Abstract: A method for aircraft maintenance scheduling includes using a scheduling environment as a reinforcement learning (RL) environment to simulate an operational concept, train an RL agent, and generate aircraft maintenance decisions and explanations; providing a decomposed reward Deep Q-Network (drDQN) algorithm, wherein the drDQN algorithm includes a first Deep Q-Network (DQN) and a second DQN; using the first DQN to maximize a mission accomplishment objective; using the second DQN to minimize a maintenance cost objective; providing a trained drDQN agent; using the trained drDQN agent to obtain the aircraft maintenance decisions and corresponding mission accomplishment and maintenance cost rewards; using a scheduling module to arrange aircraft maintenance activities; and using an explainable module to get reasons to detail why the decisions are made and present tradeoffs between the decisions and non-selected alternatives.

Abstract: A method for recognizing a low-probability-of-interception (LPI) radar signal waveform includes: obtaining, by a radar signal receiver, an LPI radar signal s(t), s(t) varying with time t; extracting, by a radar signal processor, an adaptive feature and a pre-defined analytical feature from the LPI radar signal s(t); combining, by the radar signal processor, the adaptive feature with the pre-defined analytical feature to generate a constructed adaptive feature; and applying, by the radar signal processor, a convolutional neural network (CNN) model to classify the constructed adaptive feature to recognize the LPI radar signal waveform.

Abstract: A method for detecting GPS spoofing attacks includes providing a trained deep learning (DL) model based on neural networks, feeding GPS signals into the trained DL model, using asymmetric Shapley values (ASVs) to calculate feature contributions, using the ASVs to assign a non-uniform distribution over an ordering of features, obtaining causal structures among the features, applying the ASVs to causal Shapley additive explanation to obtain Shapley attributions, incorporating the Shapley attributions and the causal structures, and detecting GPS spoofing attacks by running the trained DL model and using the causal structures, non-uniform distribution, feature contributions, and Shapley attributions.

Abstract: Embodiments of the present disclosure provide a method, a device, and a storage medium for domain adaptation for efficient learning fusion (DAELF). The method includes acquiring data from a plurality of data sources of a plurality of sensors; for each of the plurality of sensors, training an auxiliary classifier generative adversarial network (AC-GAN) by a hardware processor with data from each data source of the plurality of data sources, thereby obtaining a trained feature extraction network and a trained label prediction network for each data source; forming a decision-level fusion network or a feature-level fusion network; and training the decision-level fusion network or the feature-level fusion network with a source-only mode or a generate to adapt (GTA) mode; and applying the trained decision-level fusion network or the trained feature-level fusion network to detect a target of interest.

Abstract: A method for detecting fake images includes: obtaining an image for authentication, and hand-crafting a multi-attribute classifier to determine whether the image is authentic. Hand-crafting the multi-attribute classifier includes fusing at least an image classifier, an image spectrum classifier, a co-occurrence matrix classifier, and a one-dimensional (1D) power spectrum density (PSD) classifier. The multi-attribute classifier is trained by pre-processing training images to generate an attribute-specific training dataset to train each of the image classifier, the image spectrum classifier, the co-occurrence matrix classifier, and the 1D PSD classifier.

Abstract: A method for detecting spoofing attacks on a UAV includes receiving signals from the UAV that is an edge device of a decentralized network, transmitting the signals to a fog node of the decentralized network, and detecting the spoofing attacks by detecting slow shifting patterns and running a trained attack detection model at the fog node for the signals. The decentralized network is based on an edge-fog-cloud computing paradigm.

Abstract: A system including at least one antenna configured to (i) receive radio frequency (RF) signals emitted from a region of interest (ROI) and (ii) receive RF signals from the ROI having an RF emitting object of interest (OOI) therein; and at least one computer readable storage medium including instructions. In an embodiment, the instructions may, inter alia, create a synthesized altered ROI RF signature based on an ROI RF signature and an unaltered OOI RF signature; compare the synthesized altered ROI RF signature with the altered ROI RF signature; and determine if the RF emitting OOI in the ROI is the same OOI that was within the shield environment based on the comparison of the synthesized altered ROI RF signature with the altered ROI RF signature. In other embodiments, the system can provide, inter alia, tracking information using RF signals. Methods of using RF signal reception are also disclosed.

Abstract: The present disclosure provides a method, a system and a storage medium of PID-based automatic gain control for a satellite transponder system. The method includes receiving a sequence of sample signals; determining two different block sizes where a block size of a first block is greater than a block size of a second block; using the block size of the first block to compute a first signal-amplitude-ratio (SAR)-based gain value and using the block size of the second block to compute a second signal-amplitude-ratio (SAR)-based gain value by the AGC processor through the sequence of sample signals; calculating a to-be-applied gain control value of an m-th transmitted symbol at a n-th time step; and calculating a new AGC gain at the n-th time step according to the to-be-applied gain control value at the n-th time step and a corresponding AGC gain at the (n?1)-th time step.

Abstract: The present disclosure provides a method, a system and a storage medium of PID-based automatic gain control for a satellite transponder system. The method includes receiving a sequence of sample signals; determining two different block sizes where a block size of a first block is greater than a block size of a second block; using the block size of the first block to compute a first signal-amplitude-ratio (SAR)-based gain value and using the block size of the second block to compute a second signal-amplitude-ratio (SAR)-based gain value by the AGC processor through the sequence of sample signals; calculating a to-be-applied gain control value of an m-th transmitted symbol at a n-th time step; and calculating a new AGC gain at the n-th time step according to the to-be-applied gain control value at the n-th time step and a corresponding AGC gain at the (n?1)-th time step.

Abstract: The present disclosure provides a method for model predictive automatic gain control under additive white Gaussian noise jamming. The method includes predicting a plurality of consecutive signal values by an autoregressive integrated moving average model; calculating a signal average value of the plurality of consecutive signal values; calculating a gain control value using the signal average value of the plurality of consecutive signal values; if the gain control value is greater than a maximum control capability of a AGC processor, using the gain control value as a desired gain control value; or if the gain control value is equal to or less than the maximum control capability, using a minimum difference between an estimated amplitude and each of reference amplitudes in the LUT as the desired gain control value; and calculating a new AGC gain for a current time step according to the desired gain control value.

Abstract: A radar system includes a signal generator for generating an LFMCW signal, an intermediate frequency (IF) signal generator for generating an IF signal, an upper converter for increasing the frequency of the LFMCW signal by the IF to generate a radar signal, a mixer for decreasing the frequency of a received radar signal to generate an output signal, an IQ demodulator for decreasing the frequency of the output signal to generate a baseband signal, an analog-to-digital converter for transforming the baseband signal into a digital signal, and a micro controller for processing the digital signal.

Abstract: The present disclosure provides a method, a system and a storage medium of resilient human-on-the-loop range-only cooperative positioning of a plurality of unmanned aerial vehicles (UAVs). The method includes computing an initial exploitability using an initial distribution and an initial policy; performing a forward updating of a distribution of a portion of the plurality of UAVs, and performing a backward updating of a Q function of each UAV of the plurality of UAVs; for each time step, calculating a dual variable at an (i+1)-th iteration and calculating a policy at an (i+1)-th iteration; computing a ratio of an exploitability at the (i+1)-th iteration over the initial exploitability; and if the ratio is less than or equal to a pre-defined tolerance value, maintaining a policy at the i-th iteration; and if the ratio is greater than the pre-defined tolerance value, using the policy at the (i+1)-th iteration.

Abstract: A system and method for prior and post-image analysis for damage level assessments are provided. An Image-based Prior and Posterior Conditional Probability Learning (IP2CL) system and method is provided for Human Assistance and Disaster Response (HADR) and damage assessments (DA) situational assessment and awareness (SAA). Equipped with the IP2CL, matching prior and posterior disaster/action images are effectively encoded into one image which is then ingested into deep learning (DL) approaches to determine the damage levels. Two scenarios for practical uses are provided: pixel-wise semantic segmentation and patch-based global damage classification.

Abstract: Embodiments of the present disclosure provide a method of a burst-based route discovery process. The method includes sending a probing interest packet to an NDN network; when one NDN forwarder receives the probing interest packet from a corresponding face, sending the probing interest packet to neighboring NDN forwarders; after anyone NDN forwarder receives the probing interest packet, sending back a burst of K probing data packets; as the burst of K probing data packets being received by an NDN forwarder, evaluating gaps between arrival times of the burst of K probing data packets; determining an available network throughput level of a face of the NDN forwarder; and if determined available network throughput level indicates a predefined increase in network throughput, setting the face of the NDN forwarder as a face for forwarding interest packets; and sending the burst of K probing data packets to neighboring NDN forwarders.

Abstract: A method for recognizing a low-probability-of-interception (LPI) radar signal waveform includes: obtaining, by a radar signal receiver, an LPI radar signal s(t), s(t) varying with time t; extracting, by a radar signal processor, an adaptive feature and a pre-defined analytical feature from the LPI radar signal s(t); combining, by the radar signal processor, the adaptive feature with the pre-defined analytical feature to generate a constructed adaptive feature; and applying, by the radar signal processor, a convolutional neural network (CNN) model to classify the constructed adaptive feature to recognize the LPI radar signal waveform.

Abstract: The present disclosure provides a machine-learning-based real-time task scheduling method. The method includes, for a worker node, executing a training task distributed by a master node; collecting latency time lengths of each machine learning model under different CPU utilization and memory usage; calculating a mean squared error of the latency time lengths of each machine learning model; comparing machine learning models according to mean squared errors of latency time lengths to select a desirable machine learning model installing on the worker node; providing an API for the worker node; when receiving a task by the master node, requesting the worker node to predict a latency time length; and returning the predicted latency time length to the master node; and after the master node collects predicted latency time lengths of worker nodes, assigning the task to a corresponding worker node with a lowest predicted latency time length.

Abstract: The present disclosure provides a system of distributed edge computing for cooperative augmented reality with mobile sensing capability. The system includes a plurality of nodes configured to generate a plurality of data streams; and a plurality of distributed edge servers configured to process one or more tasks using the plurality of data streams. An Apache Storm distributed stream processing platform is installed and properly configured on each distributed edge server; the plurality of distributed edge servers includes one or more service modules installed on each distributed edge server and configured to process the one or more tasks; and the plurality of distributed edge servers includes a master distributed edge server and a plurality of slave distributed edge servers; and a scheduler is installed on the master distributed edge server and configured to distribute the one or more tasks to the plurality of distributed edge servers.

Abstract: Embodiments of the present disclosure provide a method, a device, and a storage medium for targeted adversarial discriminative domain adaptation (T-ADDA). The method includes pre-training a source model including a source feature encoder and a source classifier, adapting a target feature encoder, and generating a target model by concatenating the adapted target feature encoder with the pre-trained source classifier.

Abstract: Embodiments of the present disclosure provide a method of a burst-based route discovery process. The method includes sending a probing interest packet to an NDN network; when one NDN forwarder receives the probing interest packet from a corresponding face, sending the probing interest packet to neighboring NDN forwarders; after anyone NDN forwarder receives the probing interest packet, sending back a burst of K probing data packets; as the burst of K probing data packets being received by an NDN forwarder, evaluating gaps between arrival times of the burst of K probing data packets; determining an available network throughput level of a face of the NDN forwarder; and if determined available network throughput level indicates a predefined increase in network throughput, setting the face of the NDN forwarder as a face for forwarding interest packets; and sending the burst of K probing data packets to neighboring NDN forwarders.

Abstract: The present disclosure provides a method, a device, and a storage medium for decentralized optimal control for a large-scale multi-agent system. The method includes initializing errors to obtain an initialized error of each of an actor NN, a critic NN and a mass NN; initializing error thresholds to obtain an initialized error threshold of each of the actor NN, the critic NN and the mass NN; and if the initialized error of each of the actor NN, the critic NN and the mass NN is greater than or equal to a corresponding initialized error threshold, calculating weights of each of the actor NN, the critic NN and the mass NN, and updating the actor NN, the critic NN, and the mass NN; and calculating errors of each of the actor NN, the critic NN and the mass NN, and updating the actor NN, the critic NN, and the mass NN.