Deep Reinforcement Learning-Based Task Scheduling in IoT Edge Computing.

Edge computing (EC) has recently emerged as a promising paradigm that supports resource-hungry Internet of Things (IoT) applications with low latency services at the network edge. However, the limited capacity of computing resources at the edge server poses great challenges for scheduling application tasks. In this paper, a task scheduling problem is studied in the EC scenario, and multiple tasks are scheduled to virtual machines (VMs) configured at the edge server by maximizing the long-term task satisfaction degree (LTSD). The problem is formulated as a Markov decision process (MDP) for which the state, action, state transition, and reward are designed. We leverage deep reinforcement learning (DRL) to solve both time scheduling (i.e., the task execution order) and resource allocation (i.e., which VM the task is assigned to), considering the diversity of the tasks and the heterogeneity of available resources. A policy-based REINFORCE algorithm is proposed for the task scheduling problem, and a fully-connected neural network (FCN) is utilized to extract the features. Simulation results show that the proposed DRL-based task scheduling algorithm outperforms the existing methods in the literature in terms of the average task satisfaction degree and success ratio.

Cognitive Frequency-Hopping Waveform Design for Dual-Function MIMO Radar-Communications System.

A frequency-hopping (FH)-based dual-function multiple-input multiple-output (MIMO) radar communications system enables implementation of a primary radar operation and a secondary communication function simultaneously. The set of transmit waveforms employed to perform the MIMO radar task is generated using FH codes. For each transmit antenna, the communication operation can be realized by embedding one phase symbol during each FH interval. However, as the radar channel is time-variant, it is necessary for a successive waveform optimization scheme to continually obtain target feature information. This research work aims at enhancing the target detection and feature estimation performance by maximizing the mutual information (MI) between the target response and the target returns, and then minimizing the MI between successive target-scattering signals. The two-step cognitive waveform design strategy is based upon continuous learning from the radar scene. The dynamic information about the target feature is utilized to design FH codes. Simulation results show an improvement in target response extraction, target detection probability and delay-Doppler resolution as the number of iterations increases, while still maintaining high data rate with low bit error rates between the proposed system nodes.

A RPCA-Based ISAR Imaging Method for Micromotion Targets.

Micro-Doppler generated by the micromotion of a target contaminates the inverse synthetic aperture radar (ISAR) image heavily. To acquire a clear ISAR image, removing the Micro-Doppler is an indispensable task. By exploiting the sparsity of the ISAR image and the low-rank of Micro-Doppler signal in the Range-Doppler (RD) domain, a novel Micro-Doppler removal method based on the robust principal component analysis (RPCA) framework is proposed. We formulate the model of sparse ISAR imaging for micromotion target in the framework of RPCA. Then, the imaging problem is decomposed into iterations between the sub-problem of sparse imaging and Micro-Doppler extraction. The alternative direction method of multipliers (ADMM) approach is utilized to seek for the solution of each sub-problem. Furthermore, to improve the computational efficiency and numerical robustness in the Micro-Doppler extraction, an SVD-free method is presented to further lessen the calculative burden. Experimental results with simulated data validate the effectiveness of the proposed method.

Adaptive chaotic MIMO radar based on DPMM clustering and Kalman filtering technique.

A novel scheme to optimize the adaptive transmit waveform of chaotic multiple-input multiple-output (MIMO) radar is developed. The main objective of this work is to achieve high ability in target discrimination by using a Dirichlet process mixture model (DPMM)-based clustering method based on nonparametric Bayesian theory and to improve the capability of target detection by minimizing the mean square error of radar channel response via Kalman filtering (KF) technique. The two stages are the discrimination of multiple range-extended targets and the optimization of the adaptive chaos-based waveform for transmission. The adaptive chaotic MIMO waveform optimization scheme overcomes the problem of target discrimination and detection in an intelligent transportation system, where there is a need for extracting the feature of target information achieved from vehicle-mounted sensor. As the number of iterations increases, simulation experiments demonstrate better target discrimination capability provided by the proposed DPMM-KF technique as compared with the traditional waveform design method. In addition, the proposed DPMM-KF technique leads to improved target detection probability and receiver operating characteristics in the interference environment.

Doppler Data Association Scheme for Multi-Target Tracking in an Active Sonar System.

In many wireless sensors, the target kinematic states include location and Doppler information that can be observed from a time series of range and velocity measurements. In this work, we present a tracking strategy for comprising target velocity components as part of the measurement supplement procedure and evaluate the advantages of the proposed scheme. Data association capability can be considered as the key performance for multi-target tracking in an active sonar system. Then, we proposed an enhanced Doppler data association (DDA) scheme which exploits target range and target velocity components for linear multi-target tracking. If the target velocity measurements are not incorporated into target kinematic state tracking, the linear filter bank for the combination of target velocity components can be implemented. Finally, a significant enhancement in the multi-target tracking capability provided by the proposed DDA scheme with the linear multi-target combined probabilistic data association method is demonstrated in a sonar underwater scenario.

Waveform Optimization for Target Estimation by Cognitive Radar with Multiple Antennas.

A new scheme based on Kalman filtering to optimize the waveforms of an adaptive multi-antenna radar system for target impulse response (TIR) estimation is presented. This work aims to improve the performance of TIR estimation by making use of the temporal correlation between successive received signals, and minimize the mean square error (MSE) of TIR estimation. The waveform design approach is based upon constant learning from the target feature at the receiver. Under the multiple antennas scenario, a dynamic feedback loop control system is established to real-time monitor the change in the target features extracted form received signals. The transmitter adapts its transmitted waveform to suit the time-invariant environment. Finally, the simulation results show that, as compared with the waveform design method based on the MAP criterion, the proposed waveform design algorithm is able to improve the performance of TIR estimation for extended targets with multiple iterations, and has a relatively lower level of complexity.

Adaptive Waveform Design for MIMO Radar-Communication Transceiver.

The system architecture for an adaptive multiple input multiple output (MIMO) radar-communication transceiver is proposed. A waveform design approach for communication data embedding into MIMO radar pulse using M-ary position phase shift keying (MPPSK) waveforms is introduced. A waveform optimization algorithm for the adaptive system is presented. The algorithm aims to improve the target detection performance by maximizing the relative entropy (RE) between the distributions under existence and absence of the target, and minimizing the mutual information (MI) between the current received signals and the estimated signals in the next time instant. The proposed system adapts its MPPSK modulated inter-pulse duration to suit the time-varying environment. With subsequent iterations of the algorithm, simulation results show an improvement in target impulse response (TIR) estimation and target detection probability. Meanwhile, the system is able to transmit data of several Mbps with low symbol error rates.

Cognitive Radar Waveform Optimization Based on Mutual Information and Kalman filtering.

A new strategy to optimizing the waveforms of cognitive radar under transmitted power constraint is presented. Our scheme is to enhance the performance of target estimation by minimizing the MSE (mean-square error) of the estimates of target scattering coefficients (TSC) based on Kalman filtering and then minimizing mutual information (MI) between the radar target echoes at successive time instants. The two steps are the optimal design of transmission waveform and the selection of a reasonable waveform from the ensemble for emission, respectively. The waveform design technique addresses the problems of target detection and parameter estimation in intelligent transportation system (ITS), where there is a need of extracting the features of target information obtained from different sensors. As the number of iterations increases, simulation results show better TSC estimation from the radar scene provided by the proposed approach as compared with the traditional waveform optimization algorithm. In addition, the proposed algorithm results in improved target detection probability.

Performance analysis of single-frequency MPPSK integrated system for ranging applications.

The dual-frequency MPPSK-MODEM is a flexible platform. When ranging accuracy request is low or platform is particularly limited by power, the platform would perform both data transmission and range measurement with single-frequency modes. In this paper, the ranging resolution of MPPSK pulse waveforms with the match filter and impacting filter processing are discussed, respectively. Also, the parameter selection of MPPSK modulator for ranging is considered. In particular, requirements that allow for employing such special parameter values for range measurements with high accuracy and high range are investigated. Moreover, high repetition frequency (HRF) biphase code MPPSK pulse train base on m sequence is presented, and the ranging accuracy of the proposed signal with the match filter processing is deduced. In addition to theoretical considerations, the paper presents system simulations and measurement results of single-frequency MPPSK integrated systems, demonstrating the high-performance of ranging applications.

The dual carrier ABSK system based on a FIR bandpass filter.

The special impacting filter (SIF) with IIR structure has been used to demodulate ABSK signals. The key points of SIF, including the resonance circuit's high Q value and the "slope-phase discrimination" character of the filter sideband, are demonstrated in the paper. The FIR narrow-band bandpass filtering system, which can also provide the impact-filtering effect, is proposed. A dual carrier system of ABSK signals is designed with the proposed FIR filter as its receiver. The simulation results show that the FIR filter can work well. Moreover, compared to the traditional SIF, the proposed FIR filter can not only achieve higher spectral efficiency, but also give better demodulation performance.

Nonlinear detection for a high rate extended binary phase shift keying system.

The algorithm and the results of a nonlinear detector using a machine learning technique called support vector machine (SVM) on an efficient modulation system with high data rate and low energy consumption is presented in this paper. Simulation results showed that the performance achieved by the SVM detector is comparable to that of a conventional threshold decision (TD) detector. The two detectors detect the received signals together with the special impacting filter (SIF) that can improve the energy utilization efficiency. However, unlike the TD detector, the SVM detector concentrates not only on reducing the BER of the detector, but also on providing accurate posterior probability estimates (PPEs), which can be used as soft-inputs of the LDPC decoder. The complexity of this detector is considered in this paper by using four features and simplifying the decision function. In addition, a bandwidth efficient transmission is analyzed with both SVM and TD detector. The SVM detector is more robust to sampling rate than TD detector. We find that the SVM is suitable for extended binary phase shift keying (EBPSK) signal detection and can provide accurate posterior probability for LDPC decoding.

Nonlinear demodulation and channel coding in EBPSK scheme.

The extended binary phase shift keying (EBPSK) is an efficient modulation technique, and a special impacting filter (SIF) is used in its demodulator to improve the bit error rate (BER) performance. However, the conventional threshold decision cannot achieve the optimum performance, and the SIF brings more difficulty in obtaining the posterior probability for LDPC decoding. In this paper, we concentrate not only on reducing the BER of demodulation, but also on providing accurate posterior probability estimates (PPEs). A new approach for the nonlinear demodulation based on the support vector machine (SVM) classifier is introduced. The SVM method which selects only a few sampling points from the filter output was used for getting PPEs. The simulation results show that the accurate posterior probability can be obtained with this method and the BER performance can be improved significantly by applying LDPC codes. Moreover, we analyzed the effect of getting the posterior probability with different methods and different sampling rates. We show that there are more advantages of the SVM method under bad condition and it is less sensitive to the sampling rate than other methods. Thus, SVM is an effective method for EBPSK demodulation and getting posterior probability for LDPC decoding.

Classification of fruits using computer vision and a multiclass support vector machine.

Automatic classification of fruits via computer vision is still a complicated task due to the various properties of numerous types of fruits. We propose a novel classification method based on a multi-class kernel support vector machine (kSVM) with the desirable goal of accurate and fast classification of fruits. First, fruit images were acquired by a digital camera, and then the background of each image was removed by a split-and-merge algorithm; Second, the color histogram, texture and shape features of each fruit image were extracted to compose a feature space; Third, principal component analysis (PCA) was used to reduce the dimensions of feature space; Finally, three kinds of multi-class SVMs were constructed, i.e., Winner-Takes-All SVM, Max-Wins-Voting SVM, and Directed Acyclic Graph SVM. Meanwhile, three kinds of kernels were chosen, i.e., linear kernel, Homogeneous Polynomial kernel, and Gaussian Radial Basis kernel; finally, the SVMs were trained using 5-fold stratified cross validation with the reduced feature vectors as input. The experimental results demonstrated that the Max-Wins-Voting SVM with Gaussian Radial Basis kernel achieves the best classification accuracy of 88.2%. For computation time, the Directed Acyclic Graph SVMs performs swiftest.

Crop classification by forward neural network with adaptive chaotic particle swarm optimization.

This paper proposes a hybrid crop classifier for polarimetric synthetic aperture radar (SAR) images. The feature sets consisted of span image, the H/A/α decomposition, and the gray-level co-occurrence matrix (GLCM) based texture features. Then, the features were reduced by principle component analysis (PCA). Finally, a two-hidden-layer forward neural network (NN) was constructed and trained by adaptive chaotic particle swarm optimization (ACPSO). K-fold cross validation was employed to enhance generation. The experimental results on Flevoland sites demonstrate the superiority of ACPSO to back-propagation (BP), adaptive BP (ABP), momentum BP (MBP), Particle Swarm Optimization (PSO), and Resilient back-propagation (RPROP) methods. Moreover, the computation time for each pixel is only 1.08 × 10(-7) s.

Mobile location with NLOS identification and mitigation based on modified Kalman filtering.

In order to enhance accuracy and reliability of wireless location in the mixed line-of-sight (LOS) and non-line-of-sight (NLOS) environments, a robust mobile location algorithm is presented to track the position of a mobile node (MN). An extended Kalman filter (EKF) modified in the updating phase is utilized to reduce the NLOS error in rough wireless environments, in which the NLOS bias contained in each measurement range is estimated directly by the constrained optimization method. To identify the change of channel situation between NLOS and LOS, a low complexity identification method based on innovation vectors is proposed. Numerical results illustrate that the location errors of the proposed algorithm are all significantly smaller than those of the iterated NLOS EKF algorithm and the conventional EKF algorithm in different LOS/NLOS conditions. Moreover, this location method does not require any statistical distribution knowledge of the NLOS error. In addition, complexity experiments suggest that this algorithm supports real-time applications.

Novel anti co-channel interference scheme for sensor networks.

With improvement of the automation level, wireless sensors are widely used, but various kinds of interference lead to problems in the application. In order to deal with co-channel interference, a throughput efficient scheme based on Extended Binary Phase Shift Keying (EBPSK) modulation is introduced in physical layer. On this basis, the corresponding transmission scheme and the important impacting filter are presented. Effects of co-channel interference on the EBPSK waveform in different cases are analyzed. Simulation results illustrate the excellent anti-interference performance of the EBPSK system itself, when the initial phase of the co-channel interference is small.

On BER performance of EBPSK-MODEM in AWGN channel.

In order to satisfy the higher and higher demand for communication systems, an Extended Binary Phase Shift Keying (EBPSK) system with very high spectra efficiency has been proposed. During the research, a special kind of filters was found, which can amplify the signal characteristics and remove utmost noise, i.e., at the point of the phase jumping corresponding to code "1", produce the amplitude impulse much higher than code "0", therefore, the aim of our study was to analyze the BER performance of the impacting filter assisted EBPSK-MODEM. Considering the receiver filtered "0" and "1"signal with Rice amplitude distribution, just having different mean values, so the BER performance of EBPSK is deduced based on the classic detection theory, and compared with the traditional BPSK modulation both in spectra efficiency and in BER performance, which lays the theoretical foundation for the feasibility of Ultra Narrow Band communications based on EBPSK modulation.

Remote-sensing image classification based on an improved probabilistic neural network.

This paper proposes a hybrid classifier for polarimetric SAR images. The feature sets consist of span image, the H/A/α decomposition, and the GLCM-based texture features. Then, a probabilistic neural network (PNN) was adopted for classification, and a novel algorithm proposed to enhance its performance. Principle component analysis (PCA) was chosen to reduce feature dimensions, random division to reduce the number of neurons, and Brent's search (BS) to find the optimal bias values. The results on San Francisco and Flevoland sites are compared to that using a 3-layer BPNN to demonstrate the validity of our algorithm in terms of confusion matrix and overall accuracy. In addition, the importance of each improvement of the algorithm was proven.

Pattern Recognition via PCNN and Tsallis Entropy.

In this paper a novel feature extraction method for image processing via PCNN and Tsallis entropy is presented. We describe the mathematical model of the PCNN and the basic concept of Tsallis entropy in order to find a recognition method for isolated objects. Experiments show that the novel feature is translation and scale independent, while rotation independence is a bit weak at diagonal angles of 45° and 135°. Parameters of the application on face recognition are acquired by bacterial chemotaxis optimization (BCO), and the highest classification rate is 72.5%, which demonstrates its acceptable performance and potential value.

Comments on "a separable low complexity 2D HMM with application to face recognition'.

The separable low complexity 2D HMM proposed in [1] builds on the assumption of conditional independence in the relationship between adjacent blocks. The authors view this assumption as the key assumption to reduce the complexity. In this communication, we show that this key assumption is entirely unnecessary.