Ronald Riley

Panchromatic Enhanced Super-Resolution of Multi-Spectral Imagery

Using high-resolution Panchromatic (Pan) imagery to sharpen Multi-Spectral Imagery (MSI) has been... more Using high-resolution Panchromatic (Pan) imagery to sharpen Multi-Spectral Imagery (MSI) has been an active area of research for a number of years. The primary innovations of our approach is casting the problem as a superresolution of the MSI to the sampling geometry of the Pan, using the Pan as a template for object boundaries and cross-calibrating the Pan/MSI spatially and spectrally. This perspective leads us to a fused image with the sampling geometry (high spatial resolution) of the Pan which preserves the radiometric calibration (high spectral resolution) of the MSI. The Pan and MSI do not provide sufficient information for this fused product; additional constrains must be imposed to obtain a unique solution. Preserving the radiometric calibration of the MSI, and our choice of additional constraints are motivated by our ultimate goal of material classification of the MSI at the Pan resolution.

Download

A New Way to Detect Cyberattacks: Extracting Changes in Register Values From Radio-Frequency Side Channels

IEEE Signal Processing Magazine, 2019

The Internet of Things (IoT) and the Internet of Everything (IoE) have driven processors into nea... more The Internet of Things (IoT) and the Internet of Everything (IoE) have driven processors into nearly every powered device from thermostats to refrigerators to light bulbs. From a security perspective, IoT/IoE create a new layer of signals and systems that can provide insight into the internal operations of a device via analog side channels. Our research focuses on leveraging these analog side channels in IoT/IoE processors to detect intrusions. Our goal is to defend against cyber attacks that insert malware into IoT devices by detecting deviations in the code running on their processors from known firmware. Our approach applies signal-processing and machinelearning techniques to collected radiofrequency (RF) emissions of code running on the processor. We describe our process for positioning a wide-bandwidth RF probe over the processor of the device under test (DuT), then implement classifiers to identify the code running on the device and detect, identify, and isolate register contents based on signatures learned during DuT characterization. The probe is placed to improve classifier discrimination between instructions running on the DuT processor. At this selected location, the signatures of registers are subsequently extracted by separating the signals into execution components (i.e., instruction fetch, and decode opcode). These techniques enable a significant reduction in feature dimensions, which improves the speed and accuracy of extracting the number of bit changes, aka the Hamming Distance (HD), in registers via the RF side channel. Using these signatures, we extract the HD of registers with a weighted average accuracy of greater than 99%.

Download

Approved for public release. Distribution unlimited

We investigated the tradeoffs between accuracy and battery-energy longevity of acoustic beamformi... more We investigated the tradeoffs between accuracy and battery-energy longevity of acoustic beamforming on disposable sensor nodes subject to varying key parameters: 1) number of microphones, 2) duration of sampling, 3) number of search angles, and 4) CPU clock speed. Beyond finding the most energy efficient implementation of the beamforming algorithm at a specified accuracy, we seek to enable application-level selection of accuracy based on the energy required to achieve this accuracy. Our energy measurements were taken on the HiDRA node, provided by Rockwell Science Center, employing a 133-MHz StrongARM processor. We compared the accuracy and energy of our time-domain beamformer to a Fourier-domain algorithm provided by the Army Research Laboratory (ARL). With statistically identical accuracy, we measured a 300x improvement in energy efficiency of the CPU relative to this baseline. We also present other algorithms under development that combine results from multiple nodes to provide m...

Power-Aware Acoustic Processing 1

Abstract. We investigated tradeoffs between accuracy and battery-energy longevity of acoustic bea... more Abstract. We investigated tradeoffs between accuracy and battery-energy longevity of acoustic beamforming on disposable sensor nodes subject to varying key parameters: number of microphones, duration of sampling, number of search angles, and CPU clock. Beyond finding the most energy efficient implementation of the beamforming algorithm at a specified accuracy, we enable application-level selection of accuracy based on the energy required to achieve this accuracy. We measured the energy consumed by the HiDRA node, provided by Rockwell Science Center, employing a 133-MHz StrongARM processor. We compared the accuracy and energy of our time-domain beamformer to a Fourier-domain algorithm provided by the Army Research Laboratory (ARL). With statistically identical accuracy, we measured a 300x improvement in energy efficiency of the CPU relative to this baseline. We present other algorithms under development that combine results from multiple nodes to provide more accurate line-of-bearing...

Generalization of algorithm recognition in RF side channels between devices

Cyber Sensing 2018

All digital devices leak information through unintended emissions into analog side channels. The ... more All digital devices leak information through unintended emissions into analog side channels. The RF side channel enables passive collection of high-bandwidth information about the digital state of the device. We collected these RF emissions with a 500-MHz Riscure probe placed in the nearfield of the device under test (DuT) and applied machine learning to detect what program is running on the processor to identify malware intrusions. We explored the applicability of a generalized algorithm classification infrastructure built from a training set of similar DuTs to a similar device from a different production batch (same model number, different serial number.) We collected RF-SC data for five programs running on 28 distinct Arduino Unos (and 28 MSP430 processors.) We trained program classifiers on RF data from all but one DuT and tested the classifiers on the device withheld from the training set. The high-SNR signal provided by the Riscure probe enabled almost perfect classification results when we trained and tested on the same device. Our classification results remained above 99% when we generalized testing to the new DuT of the same model but a different serial number. The classifier was trained on 27 of the devices and tested to determine its ability to detect deviations from a baseline algorithm on a withheld device. The worst misclassification rate was a mere 0.08%.

Block-level algorithm classification based on RF side-channel

Cyber Sensing 2018

Internet of Things (IoT) and other similar devices often have little to no security and thus can ... more Internet of Things (IoT) and other similar devices often have little to no security and thus can be readily exploited in any number of ways. In this work, we collect radio frequency (RF) emissions from simple processors on several IoT devices and apply machine learning techniques to detect modifications (corrupted or injected via malware) in ‘known’ software running on the processor. We can detect these modifications due to the correlation between RF emissions and the digital state of the devices. Every bit flip produces a small but potentially detectable electrical pulse. Our approach to developing the recognition algorithm is to adapt to the variability created by the input data by recognizing the sequences in which instruction blocks are executed. Seemingly minor changes to input values can have a detectable effect on the measured RF side channel. We collect RF data from a variety of IoT devices with clock speeds varying from 16-96 MHz. A 1-GHz Riscure RF near-field antenna probe was placed within a millimeter of the IoT device, RF emissions were acquired, and software controls triggered data collection. A classification architecture was trained using object code portioned into blocks to develop the truth data. We then applied new data to the trained block classifier. This approach detects deviations in individual blocks and block sequences as a whole, allowing a greater level of detection resolution than just binary ‘Yes/No’ classification. Initial testing results showed greater than 90% classification accuracy for block-level modifications, and we can detect deviations from truth data with 100% accuracy.

Dynamic Sensor Networks

: The DSN team investigated hardware and software technologies for unattended ground sensor appli... more : The DSN team investigated hardware and software technologies for unattended ground sensor applications. Task I: Distribution and Aggregation includes node localization techniques, low-power data link protocols, power aware routing protocols, and spatial addressing and routing. Task 2: Declarative Languages and Execution Environment includes topographical soldier interface and a sensor network simulation environment for algorithm development, deployment planning, and operational support. Finally, Task 3: Platforms include investigative hardware development to support laboratory communications, processing, or localization experiments as well as open-source COTS PDA integration and system software in support of the soldier interface.

System Level Applications of Adaptive Computing (SLAAC)

: SLAAC assembled a multidisciplinary team with expertise in signal processing, design tools and ... more : SLAAC assembled a multidisciplinary team with expertise in signal processing, design tools and adaptive architectures to develop distributed adaptive computing systems (ACS) and methodologies for Defense processing. SLAAC was organized around challenge applications that collectively guided the research, including: Synthetic Aperture Radar/Automatic Target Recognition (SAR/ATR), SONAR Beamforming, Wide-Band RF Receiver Processing Hyperspectral Image Processing, Electronic Counter Measures (ECM), and Infrared ATR. The challenges were selected to stress ACS systems in dimensions of computational density, memory bandwidth and latency, I/O bandwidth, and scalability. SLAAC had technology transitions in SAR, SONAR and ECM.

Exploiting side-channel emissions to detect changes in FPGA firmware

Field Programmable Gate Arrays (FPGAs) are increasingly vital components of electronic systems us... more Field Programmable Gate Arrays (FPGAs) are increasingly vital components of electronic systems used in numerous industries. FPGAs possess well-documented logic and hardware vulnerabilities that could allow an adversary to penetrate and manipulate FPGA-based electronic infrastructure. To detect such attacks against FPGA firmware, we developed a technique that exploits the unintended RF side-channel emitted from an FPGA. Our approach presumes that malicious modification to a trusted FPGA bitstream will result in changes in radio frequency (RF) emissions—changes that our technique can detect and measure using signal processing and machine learning. The development of our RF side-channel technique was divided into three tasks: (1) determine if firmware changes can be detected using side-channel emissions, (2) determine the minimum firmware change that can be detected, and (3) extend our approach to work across multiple devices of the same type. We used the Digilent Arty development boar...

A modular power-aware wireless microsensor architecture

Geospatial Modeling System for Colorizing Images and Related Methods

Geospatial Modeling System for Reducing Shadows and Other Obscuration Artifacts and Related Methods

Spatial and spectral calibration of a panchromatic, multispectral image pair

Approved for public release. Distribution unlimited. Power-Aware Acoustic Processing

We investigated the tradeoffs between accuracy and battery-energy longevity of acoustic beamformi... more We investigated the tradeoffs between accuracy and battery-energy longevity of acoustic beamforming on disposable sensor nodes subject to varying key parameters: 1) number of microphones, 2) duration of sampling, 3) number of search angles, and 4) CPU clock speed. Beyond finding the most energy efficient implementation of the beamforming algorithm at a specified accuracy, we seek to enable application-level selection of accuracy based on the energy required to achieve this accuracy. Our energy measurements were taken on the HiDRA node, provided by Rockwell Science Center, employing a 133-MHz StrongARM processor. We compared the accuracy and energy of our time-domain beamformer to a Fourier-domain algorithm provided by the Army Research Laboratory (ARL). With statistically identical accuracy, we measured a 300x improvement in energy efficiency of the CPU relative to this baseline. We also present other algorithms under development that combine results from multiple nodes to provide m...

Panchromatic Enhanced Super-Resolution of Multi-Spectral Imagery

Using high-resolution Panchromatic (Pan) imagery to sharpen Multi-Spectral Imagery (MSI) has been... more Using high-resolution Panchromatic (Pan) imagery to sharpen Multi-Spectral Imagery (MSI) has been an active area of research for a number of years. The primary innovations of our approach is casting the problem as a superresolution of the MSI to the sampling geometry of the Pan, using the Pan as a template for object boundaries and cross-calibrating the Pan/MSI spatially and spectrally. This perspective leads us to a fused image with the sampling geometry (high spatial resolution) of the Pan which preserves the radiometric calibration (high spectral resolution) of the MSI. The Pan and MSI do not provide sufficient information for this fused product; additional constrains must be imposed to obtain a unique solution. Preserving the radiometric calibration of the MSI, and our choice of additional constraints are motivated by our ultimate goal of material classification of the MSI at the Pan resolution.

Download

Optical flow registration of panchromatic/multi-spectral image pairs

Panchromatic Modulation of Multispectral Imagery

Non-Negative Maximum Likelihood Ica for Blind Source Separation of Images and Signals with Application to Hyperspectral Image Subpixel Demixing

Proceedings Icip International Conference on Image Processing, Oct 1, 2006

The use of independent component analysis (ICA) methods for blind source separation of signals an... more

Video Summarization Using Video Frames from Different Perspectives

by Jay Hackett and Ronald Riley

Non-Negative Maximum Likelihood ICA for Blind Source Separation of Images and Signals with Application to Hyperspectral Image Subpixel Demixing

by Adrian Peter, Ronald Riley, and Jay Hackett

2006 International Conference on Image Processing, 2006

The use of independent component analysis (ICA) methods for blind source separation of signals an... more

Uploads

Papers by Ronald Riley

Log In