Abstract:Many methods optimizing the on-chip cache utilization are dependent on the profile or identification of the access sequence characteristics, for instance, prefetcher or bypass etc. How to identify these characteristics is still an open problem. Through a detailed theoretical analysis of typical access patterns, it is shown that the frequency of stack distance has obvious features. Furthermore, according to the results of the Simics simulation, these features present persistent and stable to a certain extent, therefore, they are feasible to be identified and predicted. An online method based on the peak value of the collected stack histogram attaching to each core is provided. In addition, the storage and time overhead is small. The experimental results based on 15 benchmarks from SPEC CPU2000/2006 show that it identifies all correctly.

Abstract:Building storage array based on the low latency, low power consumption, and high reliability NAND Flash is an efficient way to implement high-performance storage system. However, adopting traditional redundant arrays of inexpensive disks (RAID) techniques to build storage array based on NAND Flash bring several problems such as wear leveling and decreasing lifecycle of storage array caused by updating parity code frequently. A cache-based reconfigurable RAID mechanism that dynamically constructs a new data stripe based on the non-volatile SCM is proposed. Experimental results show that this mechanism can reduce the garbage collection overhead, improve the performance and lifecycle of storage array based on NAND Flash efficiently.

Abstract:To reduce remote cache transfer latency and improve the performance of shared memory systems, a new last-write-touch prediction scheme that exploits the inherent write characteristics of a program is proposed and a directory protocol to support the scheme is adapted. By predicting a last-write-touch and self downgrading a cache block in advance, a processor can get the data from the memory directly and one network hop can be saved for a remote cache access. Compared with the existing instruction-based prediction technique, much storage overhead can be reduced. Experimental results show that it can achieve an average prediction accuracy of 83.1%, leading to improvements up to average 8.57% on the final application performance. Moreover, compared with the instruction-based prediction scheme, the scheme can reduce the storage overheads of the history table by 69% and the storage overheads of the signature table by 36%.

Abstract:Computing unsatisfiable subformulas of Boolean formulas has practical applications in VLSI design and verification. The unsatisfiable subformulas can help electronic design automation tools to rapidly locate the errors and inconsistency. The definitions of resolution refutation and refutation parsing tree, and a heuristic local search algorithm to extract unsatisfiable subformulas from the resolution refutation of a formula are presented. The approach directly constructs the refutation parsing tree for proving unsatisfiability with a local search procedure, and then recursively derives unsatisfiable subformulas. The algorithm combines with reasoning heuristics, dynamic pruning and subsumption elimination method to improve the efficiency. The experimental results show that our algorithm outperforms the similar algorithms on the random benchmarks.

Abstract:In order to access independent memory space of CPU and GPU directly from opposite directions, an effective approach to accessing unified memory address space of heterogeneous kilo-cores system is proposed, which is implemented by building a unified 3-level Cache and tagging blocks in Cache, and optimizing the algorithms of modifying the states of blocks. Therefore, the heterogeneous kilo-cores system avoids significant overhead of accessing memory instead of that in current discrete hybrid computer system equipped with GPUs by PCI-E. According to the results of experiments from partial programs of Rodinia benchmarks, a maximal speedup by 9.8x and maximal decrease of load/store instructions by 90% are gained. In conclusion, it’s certified that our solution is effective to decrease overhead of transferring data among computing units in heterogeneous system and significantly enhance the whole system computing performance. 

Abstract:Thin-film gated silicon-on-insulator (SOI) lateral PIN photodetector is a novel photodetector, based on standard SOI technology and CMOS process, combining the conventional bipolar and field effect photosensitive device structure. The basic structure of this novel photodetector is described and the operation principle is analyzed. Applying physical equations of semiconductor device, the gate voltage and photocurrent models can be built. Two-dimensional numerical simulations are performed in SILVACO software. In the middle-short wavelength operation period, the output photocurrent increases with gate voltage, with the obvious gated-control characteristics. Under fully depleted condition, the internal quantum efficiency can yield over 96%, even near 100% for the varied wavelengths(400nm, 450nm, 530nm, 600nm). For short wavelengths(280nm, 350nm), the internal quantum efficiency is relatively lower, the maximum is approximately 80%. And the dark current of this photodetector is low, leading to a high ratio of more than 10⁶ between illuminated to dark current, achieving high sensitivity.

Abstract:Comprehensive signal integrity simulation was introduced in the design of PCI-E3.0 and SAS3.0 channels of high density storage server, which are characterized by complicated structures, high signal speed and long and dense traces. The optimized cost-effective design was obtained by simulating the key items of high-speed PCB design such as topology, material and high-speed via structure. The system performances were then evaluated by full channel active simulation to ensure the stability of the system and reduce the manufacture risks. The signal quality of storage server system totally meets the requirement of PCI-E3.0 and SAS3.0 specifications, verifying the effectiveness and reliability of signal integrity simulation.

Abstract:Today’s advancing modern science is generating and analyzing increasing scale of datasets and makes HPC storage system facing new challenges both on architecture and software approach. In order to exploit potential benefits of emerging hybrid and hierarchy storage architecture on Milky-2 system, a I/O middleware approach named application coupled burst buffer, is introduced to make full use of the solid state disk based in-system storage resources. Application coupled burst buffer aggregates distributed in-system storage close to running tasks into single namespace during application runtime and manages it as cooperative persistent burst buffer tightly coupled with its host application. To take full advantage of cooperative burst buffer, application coupled burst buffer uses an unified shadow namespace to map application data into physical in system storage based on its real namespace of the host application. Besides that, application coupled burst buffer organizes data with locality aware layout and leverages application intent based replacement policy to fully exploit spatial and temporal locality. Furthermore, application coupled burst buffer employs concurrency aware policies to optimize data movement between different storage tiers. Evaluations on Milky-2 system show that application coupled burst buffer can improve the performance of typical data-intensive applications dramatically. It can achieve scalable burst I/O bandwidth and smooth sustained I/O bandwidth with high throughput solid state disk deployed and can be taken as an appropriate candidate for storage solution on emerging leadership supercomputer systems.

Abstract:Aiming at requirement of high speed and complete of data in space storage system, the design of a high performance NAND Flash controller is present. It concludes a pipelining-programming inside of NAND Flash chip and a non-missing invalid block method. The storage implementation is present. The calculation of storage time in different situation is discussed. The simulation modules are present and the impact of pipelining programming is simulated and discussed using Monte Carlo method. Practical application proves the pipelining programming and non missing invalid block method. The operation frequency of storage system achieves to 100MB/s, ensuring accuracy, completeness and continuity of data.

Abstract:Zero-propellant maneuver is an advanced concept of attitude control, which has been applied to the International Space Station. The dynamic models of space station using control momentum gyros are constructed, and the momentum equilibrium relation between space station and control momentum gyros is established. Then, considering different large angle attitude mission, the relation between inertia of space station and characteristic parameters of control momentum gyros is analyzed, and conditions of zero-propellant maneuver path existence are obtained. Effectiveness of the existence conditions is verified by the path planning examples. The proposed existence conditions can provide a convenient and effective method to determinate whether the zero-propellant maneuver path exists, which can provide significant theoretical argument for zero-propellant maneuver technology to apply to Chinese Space Station in future.

Abstract:The technology advantages and application prospects of satellite-based automatic identification system (AIS) were introduced, and then the system of the satellite based AIS of TT-1 was designed to overcome the problems of multi-cell signal collision, Doppler shift and large space link attenuation, originating from the development of ground-based AIS to satellite-based AIS. This design included the composition of system, the qualification and function of AIS receiver, the design of antenna and microwave switch and the link level estimation of satellite-ground AIS. For the preparation of space applications, it mainly discussed the performance and environment test for the AIS receiver.The on-orbit application situation of TT-1 satellite based AIS was introduced and the results show that the innovated AIS receiver can partly ease these technical problems and achieve a wide range of maritime surveillance requirements.

Abstract:An improved uncoupled solver of non equilibrium flow was adopted to resolve the Euler reacting equations and a chemical H₂/air mixture reaction model with 9 species and 21 equations was employed to conduct the two-dimensional numerical simulation in detonation onset process of vertical hot jets. The convection term was dispersed with fifth-order WENO format and the time integration was iterated with second-order Runge-Kutta scheme. The flow fields in the detonation tube were analyzed in detail when hot jets were injected with different parameters (i.e.incident velocities, incident positions, incident widths and incident angles). And the general rule of the impact of hot jets parameters to the detonation initiation was summarized. Moreover, interaction between shocks and flame during the formation and propagation of “hot spots” was explained, especially the promotion of reflected shocks on the formation of “hot spots”. The investigation reveals that detonation always occurs from the “hot spots” in narrow unburned region which is confined by adjacent flame front and solid wall. Triple point structures emerge during the over-driven transition of “hot spots” to detonation. In order to achieve rapid detonation onset, the hot jets should be injected towards the side wall in high velocity and with a certain angle.

Abstract:Dynamic soaring is a special flying technique designed to allow air vehicles to extract energy from atmosphere. Based on the established dynamic model of the aircraft, the energy transformation of unmanned aerial vehicles (UAVs) in a gliding cycle of the known gradient wind field was analyzed. The method of subsection analysis was introduced and a dynamic soaring cycle was divided into four phases: climbing with headwind, turning in high altitude, gliding with tailwind, turning in low altitude. Turning in high altitude was the most critical phase in the dynamic soaring cycle. By using the calculation method which combines the three-dimensional space path with the two-dimensional plane projection, the motion equation and energy transformation equation of UAV in the phase of turning in high altitude were achieved. Based on the equations, the parameters which influenced the energy extract of UAVs from gradient wind field and which resulted in the energy loss because of the wind drag were analyzed. The conclusion provides theoretically guide for the UAV’s maximum extracted energy from gradient wind field and the decreasing of its energy loss. Besides, the dynamic soaring simulation was designed and results show that the phase of climbing with headwind and the early stage of turning in high altitude are the mainly periods to absorb energy from wind field. This conclusion is meaningful to direct the dynamic soaring without thrust.

Abstract:Based on arbitrary Lagrange-Euler fluid-structure interaction method, the inflation process of a typical kind of vortex ring parachute in an infinite mass situation was simulated. The inflation process and the time-history changing curves of spinning rate and opening load were obtained, and the fluid-structure interaction characteristics at the steady stage, such as the change law of flow field around the canopy and the structural strength of the canopy fabrics, were analyzed. Results show that the stable spinning rate is about 3.1r/s and the canopy inflated shape is plump with the flow velocity of 12m/s, which are in good agreement with the tower test data. In steady state, there are a plenty of vortex cores above the parachute, and the ligature of the cores is similar to a spiral line. The drag coefficient of vortex ring parachute is greater than that of other typical parachutes with axial symmetry structure. The stress of the connected region of the canopy and the suspension lines and the canopy fringing field are obviously greater than the average stress of the canopy.

Abstract:The pseudo range and the precision of pseudo range variability from ground-based pseudolite navigation system are very important to the navigation performance of evaluation system and the fusion of pseudo range information with the navigation information from other system. At present, the evaluation methods of pseudo range precision are basically aimed for satellite navigation system and they don't take full account of the characteristics of ground-based pseudolite navigation system. By constructing the precision observations of pseudo range and its variability and utilizing some bad positioning configurations produced by ground-based pseudolite navigation system, a precision evaluation method for pseudo range and its variability which can get accurate precision estimator even the exterior ballistic measuring system has low position-velocity precision and owns arm length was proposed. The simulation results indicate that the method can evaluate the precision of pseudo range and its variability effectively and has a strong applicability. 

Abstract:Using solid-propellant nozzles is one of the best schemes for kinetic interceptor to realize the fast response and high precision of attitude control. A mixed searching algorithm for ignition combination was presented for a novel attitude control solid-propellant nozzle in kinetic interceptor. Firstly, the configuration of solid-propellant nozzles was described and spin requirements of the kinetic interceptor were analyzed. Then the mixed searching algorithm was designed by a combination of sorting method and interval searching method. Sorting method is selected when the number of available nozzles is small and interval searching method is chosen on the contrary. Results of instruction torque approximation simulation and attitude control numerical simulation suggest that the algorithm can track the instruction torque effectively and achieve attitude tracking quickly and with a high precision.

Abstract:Energy balance is the basic requirement in designing and evaluating the power system of nano-satellite, and it is often examined by numerical simulation on ground. The architecture and feature of the board nano-satellite power system was introduced first. Then the mathematical models of solar array, Li-ion battery and satellite load were developed respectively. At last, the energy balance was simulated by combining the satellite attitude dynamics. The simulation result can reflect the power change, battery capacity and state of energy balance in real time. The data of on-orbit experiment is consistent with the simulation results, which show that the models are correct and the simulation method is feasible.

Abstract:In split Hopkinson tensile bar test, it is necessary to obtain the smallest strain rate of the specimen which was fractured by a tensile stress impulse, while determining the corresponding launch pressure of the smallest strain rate needs to conduct many tentative tests. In order to solve this problem well, an empirical method which needs to carry out tensile fracture test only once was proposed to calculate the smallest launch pressure and strain rate to fracture the specimen. The method can further predict the strain rate in different launch pressure. The tensile fracture tests of four kinds of aluminum alloy testify that the proposed method is coincident with the experimental results and it is efficient to evaluate the launch pressure and strain rate.

Abstract:The propagation problem of two-dimensional elastic wave in macroscopical inhomogeneous medium was analyzed. As the physical properties of medium are variable with space, the elastic wave propagation in this medium should be described by a set of inhomogeneous wave equations which conclude the space derivatives of the medium properties. By solving the inhomogeneous wave equations with finite element method, the elastic wave propagation characteristics in inhomogeneous medium whose physical properties vary with vertical gradients were studied and discussed. Results show that the planar elastic wave converges or diverges respectively when the wave velocity ascends or descends symmetrically in the vertical direction. Moreover, when the profile of the wave velocity is arc-hyperbolic secant, the planar elastic wave focuses periodically on the symmetric axis.

Abstract:Aiming at the range-spread target detection problem in KK-distributed heavy-tailed radar clutter, the KK-distributed radar clutter was taken as a spherically invariant random vector. The Neyman-Pearson optimal integrator for the range-spread target detection with known target amplitude was derived firstly. Then by replacing the ideal target amplitude with the maximum likelihood estimates, the detector model in generalized likelihood ratio test (GLRT) was obtained. Both of the detectors are dependent on the modified Bessel function of the second kind, which makes the detectors computationally complicated, so a suboptimal generalized likelihood ratio detector based on order statistics (OS-GLRT) was proposed. The OS-GLRT takes some range unit echoes with largest amplitude in detection window as target echoes. The performance assessment conducted by Monte Carlo simulation validates that: the optimal integrator and GLRT have better performance, however, they are hard to applied, and the OS-GLRT is a more practical detector.

Abstract:Mushroom electromagnetic bandgap structure shows surface-wave bandgap and in-phase reflection, these two characteristics were simultaneously used for designing the waveguide slot array antenna. The surface-wave bandgap was used to suppress the surface-wave in antennas and arrays, which results in radiation performance improvement. In addition, the in-phase reflection could be used to reduce the RCS level of antennas and arrays, so as to improve their total performances. A sample of four-element EBG waveguide slot array antenna was made, and the indexes of mutual coupling, radiation pattern and RCS of the sample were tested. The result is similar with the theoretical expectation, which has effectively reduced the mutual coupling and the RCS level of antennas and arrays.

Abstract:According to array radar system, low-angle tracking problem was investigated from the perspective of revised mono-pulse. Firstly, based on the array signal model under multi-path conditions, the double-null mono-pulse algorithm was deduced by using the principle of Maximal Likelihood (ML) estimation. The revised sum-difference patterns were designed by digital beamforming (DBF). Simulation result shows the relation of double-null mono-pulse performance, SNR, target elevation, and complex reflection coefficient. Finally, measured data is used for validation. Result shows that the root of mean square error (RMSE) is still less than 1/40 of the beamwidth when the target elevation is less than 1/4 of the beamwidth. Therefore, the double-null mono-pulse algorithm is effective for low-angle tracking problem.

Abstract:Aimed at the problem that current deployment decision-making algorithms for software defined radio waveforms with throughput and latency constraints were paid little attention to the memory overhead, a more general system model for waveform deployment decision-making was proposed, by which the processing, memory and communication bandwidth were all incorporated in the process of waveform deployment decision-making. Besides, a dynamic programming-based algorithm for waveform deployment decision-making was devised to minimize the sum of various kinds of computing resource occupancies of waveform. Simulation results show that the memory has an important effect on the waveform deployment decision-making, while comparing with the algorithm that does not consider the memory overhead, the average maximum waveform number and the average processing resource usage supported by platform are both decreased about 40%.

Abstract:In order to overcome the disadvantages of the traditional LP (Linear Prediction) algorithm that brought the cumulation of prediction errors in the process of building a virtual array, a kind of improved LP algorithm was proposed, with which we could make full use of the information of the real array in the virtual array building. Respectively through the actual array method, the traditional LP virtual array method and the improved LP virtual array method, the beam performance of assumed array was simulated and the data from lake trial was dealt with to form the imaging of echo highlights. Result shows that the improved LP method is more effective than the traditional LP method in suppressing the side-lobe of wave beams, and further improves the angle resolution and array gain. And when the improved LP method is applied to practical engineering for bottom target detection under specific experimental background, the directional resolution accuracy can be increased by 75%.

Abstract:Emitter identification is brought into research hotspot in the field of electronic warfare. To solve the problem of radar emitter identification in time-variation channel, a local correlation testing based on Likelihood Test was proposed. The method used the radar waveforms with large diversity parts to form reference waveform, identified signals by building a hypothesis model, and adjusted the variability of reference waveforms by a scale coefficient. Simulation result shows that the recognition performance and the multipath interference performance of the method are better than likelihood-ratio test with the incremental variability of the reference signal.

Abstract:Multimodal image registration between visible, infrared and SAR images is the key technology to all-weather high-precision navigation and guidance. Based on an invariant feature designed at the pixel level, the orientation-moment, which describes the similarity of a pixel with its neighbors in different orientation, a multimodal image registration algorithm, that matches multisensor images by analyzing such similarities, is proposed. In matching tests, including visible, infrared and SAR images, it has reached a success rate of more than 90%. Comparing with traditional multimodal image registration algorithms, this method greatly improved the success rate, and had broad application prospects in navigation and guidance.

Abstract:Aimed at the problem of position and attitude estimation for UAV, a relative position and attitude estimation algorithm based on target features in image was proposed. The initial location of the target was obtained with Camshift algorithm. The feature points in tracking area based on the nonlinear scale space were picked up with KAZE features, which were used to match with the feature points of the source target. The exact location of the target could be obtained and the target could be tracked quickly in the picture plane. The estimation of relative position and attitude between the unmanned aerial vehicle and target was conducted in the body frame of axes. Some experiments were fulfilled for the verification of the algorithm. Results show that the proposed algorism has strong tracking and real-time performance. 

Abstract:Parabola modification method of double helical gears was proposed, tooth contact analysis of modified gears including typical install errors was established based on tooth contact analysis of helical gears. Considering the transmission errors and meshing performance, transmission behavior of double helical gears was optimized from using the fast non-dominated sorting genetic algorithm to obtain the best parabola modification coefficient which could also avoid the local convergence. Examples show that path of contact and transmission errors of optimized gears are preferable distinctly. It has an important significance in improving distribution of load of tooth surface as well as reducing gear meshing impact. In the end, the noise value of original gears and optimization ones are obtained by the loaded test, further verifies the optimal method has certain effect on the noise reduction of gears.

Abstract:Combat system is an aggregation of various relatively independent subsystems, organizing in a large-scale hierarchical structure. The structure with multilevel, multisystem, multi decision-makers and multi-objectives complicated the allocation process of weapons and equipment to each subsystem when facing an attack. For optimizing the complicated structure, a model with multi-objectives and multi decision-makers based on risk management was built, which could deal with the hierarchical couplings between subsystems and conflicts between multi-objectives, especially in uncertain situations. In order to solve the air defense and anti-missile weapons allocation problem of the MOMDRA model, the idea of system decomposition and weighting method are used to generate Pareto optimal allocation solutions from the aspect of different decision makers’ concerns. In the end, the feasibility of this method is illustrated by an example.

Abstract:To solve the problem of air defense system modeling and assessing which is partial to determine the nature, only paying attention to system structure and overlooking entity and relation property in system, some efforts have to be done. On the basis of recent model and assess method, put forward air defense system model based on adjacency matrix, define capacity assess vector, use the target value evaluation method that was based on target contribution degree to system capacity and put forward target rank and discrete degree based on Pareto dominance and crowding distance so as to form Non-inferiority rank algorithm, and use the weighting and cost performance and other network access methods as the comparison, establish random examples for simulation, simulation result proves that the air defense system model based on Markov logic and Non-inferiority value rank algorithm are efficiency.