Abstract Chaotic time series prediction is a prediction method based on chaos theory, and has important theoretical and application value. At present, most prediction methods only pursue digital fitting and do not consider the directional trend. In addition, using the single model will not achieve better prediction results. Therefore, a chaotic time series combined prediction model for improving trend lagging (ITL) is proposed. An improved dual‐stage attention‐based long short‐term memory model with the improved training objective fuction is designed to solve the trend lagging problem. Then, an auto regressive moving average model with the sliding window is established to mine other characteristics of the time series except nonlinear characteristic. Finally, the idea of optimization algorithm is introduced to construct a time series combined prediction model with high accuracy based on the above two models, so as to perform the chaotic time series prediction from multiple perspectives. Multiple datasets are selected as experimental datasets, and the proposed method is compared with common prediction methods. The results show that the proposed method can achieve single‐step prediction with high accuracy and effectively improve the lagging of chaotic time series prediction. This research can provide theoretical support for the complex chaotic time series prediction.

Abstract With the increasing demand of applications for the spread spectrum technique, especially the demand for data transmission rates and spectral efficiency, the advantages of the traditional direct sequence spread spectrum (DSSS) system are limited. Therefore, multi‐ary spread spectrum (M‐ary) technology, parallel combinatory spread spectrum (PCSS) technology, and code index modulation (CIM) technology have been proposed. Although these three new technologies can improve the data rate, they all face the problem of the large consumption of pseudo‐code resources. In order to solve the problem of pseudo‐code resources, a bit‐activation code index modulation (BA‐CIM) method is proposed. At the transmitter, considering the good correlation among multiple pseudo‐codes, the corresponding pseudo‐code activation principle is established, and the corresponding spreading pseudo‐code is activated by using the status of each bit of the index data according to the pseudo‐code activation principle. Then, multicode superposition processing is carried out to spread the modulation data. At the receiver, the corresponding activation pseudo‐code is obtained using the maximum peak‐to‐average ratio (MPAR) and secondary peak‐to‐average ratio (SPAR) judgement mechanisms to decode the multibit index data. Compared with existing methods, the proposed BA‐CIM method can not only achieve a better bit error rate performance but also use the least pseudo‐code resources. Moreover, BA‐CIM has the best comprehensive performance improvement and is far superior to other methods. This research can provide technical support for the application of efficient spread spectrum communication.

Due to the harsh deployment environment of the underwater coustic sensor networks (UASNs), a reliable and energy-saving routing algorithm has always been an important challenge and a hot topic. A Nonuniform clustering (NC) algorithm is designed first in which clusters are generated according to different node densities. Based on NC, the backbone of the underwater acoustic sensor network is formed in UASNs. To guarantee the reliability of data transmission of the backbone network, an NC routing algorithm based on a virtual gravitational potential field (NC_RVGPF) is proposed. This algorithm: 1) establishes a virtual gravitational potential field model to allow data transmission by 3-D underwater nodes; 2) designs the virtual gravitational potential energy by combining the transmission distance between the nodes, the residual energy of the nodes, and other parameters; and 3) selects the path of the highest average potential energy as being the optimal path of data transmission. The simulation results show that compared with the classical routing algorithm, the NC_RVGPF algorithm has higher transmission efficiency, less energy consumption, and can more effectively extend the network’s lifetime.

Abstract To improve the estimation accuracy of chaotic time series and reduce the computational complexity, a high‐precision adaptive blind estimation method is proposed. By introducing the pre‐treatment method, the adaptive identification model is established, and the linear estimation equation is derived. Through the organic combination of the model and the least square mechanism, the complexity of the model is limited. When changing the type and length of the sequence, the parameters can be changed adaptively, so as to construct the corresponding trajectory equation and realize the blind estimation of chaotic time series. The analysis and experimental results show that the estimation accuracy of this method can reach 10 –16 order of magnitude.

Abstract The weighted fractional Fourier transform (WFRFT) technology can substantially change the statistical characteristics of the signal, diversify the signal, and effectively hide the communication information. However, considering the uncertainty of parameters in unknown or harsh environment, the setting of demodulation parameters in the receiver cannot meet the requirements of restoring the original data, so the blind scanning technology is needed. Thus, a high resolution blind scanning method for WFRFT communication signals is proposed. By establishing the rules between the modulation order error and the theoretical bit error rate, we derive the correlation characteristics of the left and right modulation orders and then establish a rectangular correlation function. By using the rectangular correlation boundary and the correlation mapping channel, the estimated modulation order is obtained and the real data is restored. This method solves the blind scanning problems under the condition of unknown modulation and fuzzy modulation orders in harsh environments and improves the resolution. This method is suitable for both 4‐WFRFT communication systems and multiple parameter WFRFT communication systems.

Abstract With the continuous improvement of the application requirements of Direct Sequence Spread Spectrum (DSSS), especially the high requirements of data transmission rate, the advantages of traditional DSSS system are limited, so the index modulation DSSS technology appears. However, at present, there are still some key problems such as high‐order data rate, high bit error rate and high synchronisation complexity. Therefore, in view of the above bottleneck problems, the code index modulation method with multidimensional invisible data (MID‐CIM) is proposed. On the basis of deeply mining the hidden features of DSSS communication mechanism, the index order and shift order are established, and the two‐dimensional invisible data index transmission rules are established using a bit reversed and pseudo code (PN) set mapping mechanism, and then the three‐dimensional invisible data index transmission rules are established by using the cyclic displacement and pseudo code offset mechanism. The MID‐CIM method not only ensures the effective transmission of one‐dimensional data, but also uses the transmission of low‐speed one‐dimensional data to index transmit high‐speed two‐dimensional and three‐dimensional invisible data. This research can provide the technical basis for the efficient application of DSSS system, and also provide theoretical support and technical support for the new development and application of a communication system based on the spread spectrum system.

The continuously increasing number of connected smart devices has led to the emergence of a crucial fault detection challenge to the Internet of Things (IoT). In this study, we aim to identify a method for the effective detection of faults in IoT devices. An IoT network model is first established, and a data edge verification mechanism based on blockchain is proposed; the blockchain is used to ensure that the data cannot be tampered with, and their accuracy is verified using the edge. Finally, a data set accuracy weighted random forest based on particle swarm optimization is proposed. The simulation results demonstrate that the proposed detection algorithm is both effective and efficient.

Considering its simplicity and the uniform distribution of signal energy and interference after transformation, the weighted fractional Fourier transform (WFRFT) has been gradually applied to the field of communication. With in‐depth study of WFRFT, the number of weighted terms can be extended from the original four terms to any number. This is called a multi‐WFRFT ( M ‐WFRFT). As the number of weighted terms plays an important role in the receiving system, research into an M ‐WFRFT reception method compatible with different weighted terms is critical. In view of the high complexity of communication systems based on M ‐WFRFT, in order to solve the general receiving problems of the receiver when the parameters of the transmitter are not fixed or when multiple transmitters share the same receiving system, a regenerative transformation scan method for M ‐WFRFT signals is established. This is accomplished by adapting to dynamic processing conditions and avoiding the problem of high complexity. In this method, the 4‐WFRFT mechanism is introduced, regeneration weighting coefficients are constructed, and the regeneration order is given by combining the inherent relationship between the weighting coefficients and the order. The new method can achieve the purpose of receiving M ‐WFRFT signals with different number of items and different orders.

With explosive growth of malware, Internet users face enormous threats from Cyberspace, known as “fifth dimensional space.” Meanwhile, the continuous sophisticated metamorphism of malware such as polymorphism and obfuscation makes it more difficult to detect malicious behavior. In the paper, based on the dynamic feature analysis of malware, a novel feature extraction method of hybrid gram (H-gram) with cross entropy of continuous overlapping subsequences is proposed, which implements semantic segmentation of a sequence of API calls or instructions. The experimental results show the H-gram method can distinguish malicious behaviors and is more effective than the fixed-length n-gram in all four performance indexes of the classification algorithms such as ID3, Random Forest, AdboostM1, and Bagging.

The increasing sophistication of malware variants such as encryption, polymorphism, and obfuscation calls for the new detection and classification technology. In this paper, MalDeep, a novel malware classification framework of deep learning based on texture visualization, is proposed against malicious variants. Through code mapping, texture partitioning, and texture extracting, we can study malware classification in a new feature space of image texture representation without decryption and disassembly. Furthermore, we built a malware classifier on convolutional neural network with two convolutional layers, two downsampling layers, and many full connection layers. We adopt the dataset, from Microsoft Malware Classification Challenge including 9 categories of malware families and 10868 variant samples, to train the model. The experiment results show that the established MalDeep has a higher accuracy rate for malware classification. In particular, for some backdoor families, the classification accuracy of the model reaches over 99%. Moreover, compared with other main antivirus software, MalDeep also outperforms others in the average accuracy for the variants from different families.

The application-layer distributed denial of service (AL-DDoS) attack makes a great threat against cyberspace security. The attack detection is an important part of the security protection, which provides effective support for defense system through the rapid and accurate identification of attacks. According to the attacker’s different URL of the Web service, the AL-DDoS attack is divided into three categories, including a random URL attack and a fixed and a traverse one. In order to realize identification of attacks, a mapping matrix of the joint entropy vector is constructed. By defining and computing the value of EUPI and jEIPU, a visual coordinate discrimination diagram of entropy vector is proposed, which also realizes data dimension reduction from N to two. In terms of boundary discrimination and the region where the entropy vectors fall in, the class of AL-DDoS attack can be distinguished. Through the study of training data set and classification, the results show that the novel algorithm can effectively distinguish the web server DDoS attack from normal burst traffic.