Search

With the advent of big data era and the rapid improvement of raw data scale, feature selection, as the basis and critical technologies for data mining, plays an increasingly important role. However, most studies on feature selection methods, mainly directed to treat the single feature or overall feature subset, while the influence of the correlation and redundancy of features in the feature subset on the classification results is ignored. In this paper, a hybrid feature selection method based on feature subsets generated by factor analysis (FAFS_HFS) is proposed. Firstly, this method generates feature subsets from the maximum load (maximum explanatory power) of each feature through factor analysis. Then, minimal redundancy and maximal relevance (mRMR) and sequential forward selection (SFS) are used to remove the redundancy of each feature subset. Finally, fisher score based on feature subset (FSF-score) is utilized to evaluate and obtain the optimal feature subsets. Experiments are conducted on 14 datasets, the results show that FAFS_HFS method has higher classification accuracy and lower dimension on almost all datasets, especially in high-dimensional datasets, and it has competitive efficiency and classification performance compared with other methods.

Reachability analysis is one of the most basic and challenging problems in verification. We investigate this problem in multi-agent reinforcement learning (MARL) system by transforming the reachability analysis to decision-making problem tackled by mixed integer linear programming (MILP) solver Gurobi. We define syntax and semantics for multi-role strategy logic (mrSL) which is used to describe the reachability specification. The logic mrSL is a variant of SL to express properties, which provide a holistic perspective to describe reachability properties instead of specifying a specific agent reaches a certain state in the MARL system. And the algorithms to translate reachability property into MILP constraints are provided. A tool called MAReachAnalysis is introduced, which uses Gurobi to solve the corresponding reachability problem and evaluate it on the predator-prey task of multi-agent deep deterministic policy gradient (MADDPG) example, discussion of the experimental results obtained on a range of test cases.

The automatic generation of pencil drawings from natural images has been a hot spot in the field of nonphotorealistic rendering in recent years and has attracted the attention of many scholars. Aiming at the characteristics of single line drawings generated by the existing pencil drawing algorithm and stiff lines, this paper proposes a new edge generation Gradlent Morphological Edge Detector (GMED) algorithm, which uses the generated edge convolution to generate pencil line drawings, making stylized pictures more natural. First, the gradient map of the input image is extracted according to the feature of the tone rate of change around the line at the boundary. Second, to make up for the shortcomings of too thin and discontinuous lines in the gradient map, morphological operations are performed on the gradient map and convolved to generate a hand-drawn pencil line drawing. Finally, texture filling and tone mapping are performed, and the resulting line drawing is fused to obtain the final pencil drawing effect. Compared with the existing pencil drawing algorithm, the proposed GMED edge generation method is applied to the line drawing generation process to generate a hand-drawn style, fine and concise line drawing, and the texture simulation process is simple and efficient, combined with histogram matching. In the process, a more suitable GML method is used, so that the final pencil drawing not only has natural lines but also colors closer to the input image.

Ocean-moored buoys play an important role in global ocean environment monitoring. Motivated by building a sustainable ocean buoy observational network, a spatial optimization approach is proposed to site buoy stations to maximize spatial monitoring efficiency (SME). To achieve this goal, a non-linear, continuous maximum coverage location model named CMCP-Ocean was established, associated with a measurement method of the SME. Meanwhile, a heuristic framework based on the particle swarm optimization (PSO) algorithm was built to solve the CMCP-Ocean model, and optimization strategies including the multi-core parallel computing strategy, the particle velocity updating strategy based on spatial matching, and two potential station selection strategies related to the centroid-based random radiation method (CRRM) and random grid division method (RGDM) were established to improve computing performance. The effectiveness and efficiency of the PSO-based algorithm and the CMCP-Ocean model were verified by a series of experiments; the proposed computing schema named PSO-for-CMCP-Ocean has also proven to be practical and efficient. Finally, the PSO-for-CMCP-Ocean was applied to the buoy station selection of water mass monitoring in the Laizhou Bay of China, and a multi-scale sustainable site planning solution is reported.

Few-shot learning researches to learn a novel concept from a handful of labeled samples. Due to the small amount of training data, deep network has the risk of over-fitting. Although many previous approaches based on metric criterion can make significant progress to tackle this challenge, they not only ignore the association between query set and support set when learning sample representation, but also fail to focus greater attention in the target area. To cope with these issues, we propose a novel feature transformation network (FTN) for few-shot image classification. Specifically, to draw inferences about other instances from only a few examples, it is expected to learn a model that has more discriminative representation of the target attributes and robust generalization ability. To this end, we introduce an attention-based affinity matrix to transform the semantical enhanced embedding vectors of query samples by associating the support set, thereby guiding the network to learn a sample representation that embodies higher semantic information in the target area. Furthermore, aiming at highlighting the object region in the feature maps, and strengthening the pertinence of similarity measurement between samples, a global and local feature fusion module is designed to fuse the support set samples features. The comprehensive experiments validate the doable of our model, and our method achieves the state-of-the-art performance on two public benchmark datasets, namely, general object dataset mini-ImageNet and fine-grained dataset Caltech-UCSD Birds-200-2011 (CUB).

The communication infrastructure is most likely to be damaged after a major disaster occurred, which would lead to further chaos in the disaster stricken area. Modern rescue activities heavily rely on the wireless communications, such as safety status report, disrupted area monitoring, evacuation instruction, rescue coordination, etc. Large amount of data generated from victims, sensors and responders must be delivered and processed in a fast and reliable way, even when the normal communication infrastructure is degraded or destroyed. To this end, reconstructing the post-disaster network by deploying MDRU (Movable and Deployable Resource Unit) and relay unit at edge is a very promising solution. However, the optimal wireless access control in this heterogeneous hastily formed network is extremely challenging, due to the frequent varying environment and the lack of statistics information in advance in post-disaster scenarios. In this paper, we propose a learning based wireless access control approach for edge-aided disaster response network. More specifically, we model the wireless access control procedure as a discrete-time single agent Markov decision process, and solve the problem by exploiting deep reinforcement learning technique. By extensive simulation results, we show that the proposed mechanism significantly outperforms the baseline schemes in terms of delay and packet drop rate.

This paper focuses on estimating the angle and size of different-sized targets using a radar system that estimates the angles of trucks, vehicles, bicycles, and pedestrians. The problem is that the presence of a sidelobe from the larger target degrades the accuracy of estimating the angle and size of the smaller target in multiple target detection. In this paper, a novel scheme, called antenna element space interference cancelling (AIC), is proposed for reducing the influence of the sidelobe. The performance of the proposed AIC radar system was evaluated through both computer simulations and experiments. The simulation results show that the proposed AIC radar system can estimate the angles of smaller targets within a 1-degree error, and the size estimation error is 1 dB. The experimental results also show that the proposed AIC radar is effective in detecting smaller targets that cannot be detected with traditional methods due to the sidelobes of larger targets.

Radar signal sorting is the core part of electronic support measures, which is responsible for deinterleaving the overlapping pulse sequences received by the receiver from the complex environment, separating different radiation source signals, and providing support for radiation source identification. Particle swarm optimization (PSO) is a population-based global optimization algorithm with great advantages in intelligent sorting of complex signals, which can adapt to the electromagnetic environment with complex and variable radiation source signals and high pulse stream density. However, the PSO-based sorting method is prone to premature convergence and cannot adaptively adjust particle swarm parameters and positions. In this paper, a dynamic modified chaotic PSO algorithm (DMCPSO) is proposed. Chaotic search is used to increase the diversity of particle swarm in the later iteration to avoid premature convergence and falling into local optimum. Adaptive adjustment parameters related to the particle fitness value are adopted to balance the ability of global search and local search. A new fitness function is proposed and the particle position is dynamically corrected by clustering analysis to improve the accuracy of particle position optimization and avoid the influence from the distribution of feature parameters. The simulation results show that the DMCPSO algorithm provides stable and fast performance with excellent sorting indexes in complex, variable, and dense signal environment.

In many optical communication systems, in order to make the optical networking convenient and meet the application requirements, the optical terminals need to work in the same band and at the extremely closely adjacent working wavelengths. In practice, because of the light reflection and scattering of the wireless channels with the suspended particles, parts of the emitted optical signals will be reflected to the focal plane array (FPA). On the other hand, because the bandwidth of practical optical filters cannot be made narrow enough, or the sensitivity of FPAs is usually very high, the isolation between the transmitting and receiving optical paths in the terminal cannot be made high enough even if the optical filtering is used. Both the received optical signal from the remote terminal and the transmitted optical signal reflected by the terminal or wireless channels will be detected by the FPA at the same time. The reflected optical signal will interfere with the detection of the received optical signals from the far end, which seriously reduces the detection performance and the accuracy of the light spot centroid calculation. This paper proposes an integral control and increase-decrease counting statistical method for FPAs to eliminate the influence of the reflected optical signal on the signal detection and the noise mean value in the mean value of the total detection signal. The detection sensitivity of the received signal from the remote terminal, the isolation of the transmitting and receiving paths and the accuracy of the light spot centroid calculation are improved by this method. The Cramer-Rao lower bound is calculated and analyzed. In addition, the integral control and increase-decrease counting statistical algorithm is verified by the two software co-simulation, and the simulation results are given to illustrate the feasibility and performance of this method.

In view of the current situation, that physiological parameter monitoring systems can only achieve local monitoring, and the multi-physiological parameter monitors are large, expensive, and disadvantageous to remote monitoring. This paper combines embedded and mobile communication technologies to develop a new type of multi-physiological parameter medical monitoring system with remote data transmission function. First, through the analysis of embedded system principles, an embedded computer system based on ARM is designed. Secondly, the human-computer interaction interface, data acquisition, and analysis module are designed. Finally, by connecting to the Internet network to communicate with the medical center server, the remote transmission of local detection data and the issuance of alarm signals when dangerous situations occur are realized. The system can collect and display multiple physiological parameters such as heart rate, blood pressure, blood oxygen saturation, and body temperature in real time. The simulation experiment results show that the system's monitoring function and remote data transmission function meet the design requirements, can quickly and accurately find out-of-standard data, and perform remote alarm. The system is small, easy to expand, stable in data transmission, high in reliability, convenient for remote monitoring and data sharing, and is an ideal monitoring device for hospitals and community medical centers.

Internet of Things (IoT) devices are forecasted to increase to 20 billion by the year 2020. Questions are arising as where the bandwidth and radio channels come from to make it all work. It is a common belief that the spectrum under 6GHz is easy to use for various IoT applications. However, currently it has been almost fully allocated to different kinds of systems, some of which suffer from an extremely low spectrum utilization ratio. To this end, spectrum sharing approach has been proposed recently, which dynamically redistributes the incumbents' underutilized spectrum to secondary users and exchanges for profit. To incentivize the incumbent to share the spectrum in efficient and profitable ways, auction based mechanisms are extensively discussed. Most of the existing researches propose to share the spectrum by utilizing truthful auction mechanism, in which the secondary users' dominant strategies are to bid their true valuations. However in reality, exposing true valuation is very risky and thus keeping user's bidding value confidential is of great importance. In this paper, we propose an efficient privacy preserving spectrum sharing framework for IoT by utilizing ElGamal cyptosystem. The proposed scheme hides the user's bid from both the auctioneer and other users. By extensive evaluation results, we show that the proposed framework could achieve good spectrum utilization efficiency and user satisfactory ratio, at the cost of acceptable communication and computation overheads.

Trust management in a decentralized vehicular network, such as vehicular ad hoc network, is particularly challenging due to the lack of centralized communication infrastructure and a fast varying feature of the vehicular environment. In this paper, we propose a decentralized trust management scheme for vehicular networks. The proposed scheme uses a fuzzy logic-based trust calculation approach to evaluate the direct trust where trustee nodes are located within the transmission range of a trustor node. A reinforcement learning-based approach is also employed to estimate the indirect trust where the behaviors of trustee cannot be observed directly. The extensive simulations are conducted to show the advantage of the proposed scheme over other baseline approaches.

With the development of wireless communication systems, it is particularly essential to maximize the quality of experience (QoE) of machine-to-machine (M2M) communication. In this paper, we propose a new QoE-oriented uplink rate control and resource allocation scheme for the Internet of Things (IoT) network, by introducing an evaluation model based on mean opinion score (MOS) for different machine-type communication (MTC) devices. The existing works are only dedicated to solving the short-term resource allocation problems by considering the current transmission time slots, which cannot handle long-standing problems. To this end, based on the recently developed Lyapunov optimization, we convert the original long-term optimization problem into the admission rate control subproblem and the resource allocation subproblem in each time slot. To solve the joint power optimization and sub-channel selection subproblems, Gale-Shapley algorithm is utilized to formulate it as a two-dimensional matching problem, and the preference lists are established by the transmission rate and signal to interference plus noise ratio (SINR). In the proposed algorithms, a priority mechanism is employed to ensure fairness. The simulation results demonstrate that without prior knowledge of the data arrivals and sub-channel statistics, the proposed algorithms can significantly improve the overall perceived quality from the users' perspective.

Broadband Internet access in the building has fundamentally changed almost every aspect of our lives. As the growing use of indoor portable systems and devices, saving their energy consumption becomes an interesting and important issue for smart buildings. Recently, multi-band WLAN where 2.4/5-GHz and 60-GHz bands coexist is a promising solution to offer both ultra-high speed and robust wireless connections. For a multi-band WLAN end device, detecting the available service areas in an energy efficient way is of great importance. In the existing systems, the RF units of device need to be turned on and kept listening all the time, which leads to substantial energy consumption overhead. To solve this problem, this paper proposes an energy efficient learning-based indoor multi-band WLAN system, in which the end device predicts the distinct service areas by learning the influences of reflected waves in buildings. We have performed extensive experiments in different indoor environments, and the evaluation results demonstrate that the proposed mechanism could substantially improve the performance compared with the existing approaches.