Abstract:Nowadays, the research work about data center networking (DCN) mainly focuses on the improvement of the topological properties of DCN. Unfortunately, it has been ignored whether the topologies of DCN are suitable for the cloud computing mechanisms running on DCN. To address this problem, a novel DCN topology for cloud computing was proposed. The effective method of running a mainstream cloud computing mechanism on the new topology was researched. The size and diameter of the new topology was analyzed. Simulation results demonstrate that the new topology is feasible for cloud computing.

Abstract:In many embedded real-time systems, the actual execution time of tasks is usually less than their worst-case execution time (WCET), which produces lots of dynamic slack time. Based on this fact, we proposed an energy-efficient real-time scheduling algorithm LRE-DVFS-ALL for sporadic tasks deployed on multi-cores, which is based on optimal online scheduling algorithm LRE-TL. The main idea of the algorithm is to reclaim dynamic slack time, during which DVFS techniques can be used to reduce the execution frequency of future tasks to eliminate the energy consumption. Meanwhile, our algorithm also dynamically scales the voltage and frequency at the initial time of each TL plane and the release time of a sporadic task in each TL plane to guarantee the optimal schedulability of sporadic tasks. Systematic mathematical analysis and extensive simulation results show that LRE-DVFS-ALL can always save more energy than the existing algorithms when the total workload of the system exceeds a threshold, and it can also guarantees the optimal schedulability at the same time.

Abstract:The power leakage covers more and more of the consumption of power, especially when the production of highly integrated circuit has reached the level of very deep submicron, thus it becomes the main source of the power leakage of the microprocessor. Power leakage is closely related to voltage, leakage current and the amount of transistors. Cache is the sizable fraction of the total microprocessor, and its leakage power optimization must be firstly considered in low power microprocessor design. Besides process improvement, the leakage power of caches can be adaptively reduced by monitoring and controlling its operating states at architectural level. In light of this idea, a dynamic resizing policy based on cache replacement algorithm was proposed. The cache was dynamically resized on-so-called logical way granularity according to cache operating states. Simulation results show that dynamic resizing policy can reduce cache leakage power by 76.6% without obviously performance drop, especially for high associative caches.

Abstract:As criterions and algorithms evolve and become more complex, high performance embedded application demands the high performance and energy efficiency. The challenge, however, is how to turn the VLSI capability into the actual computing performance. This research proposed an energy efficient processor architecture named ET (Embedded Tera-scale Computing), which is composed of many lightweight VLIW processor cores, also named small cores. Each core executes a thread with the mechanisms for explicitly managing the data and instructions. ET uses a hierarchical data registers to reduce the cost of delivering data, and the asymmetric and distributed instruction registers to deliver the instructions. In order to further reduce the energy, ET employs non-deep pipeline and simple control flow and optimizes the execution of loop body of applications. The primary result shows that ET can achieve the 1TOPS performance and the 100GOPS/W efficiency when scaled to 40nm.

Abstract:This paper presents a programmable shuffle unit with the efficient shuffle pattern memory for vector DSPs. The shuffle instructions can efficiently execute without occupying the system’s key resource such as the general registers or the memory bandwidth. We compress the switch-matrix by differentiating the shuffle granularity and indexing the elements. The memory efficiency of our scheme is higher than the state-of-art methods. Programmers can design the shuffle patterns ahead of time and load them into the shuffle pattern memory by the DMA or other ways. Experimental results show that our scheme can reduce the execution cycles by 7.4%～17.4% for the applications with the shuffle instruction requirement, at the cost of 0.6% additional chip area.

Abstract:Networks-on-Chip (NoC) has become a promising interconnect solution between processor cores and cache banks in many-core-processor. Routers in the network should support multicast efficiently, due to the fact that the invalidation requests in the directory-based cache coherent protocols often cause multicasting packets. Throughput and energy model of multicast communication proposed in this paper can evaluate the network ideal throughput and energy under the different multicasting routing algorithms. Network can get higher throughput under Balanced Dimension order Routing algorithm (BDoR) and Minimal Path Dimension order Routing algorithm (MPDoR) proposed in the paper than the conventional multicast routing algorithms, because they can balance the link load between X direction and Y direction. Simulation results show that the throughput under MPDoR is 1.6 times than that under the XY multicasting routing algorithm under random broadcast traffic in the 4×4 mesh network; the link will not get saturated if the output speedup of the network is less than two; and the energy consumption under MPDoR is almost the same as that under OPT alogithm.

Abstract:Recently, bufferless router, which does not need buffers, has become a low-cost solution for Network-on-Chip. To improve the performance of the bufferless router, a 1-cycle high-performance bufferless router was proposed for Network-on-Chip. The router used a simple permutation network instead of the serialized switch allocator and the crossbar to achieve high performance. Compared with the virtual channel router and the baseline bufferless router, the proposed bufferless router can achieve the frequency of 2GHz with small area cost under TSMC 65nm technology. Simulation results under both synthetic and application workloads demonstrate that the proposed bufferless router achieves much less average packet latency than the virtual channel router and other bufferless routers.

Abstract:Rolling cyclic redundancy code (Rolling CRC) uses two generator polynomials for code generation. The degree of error detection capability provided by current implementations is not strong enough to check missing of a whole packet in a series of identical ones. A modified Rolling CRC, named MR-CRC, was presented for the data error checking, which adopted higher degree polynomials, chose the proper combination and revised the schemes used in original generator and detector. The result of FPGA implementation shows that this method has low logic complexity, can improve error detection and provide advanced reliability with little performance depression. Compared with the FPGA results of 16B rolling CRC and traditional 32B CRC, the former can improve the frequency by 25% with 10% less LUTs.

Abstract:With the development of the high performance servers and very large scale super computers, the requirement for board-level high-speed data transportation bus is higher than before. How to reduce the transfer delay between chips and improve the ratio of computation to communication is very important. In light of this, the characteristics of link layer in Hyper Transport and PCI Express buses which are very popular in recent years was studied. On the basis of this, link layer architecture for a 64-bit high-speed data transportation bus was proposed and some key technologies were researched. A 16-bit scrambler/descrambler, which can scramble or descramble a 16-bit data in one cycle, was designed. A lane-to-lane deskew logic, which can correct 5 cycles delay skews at most between two lanes, was also proposed. The verification results show that the function of our designs is correct.

Abstract:For base timing-driven PathFinder router, it is found that the High-Fan-Out-Net routing spends around half of the time to insert the previous routed routing tree into routing resource node priority queue, but not all the nodes inserted are useful. In light of this finding, we proposed a pruning tree based priority queue initialization algorithm by just inserting the routing resource node which shares the same direction with next routing sink. For the High-Fan-out Net benchmark, results show our algorithm can shorten the initialization time by 5.23 times, and achieve 1.55 times speedup with almost the same quality of result.

Abstract:The correctness and reliability of parallel discrete event simulation models play an important role in the development of PDES. Most of the existing PDES developing environments do not support model verification at present. The completeness and high effectiveness of parallel model checking helps it successful in making its way into industrial tools. While PDES and PMC are implemented in different modeling languages, in the current practice, to check whether or not a PDES model contains any errors, a prototype that is solely written for model checking purposes has to be built. This process is an onerous, time-consuming and error-prone task. The current research presented a modeling language, namely Extended Event Graph (EEG) to combine PDES and PMC, and it extended the classical event graph in aspect of synchronization. This modeling language makes it possible for users to achieve PDES and PMC with only one model via transformation mechanisms, thus saving both the time and effort of developers. The experimental results confirm the validity of this language and it can support both PDES and PMC.

Abstract:Low frequency wide band SAR/GMTI has the capabilities of reconnaissance, surveillance and imaging for stationary and moving targets in the area open or concealed by foliage. A low frequency wide band SAR/GMTI system, developed by National University of Defense Technology, was introduced in detail as well as the information processing procedures. Important issues in the procedures such as spectrum correction, imaging and channel equalization were discussed. Three GMTI methods-STAP, DPCA combined with ATI and image domain STAP-were chosen in consideration of the low frequency property in the system. It is verified through the outfield experiment that the system can effectively suppress clutter, detect moving target and estimate its motion parameters.

Abstract:Accurate PSF (Point Spread Function) estimation and image deconvolution are two key point problems during motion blur image deblurring. Aimed at these key problems, a motion deblurring framework based on hybrid coded exposure camera was proposed. The hybrid coded exposure camera is made up of a high-speed stereo camera and a high-resolution coded exposure camera. The coded exposure camera is the main camera which uses a kind of binary code to control the shutter during the exposure time. As a secondary motion sensor, the stereo camera can capture three image pairs in the beginning, middle and the end time of the coded exposure camera’s exposure time. An algorithm based on the binary code and the motion parameters was proposed to estimate the accurate PSFs of the coded exposure images. The motion parameters are obtained by a motion measurement algorithm based on the stereo vision. Direct deconvolution algorithm was used for deblurring the blur images. The experiment results show that the technique aforementioned can deblur constant acceleration motion blur images effectually and quickly.

Abstract:An improved mutual coupling modified projection method was proposed, based on the two stacked patches back cavity microstrip conformal array model. The method was applied for synthesizing the patterns of this kind of conformal array successfully. The amplitude and phase of the element exciting can be solved at one time, by means of introducing the generalized element pattern including the location information. Meanwhile, the concept of the element active pattern was combined with the least square method to simplify the solved process for mutual coupling matrix. Moreover, the phase pattern was also considered, which improved the reliability of the synthesis method. Besides, the perfect agreement between the synthesized results and full-wave method confirms the validity of the proposed method.

Abstract:The inner-formation gravity field measurement satellite was proposed to measure Earth's gravity field with high accuracy and spatial resolution through precise orbit determination and relative state measurement. Both techniques relied on precise formation control and accurate attitude control, during steady-state phase, where the relative displacement had to be suppressed down to centimeter, and the outer-satellite attitude had to be earth-oriented with 0.1 degree accuracy. The integrated orbit and attitude control aims to guarantee the scientific measurement during steady-state phase, applying micro-thrusters. The all-propulsion design was processed on a coupled linear model with thruster layout; the integrated orbit and attitude control was implemented based on model predictive control (MPC) algorithm. Simulation results demonstrate the feasibility and validity of this integrated control scheme with micro-thrusters.

Abstract:It is well known that the use of aeroassisted orbital transfer can save fuel efficiently, thus representing one of development trends of new orbital transfer technology. This paper is concerned with the problem of coplanar orbital transfer of spacecraft from high earth orbit to low earth orbit based on Gauss pseudospectral method (GPM). With the consideration of heating rate limit and control constraints, sub-optimal trajectory of spacecraft’s flight in atmosphere was obtained. On the analysis of trajectory characteristic the first-order necessary optimality conditions of optimized results was demonstrated. Indirect method was also adopted to solve the problem, then the calculation results of the two methods were compared. Simulation results show that it is a simple way to handle constraints with GPM, and it has high precision solutions and can overcome the disadvantages of the indirect method, such as initial value sensitivity and poor robustness. Therefore, GPM is suitable to be applied in the problem of atmospheric spacecraft trajectory optimization.

Abstract:The thermal dynamic response of structures irradiated by the intensive laser beam has been an important issue in the area of laser application. Adopting three-dimensional finite element model, numerical simulation and analysis were performed in order to make out the temperature, stress and displacement fields of two-layered cylindrical shell irradiated by intensive laser beam. In the study, the several parameters of materials corresponding to temperature were considered as variables. The influence on distribution of temperature field brought by various laser spot diameters was also studied. And the rules of temperature rise, stress and displacement of inner layer at the spot center were investigated, while the outer layer’s thickness, expansion factor, the elastic modulus and the Poisson ratio were changing. It is shown that, the diameter of laser spot plays an important role in the temperature field distribution, keeping in touch with conductivities of outer layer; expansion factor is the dominant parameter affecting stress field, thickness is a lesser dominant one, and elastic modulus and Poisson ratio are the least. The conclusion can provide a substantial reference for the laser breakage analysis and anti-laser reinforcing technology.

Abstract:A far-field Direction-of-Arrival (DOA) estimation algorithm with attitude correction based on a single vector hydrophone was proposed. In practical application, both the direction finding and the signal-to-noise ratios (SNR) of three components in the acoustic particle velocity vector suffered from the attitude deflection of a vector hydrophone. Two components of a particle velocity vector combined with pressure signal could be used to estimate the two-dimensional bearing of a far-field target. Therefore, a far-field DOA estimation algorithm based on two properly chosen velocity components with attitude correction was proposed to increase the DOA estimation accuracies. Both the experiment data and numerical simulations were used to verify the feasibility and validity of the method.

Abstract:Aimed at the problem of sea-sky-line detection in complicated sea-sky background image, a novel method, based on phase grouping and gray statistics, was presented. Firstly, a lot of line segments were obtained by using phase grouping method on edge image. Secondly, the line segments were classified according to two parameters, the slope angle and the distance to image origin. Several classes which have the largest sum of length of the line segments were selected, and edge points in each selected class were used to fit a straight line, thus several candidate sea-sky-lines could be gained. Finally, the correct sea-sky-line could be elected through comparing statistical gray values of the regions which correspond to the same horizontal pixel coordinates and take the points on the candidate sea-sky-lines as center points. Many experimental results show that the proposed method can detect the sea-sky-line under complicated sea-sky background effectively and has strong robustness.

Abstract:Chemiluminescence is an important method of measuring parameter in optical cavity flow field for HF/DF chemical laser. In the process of increasing H₂ flow rate, the changing of the chemiluminescence in visible range in the optical cavity flow field of discharge driven HF laser was observed and recorded, and the spectrum of the chemiluminescence on the optic axis was measured. The component particle of the gas in the optical cavity was analyzed theoretically, and the high order overtone spectral line of HF molecule was computed. It was found that the spectral line of the transition from 3 order vibration excited state to ground state was in infrared region. The measured spectrum of the chemiluminescence and the spectral line of the particle was compared, discovering that the yellow-green chemiluminescence was formed by the radiation of the excited F, He, H, N, HF and other particle, which corrected the explanation in some classic literature that the yellow-green chemiluminescence was formed by the 3,4,5 order overtone radiation of HF molecule. This further explains the color forming of the chemiluminescence on the optic axis.

Abstract:The modified smoothed particle hydrodynamics method (MSPH) has obvious advantages on simulating the radiation effects of resin composite irradiated by laser, but the serial computation is difficult to meet the demand. In order to improve the computational efficiency of the serial MSPH program, the parallelization of the serial program was carried out by using the message passing interface (MPI) standard. The parallel computing results were compared with the serial computing results given by a personal computer, and the performances of the parallel program involving speed-up and efficiency are tested on a 5 teraflop computer cluster. Numerical calculations show that the parallel computing results are consistent with the serial computing results, thus obtaining a very good parallel performance. Therefore, the parallelization based on MPI is successful.

Abstract:Based on the flow characteristics of ramjet combustion chamber, combustion characters match of the flameholder and injectors is investigated by means of photography of using OH spontaneous radiation，and the two-phase flow numerical simulation. The influence of distance between the flameholder and the injectors on the fuel-air equivalence ratio distribution behind the flameholder is analyzed. The experiment results indicate that the space between the flameholder and injectors has signiticant influence on the reaction flow. When the space between the flameholder and injectors increased, the flame was closed to flameholder. And the flame flashed back when the velocity of air was low. when the space between the flameholder and injectors shortened, the flame flash back could avoided and the lift-off distance of the flame became small. The flame located close to the flameholder was due to improved vaporization and mixing at a big space between the flameholder and injectors.

Abstract:Three-dimensional carbon fiber reinforced SiC-ZrC composites (C/SiC-ZrC) were prepared by chemical vapor infiltration (CVI) combined with precursor infiltration and pyrolysis (PIP) process for ultra high temperature application, and the properties of C/SiC-ZrC composites with different SiC contents were investigated. The results show that with the of increased the SiC content, the flexural strength of the composites decreased gradually, the oxidation properties increased, and the mass loss rate and linear recession rate in oxyacetylene torch environment decreased first and then increased. The changes of the oxidation and ablation properties were caused by the formation of SiO₂ glass.

Abstract:Gain-scheduled robust control design method based on LMI was adopted to resolve the control problem of speed-dependent model varying for AMB-rotor system with strong gyroscopic effect. Speed-dependent LPV model was derived, and gain-scheduled robust controller was designed which can guarantee the robust stability and performance in the whole speed range. As for reducing the design conservation, controller design should have the speed range limit. Compared with the robust controller based on LTI model, LPV gain-scheduled robust controller guarantees the performances and realizes the adaptive parameter adjustment for the varying speed. The simulation validates the effectiveness and advancement of the designed controller.

Abstract:A new uncoupled algorithm of simulating chemical non-equilibrium flow originally realized by finite difference method was extended to the unstructured finite volume method, which makes the simulation of reaction flow in complex configurations possible. The H₂/Air shock-induced oscillating combustion experiment conducted by Lehr was simulated, and the computed oscillating frequencies conformed well with the ones acquired by experiment, which indicates that the present method is temporally and spatially second order accurate. It is also demonstrated that geometrical configuration plays a definitive role in the combustion regime. If the truncated angle is less or equal to 15 degrees, high frequency oscillating combustion in regular regime occurs, while if the angle is larger or equal to 20 degrees, low frequency oscillating combustion in large-disturbance regime is observed.

Abstract:A satellite antenna positioning mechanism based on spherical gear (PMSG) was presented, and its kinematics character was also researched. The kinematics model of PMSG itself and a satellite were established separately, and the PMSG kinematics model was established based on the kinematics models. Considering the characters of intersatellite communication, the kinematics characters of two middle orbit satellites were simulated and researched for example. Results of the simulation verified the theory model established above, and the direction rule and motors movement rule were acquired, which were the theory basis in intersatellite communication. Conclusions acquired can be the basis of the engineering applies of PMSG, and it can be beneficial to the other correlative applications.

Abstract:A method of particle coding based on progression was applied to the simplification of constraints in Weapon-Target Assignment (WTA) model. The original EM algorithm was modified in all steps, including the initialization of the population, the local searching, the calculation of total force vector and movement along the total force, and the WTA problems with integer solution space was settled by the algorithm. Finally, the numerical experiments verified the effectiveness of the algorithm for solving WTA problems.

Abstract:In the field of M&S of Component, the component modeling and simulation environment KD-SmartSim based-on BOM specification has been used widely. The application composed of components in KD-SmartSim employs RTI to achieve parallelism, but its performance is inefficient. In this study, the deficiencies of KD-SmartSim were discussed, and a new parallelization structure for component which inherits the concept of BOM component modeling was presented. A parallel simulation engine for components in the form of SPMD was designed, which implements various services required by model computing. In the prototype system, the sequence and conservative time advancing were completed now. The experiment shows that the engine has all capacities which can support BOM component executing and can improve the performance of simulation system composed of components greatly.

Abstract:The magnetic flux density in the clearance and the structural parameters were designed for the U-shaped electromagnet of Low-speed EMS Maglev Trains. The magnetic flux density in the clearance was chosen as the magnetic flux density when the permeability of the iron core material decreased to ninety percent of its maximum value. The related structural parameters are the length and the width of the magnetic pole, the width and the height of the window, and the depth of the yoke. It was represented that the length of the magnetic pole has effects on the turning capability of the maglev train, the width of the magnetic pole has effects on the carrying capacity and the guiding stiffness of the train, the width and height of the window has influence on the weight of the train, which should be determined by the width of the guide way, and the depth of the yoke should be chosen according to the principle that the magnetic flux density in the magnetic yoke equals to that in the magnetic pole.

Abstract:Poor reusability, semantic inconsistency and difficult implementation really exist in weapon system of systems capability (WSoSC) modeling due to capability concept and complicated relationships. To solve this problem, an ontology-based capability views construction method and six capability meta-models were proposed. The first step is to determine capability domain problems and specify core ontology of WSoSC domain and its building guidance. Next is to describe ontology in formal way and construct their relationships. The final step is to develop meta-model and create views production using relative tools. Armored equipment system-of-systems capability views were built as an illustration based on previous approach.

Abstract:Some effective modifications upon NSGA-II were proposed to further improve its optimization ability, which results in a hybrid multi-objective optimization algorithm. In the hybrid algorithm, Taguchi-method was incorporated into the crossover and mutation options of NSGA-II, whose effectiveness was approved by experiments on typical test functions, and the hybrid algorithm can be easily implemented since it makes no change on the framework of NSGA-II.

Abstract:The quantification of human error probability in CREAM was improved based on available meterials. First, the basic method of CREAM was introduced. Second, two methods to determine probabilistic control modes were discussed, which are Bayesian nets and fuzzy logic. The necessity to quantify human error probability in probabilistic control modes was emphasized. Last, the method was proposed to quantify the human error probability in probabilistic control modes by theoretical inference. In order to improve the computing effectiveness, the Monte Carlo algorithm was provided. An example was presented and the validity of the proposed method was proved.