the four output terminals, respectively. On the other hand, it also can act as a universal voltage-mode filter with three inputs and a single output and can realize five generic voltage-mode filter signals without requiring any inverting input voltage signals and component-matching conditions. The proposed circuit employing single FDCCII, two grounded capacitors and two resistors, which are the minimum components necessary for realizing a second-order voltage-mode universal filtering response from the same topology. Both its active and passive sensitivities are low.
Effects of an extra PdO layer in resistance-based hydrogen sensors are considered. P-type Si substrates were subjected to porous Si by electrochemical etching at room temperature. One category of samples has PdO layer in its structure while the other one does not. We have used electron-beam method for Pd deposition. Results show that when electron-beam technique is used for Pd deposition, existence of PdO layer will increase the range of sensors' operation up to 1% of hydrogen concentration. These samples will be saturated after 186seconds, while the samples without this layer can detect hydrogen up to 0.6% in 150 seconds response time.
The conventional second-order Delta-Sigma modulator (DSM) based switching converter (SC) cannot be used for full range (-1 to +1) of input signal. The DSM becomes unstable when the normalized input signal is about 0.4. The output of the second integrator saturates the operational amplifier which is used as comparator in the single-bit quantizer. The three- phase switching converter, based on DSM with hexagonal quantizer, improves the signal to noise ratio as well as the spectral response without improving the range of the input signal. This paper proposes three-phase signal dependant DSM based switching converter with vector quantizer. The proposed three-phase SC increases the input signal to full scale and maintains its stability. The SNR and spectral response are better than the conventional DSM based three-phase switching converter with hexagonal quantizer. The overall maximum percentage of difference in phase voltages is 0.4%.
This study presents a microelectromechanical voltage controlled oscillator based on the pull-in electromechanical contact of an electrostatic actuator. Micromachining process of nickel electroplating on a silicon substrate was used to develop an electrostatic torsion plate mechanism (200µm × 200µm in area, 15µm in thickness) that could be mechanically movable (0.1rad) by the electrostatic torque of applied voltage. An electromechanical stopper was designed underneath the movable plate such that the effective drive voltage was discharged upon the electrostatic pull-in contact. Peripheral electrical circuits were designed to compose a ring oscillator, in which the temporal response of electrostatic mechanism determined the oscillation frequency. The frequency became tunable as a function of the drive voltage. Frequency tuning range of 4.39∼9.6kHz was experimentally observed using the control voltage of 7.4∼10.6V.
In this paper, we analyze the bit error rate (BER) performance of a new modified structure of multitone code-division-multiple-access (MT-CDMA) system that is serial-to-parallel (S/P)-Replica MT-CDMA. It is shown that with the same bandwidth, the S/P-Replica MT-CDMA system outperforms the conventional system that we name it S/P MT-CDMA. We also propose a novel bandwidth-efficient-channel-coding-scheme (BECCS) based on super-orthogonal-code (SOC)-assisted serially concatenated turbo code (SCSOC) for S/P-Replica MT-CDMA system so that by using it, the coded system without extra bandwidth, significantly improves the performance of an uncoded system over fading channel.
We experimentally demonstrated phase-noise-tolerant 20-Gbit/s QPSK self-homodyne transmission over 80km using a spectrum-sliced lightwave from an inexpensive Fabry-Perot laser diode (FP-LD).We employed an optical phase-noise cancellation technique based on self-homodyne detection using a polarization-multiplexed pilot-carrier. A BER of < 10 -8 after 80-km transmission was successfully attained without using error correction. To the authors' best knowledge, this is the first experimental demonstration of multi-level modulation/demodulation using an FP-LD.
We experimentally investigated DGD- and dispersion-tolerance of QPSK self-homodyne detection based on a polarization-multiplexed pilot carrier for both NRZ and RZ data formats. The results show that the tolerance is large enough to be used in deployed network systems. We also investigated receiver sensitivity penalty versus the ratio of pilot-carrier power to total optical power. The ratio had a detunable margin of as large as ±15% at around the optimum point of 50%.
An analytical approach is presented to investigate the performance of matrix amplifiers containing both hetero-junction bipolar and high-electron-mobility transistors. An analytical relation for attenuation function of the central line of the matrix amplifier is developed which can be employed in the consideration of the influence of several parameters on the device performance. It is shown that this structure may have lower attenuation respect to the traditional HEMT_HEMT and HEMT_HBT matrix amplifiers. The proposed technique is applicable in improving the performance of Microwave and mm-wave ICs.
This paper presents a high speed 1Gb GDDR3 Graphics DRAM using data bus inversion (DBI) DC mode in order to achieve low power and low noise in DRAM. A DBI, digital majority voter (DMV) circuit and the Global I/O (GIO) control circuit on the DBI DC mode are newly proposed. In this DMV, The current of GIO toggle pattern is consumed less than 47% compared with the analog majority voter (AMV). The voltage fluctuation wave form of the data eye is also reduced in accordance with DBI on the operation mode. Using the proposed DBI scheme can produce almost stable signal integrity of the DQs against high speed operation. The DBI is fabricated using 54nm technology.