Jingyang Lu

Machine learning based Intelligent cognitive network using fog computing

In this paper, a Cognitive Radio Network (CRN) based on artificial intelligence is proposed to distribute the limited radio spectrum resources more efficiently. The CRN framework can analyze the time-sensitive signal data close to the signal source using fog computing with different types of machine learning techniques. Depending on the computational capabilities of the fog nodes, different features and machine learning techniques are chosen to optimize spectrum allocation. Also, the computing nodes send the periodic signal summary which is much smaller than the original signal to the cloud so that the overall system spectrum source allocation strategies are dynamically updated. Applying fog computing, the system is more adaptive to the local environment and robust to spectrum changes. As most of the signal data is processed at the fog level, it further strengthens the system security by reducing the communication burden of the communications network.

Markov logic network based complex event detection under uncertainty

In a cognitive reasoning system, the four-stage Observe-Orient-Decision-Act (OODA) reasoning loop is of interest. The OODA loop is essential for the situational awareness especially in heterogeneous data fusion. Cognitive reasoning for making decisions can take advantage of different formats of information such as symbolic observations, various real-world sensor readings, or the relationship between intelligent modalities. Markov Logic Network (MLN) provides mathematically sound technique in presenting and fusing data at multiple levels of abstraction, and across multiple intelligent sensors to conduct complex decision-making tasks. In this paper, a scenario about vehicle interaction is investigated, in which uncertainty is taken into consideration as no systematic approaches can perfectly characterize the complex event scenario. MLNs are applied to the terrestrial domain where the dynamic features and relationships among vehicles are captured through multiple sensors and information sources regarding the data uncertainty.

Intelligent path loss prediction engine design using machine learning in the urban outdoor environment

Due to the progressive expansion of public mobile networks and the dramatic growth of the number of wireless users in recent years, researchers are motivated to study the radio propagation in urban environments and develop reliable and fast path loss prediction models. During last decades, different types of propagation models are developed for urban scenario path loss predictions such as the Hata model and the COST 231 model. In this paper, the path loss prediction model is thoroughly investigated using machine learning approaches. Different non-linear feature selection methods are deployed and investigated to reduce the computational complexity. The simulation results are provided to demonstratethe validity of the machine learning based path loss prediction engine, which can correctly determine the signal propagation in a wireless urban setting.

A joint manifold leaning-based framework for heterogeneous upstream data fusion

A joint manifold learning fusion (JMLF) approach is proposed for nonlinear or mixed sensor modalities with large streams of data. The multimodal sensor data are stacked to form joint manifolds, from which the embedded low intrinsic dimensionalities are discovered for moving targets. The intrinsic low dimensionalities are mapped to resolve the target locations. The JMLF framework is tested on digital imaging and remote sensing image generation scenes with mid-wave infrared (WMIR) data augmented with distributed radio-frequency (RF) Doppler data. Eight manifold learning methods are explored to train the system with the neighborhood preserving embedding showing promise for robust target tracking using video–radio-frequency fusion. The JMLF method shows a 93% improved accuracy as compared to a standard target tracking (e.g., Kalman-filter based) approach.

Dynamic multi-arm bandit game based multi-agents spectrum sharing strategy design

For a wireless avionics communication system, a Multi-arm bandit game is mathematically formulated, which includes channel states, strategies, and rewards. The simple case includes only two agents sharing the spectrum which is fully studied in terms of maximizing the cumulative reward over a finite time horizon. An Upper Confidence Bound (UCB) algorithm is used to achieve the optimal solutions for the stochastic Multi-Arm Bandit (MAB) problem. Also, the MAB problem can also be solved from the Markov game framework perspective. Meanwhile, Thompson Sampling (TS) is also used as benchmark to evaluate the proposed approach performance. Numerical results are also provided regarding minimizing the expectation of the regret and choosing the best parameter for the upper confidence bound.

Smart antenna design for real-time multi-channel power spectral density estimation and target localization

Situational awareness is dependent on efficient spectrum use for data communication. In this paper, we describe spectrum band selection based on the target operations and localization. For wireless spectrum detection, given the system noise and signal information, the Neyman-Pearson based likelihood ratio test can provide the optimal detection performance under a certain probability of false alarms. However, in practice not all the information of alternative hypotheses are available. In this paper, a robust generalized likelihood ratio test (RGLRT) based detection is proposed without knowing channel information and signal information. An online subspace learning algorithm for direction of arrival (DOA) is introduced, which only uses fixed partial observation of antennas to estimate the subspace of the steering matrix. The subspace rank is not necessarily known at the beginning. The simulation results show that only partial observations can achieve a good DOA estimation performance with comparatively smaller estimation error.

Cognitive radio testbed for Digital Beamforming of satellite communication

Digital Beamforming (DB) in satellite communication (SATCOM) has drawn significant attention in recent years. DB design is flexible and adaptive among frequencies, bandwidths, and interferences. In this work, we design a cognitive radio testbed for SATCOM for digital beamforming design and implementation. The testbed is capable of simultaneously capturing signals from multiple antennas. Considering the conditions of jamming or interference, the design implements spectrum detection and compressive sensing for channel condition monitoring in a real time. Using the analyzed jammer and interference information, the incoming signal can be moved into a different frequency where the channel is vacant or at least meets acceptable conditions. The beamforming design is also adaptive to the frequency where each beam is weighted differently for a given channel condition. In addition, the satellite antenna array is used to determine the signal direction of arrival (DoA), The DB is capable of forming the beam in the direction that avoids interference and jamming. Moreover, the detection of the jamming signal direction can help localize the jammer. In the testbed design, a Universal Software Radio Peripheral (USRP) is used as a radio frequency (RF) front end component and GNU radio connecting to computer as signal processing back end. The system architecture includes the transponder receiver. And a USRP ×310 with twinRx receiver is capable of obtaining multiple copies of the signal captured by multiple receiving antennas.