A real-time image-tracking algorithm is proposed, which gives small weights to pixels farther from the object center and uses the quantized image gray scales as a template. It identifies the target s location by the mean-shift iteration method and arrives at the target s scale by using image feature recognition. It improves the kernel-based algorithm in tracking scale-changing targets. A decimation method is proposed to track large-sized targets and real-time experimental results verify the effectiveness of the proposed algorithm.

In order to satisfy the need of diagnoses, based on the characteristic of medical images that a sequence of frames are formed in one body inspection, a new strategy for medical images compression is proposed. The 3-D wavelet is adopted and the planar zerotree is extended to the 3-D zerotree. By making use of the 3-D zerotree structure, a simple method for region of interest (ROI) mask generation is put forward. Medical images are compressed by three-dimensional embedded coding with the compression of regions of interest. Simulation results have shown that it can efficiently improve the compression ratio without affecting the diagnoses.

Obtaining the location of an unknown node accurately is a key problem of a locating service under a ubiquitous computing environment. The paper proposes and proves three theorems of location reference node placement according to the analysis of the location error produced during location using a polygon location method and three important characteristics of chaos dynamics. Based on the three theorems, the location reference node selection (LRNS) algorithm is proposed by improving on the traditional polygon location algorithm. The simulation results indicate that the reference node placement theorems and the LRNS algorithm can meet the requirements of a ubiquitous terminal s real-time location and possess a preferable precision in location.

Question-answering systems provide short answers with the use of available information. The implementation mechanism for a question answering system is presented in this paper and is based on concepts and statistics. The system determines the question and focuses on the answer types, making different conceptual expansions for different questions. It applies the latent semantic indexing (LSI) method to retrieve relevant passages. It uses matching algorithms to find a match between questions and sentences stored in a database. It also extracts answers from a frequently asked questions (FAQ) database by finding matching or similar sentences. The answering ability of the system has been improved with the use of LSI and FAQ. The question-answering system introduced in Chinese universities is a developed and proven system capable of precise results.

Even if two implementations of a protocol pass the conformance testing, it cannot guarantee that they can interoperate properly; so direct testing of interoperation is considered indispensable. During the interoperability testing, a minimal number of test sequences are expected to check as many as possible implementation errors. By using minimalcomplete-coverage criterion, the test sequence generation based on digraph can produce more effective test sequences.

By studying the core algorithm of a three-dimensional discrete wavelet transform (3-D DWT) in depth, this paper divides it into three one-dimensional discrete wavelet transforms (1-D DWTs). Based on the implementation of a 3-D DWT software, a parallel architecture design of a very large-scale integration (VLSI) is produced. It needs three dual-port random-access memory (RAM) to store the temporary results and transpose the matrix, then builds up a pipeline model composed of the three 1-D DWTs. In the design, the finite state machine (FSM) is used well to control the flow. Compared with the serial mode, the experimental results of the post synthesized simulation show that the design method is correct and effective. It can increase the processing speed by about 66%, work at 59 MHz, and meet the real-time needs of the video encoder.

In this paper, an integrated system of segmenting unconstrained handwritten numeral strings with unknowing number of digits is proposed, which consists of the extraction of connected components based on vertical projection and isolated components analysis, the length estimation of connected components using syntax analysis and waveform analysis and the segmentation of unconstrained connected handwritten numeral strings using innovative reverse drop-falling  algorithm. This segmentation system which has promising results is then incorporated into a complete bank check character recognition system.

The modeling and simulation (M&S) architecture describes and defines the relationship between the different parts of a simulation. The simulation system architecture and simulation support platform architecture are discussed separately. The simulation support platform architecture consists of the management layer, the resource layer, the communication layer, the application layer and the infrastructure layer. The best way is to design and realize the M&S collaborative environment for simulation support platform in the resource-communication-application three-dimension space.

Conventional adaptive transmission schemes perform poorly in wireless correlated slow-fading channels. A cross-layer adaptive transmission scheme combined with selective repeat automatic repeat request (SR-ARQ) is proposed. We apply a multi-state Markov system model for analyzing the performance of systems and optimizing the selection of modulation levels and packet sizes in correlated fading channels, which is also described by a finite-state Markov chain. A general closed-form expression of the average throughput for our suggested scheme is presented. Numerical results show that our adaptive scheme combined with SR-ARQ can obtain good performance in correlated fading channels.

Based on the transmitting power of a mobile station and the maximum cell capacity, the analysis of a wideband code-division multiple access (WCDMA) standard and the comparison of the corresponding parameters with CDMA2000 or evolution-data only (EV-DO) are given in this paper. The result proves that the mobile station of WCDMA cannot provide the required transmitting power on the specified frequency band and data rate, and the symmetrical capacity in an up-downlink cannot meet the requirement of a mobile Internet. It is questionable for the International Telecommunication Union (ITU) to select the frequency band of 2 GHz for vehicular mobile telecommunication and unreasonable to distribute it over the downlink and uplink channels symmetrically. Only CDMA2000 1X EV-DO or enhanced data rate for GSM evolution (EDGE), which adopts the Time Division Multiple Access (TDMA) mode, can provide the high-speed data rate for a wireless access.

In this paper, a distributed topology control algorithm is proposed. By adjusting the transmission power of each node, this algorithm constructs a wireless network topology with minimum-energy property, i.e., it preserves a minimum-energy path between every pair of nodes. Moreover, the proposed algorithm can be used in both homogenous and heterogeneous wireless networks, and it can also work without an explicit propagation channel model or the position information of nodes. Simulation results show that the proposed algorithm has advantages over the topology control algorithm based on direct-transmission region in terms of average node degree and power efficiency.

This study proposes a new multiple access control protocol named distributed synchronous reservation multiple access control protocol, in which the hidden and exposed terminal problems are solved, and the quality of service (QoS) requirements for real-time traffic are guaranteed. The protocol is founded on time division multiplex address and a different type of traffic is assigned to different priority, according to which a node should compete for and reserve the free slots in a different method. Moreover, there is a reservation acknowledgement process before data transmit in each reserved slot, so that the intruded terminal problem is solved. The throughput and average packets drop probability of this protocol are analyzed and simulated in a fully connected network, the results of which indicate that this protocol is efficient enough to support the real-time traffic, and it is more suitable to MANETs.

As the traditional methods were not suitable for the detection of small distribute denial of service (DDoS) attack and identification of busy traffic, on the basis of the influence of DDoS attack, one wavelet analysis method was proposed. Wavelet method of coefficient variance analysis was deduced and a software model for the method was designed. In addition, key issues of the choice of wavelet and calculation of Hurst were resolved. The experimental results show that the proposed method has more advantages in accurately identifying busy traffic and detection of small DDoS attack.

The experimental results of a 2?2 integrated thermooptic Mach-Zehnder interferometer (MZI) switch based on multimode interference (MMI) couplers are reported. The insertion loss (IL) is 3.40 dB, the polarization dependent loss is 0.47 dB, the extinction ratio (ER) at bar state and cross state is 32.01 dB and 16.42 dB respectively, the response time is less than 3 ms, and the power consumption is 658 mW. Regarding the asymmetry of extinction ratio at bar state and cross state, theoretical analysis is first presented. The coupling ratio of the two MMI couplers deviates from 50 : 50 but the deviation is nearly the same, which is decided by good uniformity of semiconductor planar lightwave circuit (PLC) process. This property affects the extinction ratio at bar state and cross state in different mechanisms. Adding a metal electrode on the waveguide affects its effective index and analysis based on experimental results is given. The increment of waveguide effective index is in the order of 2?10^-4.

On the basis of opto-mechanical effect and micro electromechanical system (MEMS) technology, a novel substrate-free focal plane array (FPA) with the thermal isolated structure for uncooled infrared imaging is developed, even as alternate evaporated Au on SiN cantilever is used for thermal isolation. A human thermal image is obtained successfully by using the infrared imaging system composed of the FPA and optical detecting system. The experiment results show that the realization of thermal isolation structure in substrate-free FPA increases the temperature rise of the deflecting leg effectively, whereas the noise equivalent temperature difference (NETD) is about 200 mK.

Analytical delay models for Resistance Inductance Capacitance (RLC) interconnects with ramp input are presented for different situations, which include overdamped, underdamped and critical response cases. The errors of delay estimation using the analytical models proposed in this paper are less by 3% in comparison to the SPICE-computed delay. These models are meaningful for the delay analysis of actual circuits in which the input signal is ramp but not ideal step input.

As the strain sensing element of a structural health monitoring, the study and the application of the fibre-optic bragg grating (FBG) have been widely accepted. The accuracy of the FBG sensor is highly dependent on the physical and the mechanical properties of the strain interface transferring characteristics among the layers of bare optical fibre, protective coating, adhesive layer and host material. In this paper, firstly, the general expression of the multilayer interface strain transferring mechanism is derived. Secondly, based on the defined average strain, the error-modified equation of the FBG sensor is obtained. Finally, in the light of the embedded tube-packaged FBG and the fibre reinforced polymer-optical fibre bragg grating (FRP-OFBG) strain sensors, developed in the Harbin Institute of Technology (HIT), the corresponding strain transferring laws have been studied, and the corresponding error modification coefficients have also been given, which are validated by experiments. The research results provide theories for the development and application of the embedded FBG sensors.

The single-input single-output (SISO)j-stepahead predictor for generalized predictive control (GPC) controllers was traditionally derived using the polynomial approach through the Diophantine equations. An equivalent version of the predictor in a state-space form is available in the literature. In this paper, a z-domain analysis of the multiple input multiple output (MIMO) extension of the statespace predictor is carried out, and then an MIMO j-step-ahead predictor in polynomial form based on the controlled autoregressive moving average model is derived. The predictor enables us to simplify the GPC algorithm design for multivariable systems. In the SISO case the predictor is just the traditional GPC predictor, therefore this paper gives rigorous proof of the equivalence between the traditional GPC predictor and the state-space predictor.

Master-slaver dual-leg coordination control was proposed for the biped robot with heterogeneous legs (BRHL), in order to reduce gait planning and to get a good tracking performance. The key to coordination is gait trajectory tracking control. Bionic knee joint with closed-chain 4links makes robot walking more humanlike, but the model is complex and tracking of the bionic leg to the artificial leg is more difficult. P-type open-closed-loop iterative learning control (ILC) is not based on model parameters and has advantages in both open-loop and closed-loop ILC; so this paper proposes a complex robot gait trajectory tracking. The convergence is proved by using functional analysis and the stability condition is given. A tracking simulation based on the virtual prototype was done. The results show that this control strategy and algorithm are effective and robust, and the convergence speed is better than separate open-loop or closed-loop ILC.

Physical defects have always played an important role in integrated circuit (IC) yields, and the design sensitivity to these physical elements has continued to increase in today s nanometer technologies. The modeling of defect outlines that exhibit a great variety of defect shapes is usually modeled as a circle, which causes the errors of critical area estimation. Since the outlines of 70% defects approximate to elliptical shapes, a novel yield model associated with elliptical outlines of defects is presented. This model is more general than the circular defects model as the latter is only an instance of the proposed model. Comparisons of the new and circular models in the experiment show that the new model can predict yield caused by real defects more accurately than what the circular model does, which is of significance for the prediction and improvement of the yield.

Considering the salient pole and high magnetic nonlinearity of the electromagnetic doubly salient (EMDS) DC generator, a 12/8 pole prototype EMDS generator is designed and calculated using a 2-D finite element method (FEM). The phenomenon is analyzed and that the phase voltage wave changes between 120° and 180°. The influence of the exciting current and armature reaction on the DC voltage ripple of the generator is discussed in detail, and the nonlinear rules are gained that DC voltage ripple changes accordingly. The theoretical analysis is verified by the simulation and experimental results. The results are helpful for the optimal design of the generator and the optimal control of exciting-winding. We conclude that the filter-capacitance of the rectifier can be designed.

This paper presents a new rotor fault diagnosis method for induction motors which is based on the double PQ transformation. We construct the PQtransformation matrix with the positive sequence fundamental voltage components and their Hilbert transformation as elements. The active power P and the reactive power Q are obtained through the PQ transformation of the stator currents. As both P and Q are constant for a healthy motor, they are represented by a dot on the PQ plane. Whereas the P and Q for a rotor broken bar motor are represented by an ellipse because they comprise an additional frequency component 2sf_s (_s is the slip and f_s is the supply frequency). Thus, by distinguishing these two different patterns, the rotor broken bar fault is detected. We use the major radius of the ellipse as the fault indicator and the distance between the point of no-load condition and the center of the ellipse on the PQ plane as its normalization value. We thus arrive at the fault severity factor which is fairly independent of the load level and the inertia value of the induction motors. Experimental results have demonstrated that the proposed method is effective in identifying the rotor-broken-bars fault and at determining the severity of the fault.

A new method based on the finite difference time domain (FDTD) method is presented to numerically analyze the transmission and reflection characteristic of composite materials with negative effective permittivity. The numerical results are compared with the results of the existing theoretical model and the experimental data. The feasibility of analyzing the composite materials using the FDTD method is validated. It is useful for the design and application of the composite materials.