Bufferbloat occurs when excessive packets that exceed the network capacity are blocked in the buffer, which increases the network transmission latency. Cellular networks generally use a large buffer for network reliability, which further aggravates this problem. This paper gives a receive window adjustment scheme based on radio channel state information to mitigate bufferbloat in cellular networks. The scheme retrieves the available bandwidth at the receiver side from the radio channel state information. The effects of channel noise and the ACK self-clocking mechanism in the transmission control protocol are modeled by a Kalman filter to predict the available bandwidth. Then the send window size is adjusted according to the calculated receive window. The scheme performance was evaluated on a network simulation platform with the results indicating that this scheme significantly reduces the queue backlog and the latency while still maintaining the original high throughput than previous mechanisms

Analyses of big data sets often require aggregate queries on uncertain data with various types of data that are computationally complex. In this paper, the results of aggregate queries on uncertain data are defined to include all possible values and their corresponding probabilities. Dynamic programming is then used to solve the Distribution Sum (DSUM) algorithm using a Greedy-based Distribution Sum and a Binary Merge based Distribution Sum approximation algorithms, which both can be applied to tuple-level and attribute-level uncertainty models. The time and space complexities of the algorithms are determined theoretically as well as the error range of the results. Tests demonstrates that these two approximation algorithms with a 1% allowable error shorten the execution times by 15%-21% and 22%-32%, respectively.

Existing detection approaches of return oriented programming (ROP) attacks cannnot simultaneously provide flexible deployment, allow portability, and allow transparent detection in the cloud environment. A hardware-assisted method was developed to detect ROP attacks in real time using the hardware features of the Intel last branch record (LBR) to record indirect branch information of a guest virtual machine (VM) to achieve rapid detection of gadget attack chains in the hypervisor. In the privileged domain, the method takes advantage of the virtual machine introspection (VMI) technology to validate the legitimacy of indirect branches to guarantee the control flow integrity of the shared link library in the process address space of the guest VM. Tests show that this demonstrate method can detect ROP attacks with an average run-time overhead of less than 7%.

This paper describes a class of rate-compatible nonbinary low-density parity-check (LDPC) codes based on an improved extension method. The check symbols for lower code rates involve not only the codewords of the highest-rate code but also the check symbols of moderate-rate codes. The construction process optimizes the masking matrix and the base matrix with random replacement of nonbinary elements. This paper also describes an algebraic method to design the parity matrices to further reduce the design complexity. The codes are not only rate compatible, but also have a quasi-cyclic structure which will benefit hardware implementations of the encoder and decoder. Numerical tests show that the codes can achieve good performance within a wide range of code rates in both the waterfall region and in the error-floor region.

Distant speech recognition remains a challenging task owning to background noise, reverberation, and competing acoustic sources. This work describes a long short-term memory (LSTM) based acoustic model with an attention mechanism and a multitask learning architecture for distant speech recognition. The attention mechanism is embedded in the acoustic model to automatically tune its attention to the spliced context input which significantly improves the ability to model distant speech. A multitask learning architecture, which is trained to predict the acoustic model states and the clean features, is used to further improve the robustness. Evaluations of the model on the AMI meeting corpus show that the model reduces word error rate (WER) by 1.5% over the baseline model.

The language recognition (LR) accuracy is often significantly reduced when the test utterance duration is as short as 10 s or less. This paper describes a method to extend the utterance length using time-scale modification (TSM) which changes the speech rate without changing the spectral information. The algorithm first converts an utterance to several time-stretched or time-compressed versions using TSM. These modified versions with different speech rates are concatenated together with the original one to form a long-duration signal, which is subsequently fed into the LR system. Tests demonstrate that this duration modification method dramatically improves the performance for short utterances.

Current manual segmented microblog-oriented corpora are inadequate, so both conventional Chinese word segmentation (CWS) systems and deep learning based CWS systems are still not very effective. This paper presents an active learning method that selects samples with high annotation values from unlabelled tweets for microblog-oriented CWS. A parameter is introduced to control the number of repeatedly selected samples that offen occur in microblog data. Three strategies (Max, Avg and AvgMax) are used to evaluate the overall values of each sample. The initial segment character is a stop character which is calculated by taking character embeddings into consideration. Tests demonstrate that this method outperforms the baseline system with F Gains of 0.84%~1.49% and state-of-the-art active learning method word boundary annotation (WBA).

PDF files are portable and widely used, so they often host malware. Traditional PDF viewers fuzzing algorithms cannot work well due to their strict format validation. Also, existing file-format based grey-box fuzzing cannot be easily used to build a uniform data model because of the limits of its descrition language. This paper presents a method for generating test cases to test the font parser of PDF viewers. The system reconstructs the font files and adds supportive information to build a uniform data model for TrueType files. A fuzzer is built into the method and evaluated on more than twenty PDF viewers to identify several vulnerabilies. Tests show that this method can effectively generate test cases and detect bugs in PDF viewers.

A causal relation support analysis method based on a causal network is presented here to identify the causal relation types in Chinese reading comprehension. Firstly, the causal events are extracted from the literature by clue phrases, the causal relation between the events is given a value, and a causal network is constructed from the causal events and the causal relation. Then, the TF-IDF (term frequency-inverse document frequency) method is used to retrieve related sentences from the document and the importance of each word in the document to characterize the whole document. Finally, the causality network and related sentences are combined to analyze the causal support of the option. The method was evaluated using 769 articles and 13 Beijing colleges entrance examination (including the source text and the selected title) as a test set. This method then gave about 11% better result than the Baseline method.

With the increasing numbers of electric vehicles, electric vehicle charging can result in traffic congestion near the charging stations during peak charging times with long wait times, power grid load fluctuations and other issues. Real-time information communication systems can be used in a charging path planning method for electric vehicles based on traffic network, charging stations and distribution network information. The system constraints include the traffic conditions, vehicle charging requirements, charging station loads, and grid loads. The multi-objective optimization function includes the travel time, number of vehicles in the charging stations, and charging loads with an improved Dijkstra method used to solve the optimization problem. Simulations show that this method improves the distribution operations, relieves traffic congestion around the charging stations, reduces the wait times and improves the usability of the charging infrastructure.

A radar sharing based energy-saving control method was developed for intelligent hybrid electric vehicles to reduce energy loss caused by designs for fixed working conditions that disregard the actual running environment. Four scenarios were developed base on the relative motion of the vehicle in front of this vehicle. The motor drive torque was optimized with regenerative braking by the motor added for some scenarios. These control strategies reduce the hybrid electric vehicle energy use without adding extra hardware cost. Tests on different road conditions show that this intelligent energy-saving control strategy provides large energy savings especially on congested urban roads.

Versatile welding robots designed for the field welding of box-type steel structures have the problems of low DOF (degree of freedom) and DOF coupling. A numerical solution method was developed for the inverse kinematics of the robot based on the layouts of actual welding processes to solve difficulties encountered in analytic methods. The numerical solution leads to a trajectory planning method based on the recursive algorithm to adjust the position and posture of the welding torch to realize smooth adjustments and transitions of the welding torch during welding of the right-angle turning point of the box-type steel structures. This approach provides trajectory planning for the all-position welding of box-type steel in an automatic algorithm that is applicable to different sizes of box-type steel structures.

A 6-DOF (degree of freedom) parallel machine was used as the reset mechanism in a wind tunnel balance calibration system because of its simple structure and stiffness. A model was developed using the D-H (Denavit-Hartenberg) convention to analyze inaccuracies in the reset mechanism caused by joint manufacturing tolerances. In this model, each closed-loop kinematic chain was treated as a RRRPRR chain. The effects of the joint tolerances and structural parameters on the kinematic accuracy were calculated based on Qusai-Newton method. Simulations indicate that the inaccuracy is on the same order of magnitude as the joint offsets. A principle is given to determine the structural parameters needed to improve the accuracy of this type of reset system.

Relief materials need to be rapidly transported to disaster areas after natural disasters. This paper describes how to minimize the transport by maximizing the number of victims supplied by the relief materials. This integer programming method involves the substitution and collocation of relief materials and vehicles to maximize the number of victims served and minimize the transport costs. Simulations show that the time complexity of the vehicle dispatch is significantly reduced after the problem is decomposed into sub-problems. The substitution and collocation reduces the transport costs while maximizing the number of victims supplied by relief materials.

An coupled control strategy was developed for proximity operations of spacecraft formation flying which tabes into account the unknown but bounded disturbance, parametric uncertainties and actuator saturation. An auxiliary system is used to compensate for the effects of input constraints. The auxiliary system states are applied to the control design and stability analysis. The system was an adaptive updating technique that does not need to know the system parameters and the unknown bound of the external disturbance in advance Lyapunov theory shows that this controller guarantees uniform ultimate boundedness of all the signals in the closed-loop system and that the tracking errors will converge to a small neighborhood of the origin. Finally, a six degree of freedom simulation demonstrates the effectiveness of this control law compared to an approach that does not consider actuator saturation.

A multi-party incentive model shows how the project characteristics, participant characteristics and incentive structure promote a project. The real effect of the participant cost, project optimization and incentive structure are illustrated through game simulations. The results show that the importance of the incentive model. High level of optimization potential, low team costs and the team members' cost variance contribute to the project success. The target cost setting, goal achievement compensation, and income distribution are critical project incentives.

Microalgae biodiesel is a complex industrial energy production method that is influenced by many factors including the complex composition. The energy consumption and generation ability and the amount of greenhouse gas emissions (net carbon emissions) are important criteria for assessing the system sustainability. A life cycle analysis of the energy and carbon balance of microalgae biodiesel were used to calculated the net energy surplus and net carbon emissions of biodiesel biofuel and the key factors influencing the energy production and carbon emissions. The results show that the energy output of the whole system is larger than the energy consumption during the production and processing of the microalgae biodiesel and that more CO₂ is fixed than released over the life time. Thus, microalgae biodiesel is a new fuel that is sustainable and environmentally friendly.

The carrier and demodulated signals of eddy current displacement sensors are both analog signals. Thus, the measurement and demodulation circuitries are complex, unstable and increase the volume and power consumption of the magnetic bearing system. This paper describes an improved digital eddy current displacement sensor based on amplitude modulation with the digital carrier signal given by a field programmable gate array (FPGA) to demodulate the digital output signal of the sensor. The measurement reliability and accuracy are improved while the complexity and power consumption are reduced in this integrated digital eddy current displacement sensor.