A5043823276- Wireless Signal Modulation Classification
- Blind Source Separation Techniques
- Digital Media Forensic Detection
- Speech and Audio Processing
- Advanced Adaptive Filtering Techniques
- Hate Speech and Cyberbullying Detection
- Radar Systems and Signal Processing
- Adversarial Robustness in Machine Learning
- Digital Filter Design and Implementation
- Internet Traffic Analysis and Secure E-voting
- Biometric Identification and Security
- Advanced Wireless Communication Techniques
- Full-Duplex Wireless Communications
- RFID technology advancements
- Face and Expression Recognition
- Neural Networks and Applications
- Image and Signal Denoising Methods
United States Naval Research Laboratory
2011-2019
K Lab (United States)
2012
With the increasing presence of cognitive radio networks as a means to address limited spectral resources, improved wireless security has become necessity. In particular, potential node impersonate licensed user demonstrates need for techniques authenticate radio's true identity. this paper, we use deep learning detect physical-layer attributes identification devices, and demonstrate performance our method on set IEEE 802.15.4 devices. Our is based empirical principle that manufacturing...
Unintentional modulations of the electromagnetic signal radio-frequency (RF) emitters are used to identify individual sources signals as unique from same type in a procedure known RF fingerprinting. It allows for identification and tracking physical threats, prevention unauthorized access, detecting cloning sensitive devices. Machine learning techniques assist fingerprinting by providing automatic recognition these aspects emitters. (RFID) tags common emitter track supplies also present...
A number of successful RF fingerprint classifiers have been demonstrated, but relatively few results evaluate the impact changing receiver used for training and inference. In this work, we record a set 25 ZigBee transmitters with 10 independent, unsynchronized receivers first show that similar performance may be achieved by neural network-based verification system on all when inference are performed same receiver. Next, significant degradation different We propose two methods to address...
As the Internet of Things (IoT) continues to expand, there is a growing necessity for improved techniques authenticate identity wireless transmitters. In this paper, we develop physical-layer authentication technique using neural network structure with both convolutional and recurrent components distinguish transmissions originating from particular target device all others. addition, demonstrate strong performance in realistic multipath channel environment, as well show that classifier...
Recently, a number of neural network approaches to physical-layer wireless security have been introduced. In particular, these are able authenticate the identity different transmitters by device-specific imperfections present in their transmitted signals. this paper, we introduce weakness training protocol approaches, namely, that generative adversarial (GAN) can be trained produce signals realistic enough force classifier errors. We show GAN learn signal without modifying bandwidth or data...
In this paper we present an approach to harness multi-spectral projections (MSPs) carefully shape and locate tones in the spectrum, enabling a new robust modulation which signal's discrete frequency support is used represent symbols. This method, called Frequency Position Modulation (FPM), innovative extension MT-FSK OFDM can be non-uniformly spread over many GHz of instantaneous bandwidth (IBW), resulting communications system that difficult intercept jam. The FPM symbols are recovered...
Products of time-series signals have found wide-spread application in many fields signal processing. For example, they are often used modeling and compensating distortions generated by analog mixed-signal components. Without excess bandwidth, the products will produce aliased artifacts that do not represent physical phenomenology components being modeled. To ameliorate effects these potentially unwanted, distortion products, we compare two multidimensional filters: first is a simple...