Defense Notices

Optical sensors are increasingly prevalent devices whose costs tend to increase with their sensitivity. A hike in sensitivity is typically associated with fragility, rendering expensive devices vulnerable to threats of high intensity illumination. These potential costs and even security risks have generated interest in devices that maintain linear transparency under tolerable levels of illumination, but can quickly convert to opaque when a threshold is exceeded. Such a device is deemed an optical limiter. Copious amounts of research have been performed over the last few decades on optical nonlinearities and their efficacy in limiting. This work provides an overview of the existing literature and evaluates the applicability of known limiting materials to threats that vary in both temporal and spectral width. Additionally, we introduce the concept of plasmonic parametric resonance (PPR) and its potential for devising a new limiting material, the plasmonic parametric absorber (PPA). We show that this novel material exhibits a reverse saturable absorption behavior and promises to be an effective tool in the kit of optical limiter design.

The Internet of Things is been a rapidly growing field that offers improved data collection, analysis and automation as solutions for everyday problems. A smart-city is one major example where these solutions can be applied to issues with urbanization. And while these solutions can help improve the quality of live of the citizens, there are always security & privacy risks. Data collected in a smart-city can infringe upon the privacy of users and reveal potentially harmful information. One example is a surveillance system in a smart city. Research shows that people are less likely to commit crimes if they are being watched. Video footage can also be used by law enforcement to track and stop criminals. But it can also be harmful if accessible to untrusted users. A malicious user who can gain access to a surveillance system can potentially use that information to harm others. There are researched methods that can be used to encrypt the video feed, but then it is only accessible to the system owner. Polls show that public opinion of surveillance systems is declining even if they provide increased security because of the lack of transparency in the system. Therefore, it is vital for the system to be able to do its intended purpose while also preserving privacy and holding malicious users accountable. 

To help resolve these issues with privacy & accountability and to allow for collaboration, we propose Corvus, an IoT surveillance system that targets smart communities. Corvus is a collaborative blockchain based surveillance system that uses context-based image captioning to anonymously describe events & people detected. These anonymous captions are stored on the immutable blockchain and are accessible by other users. If they find the description from another camera relevant to their own, they can request the raw video footage if necessary. This system supports collaboration between cameras from different networks, such as between two neighbors with their own private camera networks. This paper will explore the design of this system and how it can be used as a privacy-preserving, but translucent & accountable approach to smart-city surveillance. Our contributions include exploring a novel approach to anonymizing detected events and designing the surveillance system to be privacy-preserving and collaborative.

Interference has been a subject of interest to radars for generations due to its ability to degrade performance. Commercial radars can experience radio frequency (RF) interference from a different RF service (such as radio broadcasting, television broadcasting, communications, satellites, etc.) if it operates simultaneously in the same spectrum. The RF spectrum is a finite asset that is regulated to mitigate interference and maximum resources. Recently, shared spectrum have been proposed to accommodate the growing commercial demand of communication systems.  Airborne radars, performing ground moving target indication (GMTI), encounter interference from clutter scattering that may mask slow-moving, low-power targets. Least-squares (LS) optimal and re-iterative minimum-mean square error (RMMSE) adaptive mismatch processing recent advancements are proposed for GMTI and shared spectrum. Each estimation technique reduces sidelobes, provides less signal-to-noise loss, and less resolution degradation than windowing. For GMTI, LS and RMMSE filters are considered with angle-Doppler filters and pre-existing interference cancellation techniques for better detection performance. Application specific reduce rank versions of the algorithms are also introduced for real-time operation. RMMSE is further considered to separate radar and mobile communication systems operating in the same RF band to mitigate interference and information loss.

Managed language virtual machines (VM) rely on dynamic or just-in-time (JIT) compilation to generate optimized native code at run-time to deliver high execution performance.  Many VMs and JIT compilers collect \emph{profile} data at run-time to enable profile-guided optimizations (PGO) that customize the generated native code to different program inputs.  PGOs are generally considered integral for VMs to produce high-quality and performant native code.  Likewise, many static, ahead-of-time (AOT) compilers employ PGOs to achieve peak performance, though they are less commonly employed in practice. 

We propose a study that analyzes and quantifies the performance benefits of PGOs in both AOT and JIT enviroments, understand the importance of profiling data quantity and quality/accuracy to effectively guide PGOs, and assess the impact of individual PGOs on performance.  Additionally, we propose an extension of PGOs found in AOT compiler based on specialization and seek to perform a feasibility study to determine its viability.

As the upcoming fifth-generation (5G) and future wireless network is envisioned in areas such as augmented and virtual reality, industrial control, automated driving or flying, robotics, etc, the requirement of supporting ultra-reliable low-latency communications (URLLC) is increasingly urgent than ever. From the channel coding perspective, URLLC requires codewords being transported in finite block-lengths. In this regards, we propose novel encoding algorithms and analyze their performance behaviors for the finite-length Luby transform (LT) codes. 

Luby transform (LT) codes, the first practical realization and the fundamental core of fountain codes, play a key role in the fountain codes family. Recently, researchers show that the performance of LT codes for finite block-lengths can be improved by adding memory into the encoder. However, this work only utilizes one memory, leaving the possibilities of exploiting and how to exploiting more memories an open problem. To explore this unknown, in this work we propose an entire family of memory based LT encoders, and analyze their performance behaviors thoroughly over binary erasure channels and AWGN channels. 

Data mining is an important part of the knowledge discovery process. Data mining helps in finding out patterns across large data sets and establishing relationship through data analysis to solve problems.

Input data sets are often incomplete, i.e., some attribute values are missing. The rough set theory offers mathematical tools to discover patterns hidden in inconsistent and incomplete data. Rough set theory handles inconsistent data by introducing probabilistic approximations. These approximations are combined with an additional parameter (or threshold) called alpha.

The main objective of this project is to compare global and saturated probabilistic approximations using characteristic sets in mining incomplete data. Eight different data sets with 35% missing values were used for experiments. Two different variations of missing values were used, namely, lost values and "do not care" conditions. For rule induction, we implemented the single local probabilistic approximation version of MLEM2. We implemented a rule checker system to verify the accuracy of our generated ruleset by computing the error rate. Along with the rule checker system, the k-fold cross-validation technique was implemented with a value of k as ten to validate the generated rule sets. Finally, a statistical analysis was done for all the approaches using the Friedman test.

Mobile application security concerns safeguarding mobile apps from threats, such as malware, password cracking, social engineering and other attacks. Application security is crucial for every enterprise, as the business can be developed only with the guarantee that the apps are secure from potential threats. Open Web Application Security Project(OWASP) has compiled a list of top 10 mobile risks, and has formulated a set of guidelines for app development and testing. The objective of my project is to analyze the security risks of android application, using the guidelines from OWASP top 10. With the help of suitable tools, analysis is done to identify the vulnerabilities and threats in android applications, on API 4.4.1. Numerous tools have been used as a part of this endeavor, all of them are open source and freely available. As a part of this project, I have also attempted to demonstrate each of the top 10 risks, using individual android applications. A detailed analysis was performed on each of the top 10 mobile risks, and suitable countermeasures for mitigation was provided. A detailed survey of 100 popular applications from the Google Play store was also performed and the risks were categorized into low, medium and high impact, depending on the level of threats.​ 

In real world environments, the speech signals received by our ears are usually a combination of different sounds that include not only the target speech, but also acoustic interference like music, background noise, and competing speakers. This interference has negative effect on speech perception and degrades the performance of speech processing applications such as automatic speech recognition (ASR), speaker identification, and hearing aid devices. One way to solve this problem is using source separation algorithms to separate the desired speech from the interfering sounds. Many source separation algorithms have been proposed to improve the performance of ASR systems and hearing aid devices, but it is still challenging for these systems to work efficiently in noisy and reverberant environments. On the other hand, humans have a remarkable ability to separate desired sounds and listen to a specific talker among noise and other talkers. Inspired by the capabilities of human auditory system, a popular method known as auditory scene analysis (ASA) was proposed to separate different sources in a two stage process of segmentation and grouping. The main goal of source separation in ASA is to estimate time frequency masks that optimally match and separate noise signals from a mixture of speech and noise. In this work, multiple algorithms are proposed to improve upon source separation in noisy and reverberant acoustic environment. First, a simple and novel algorithm is proposed to increase the discriminability between two sound sources by scaling (magnifying) the head-related transfer function of the interfering source. Experimental results from applications of this algorithm show a significant increase in the quality of the recovered target speech. Second, a time frequency masking-based source separation algorithm is proposed that can separate a male speaker from a female speaker in reverberant conditions by using the spatial cues of the source signals. Furthermore, the proposed algorithm has the ability to preserve the location of the sources after separation. Three major aims are proposed for supervised speech separation based on deep neural networks to estimate either the time frequency masks or the clean speech spectrum. Firstly, a novel monaural acoustic feature set based on a gammatone filterbank is presented to be used as  the input of the deep neural network (DNN) based speech separation model, which shows significant improvement in objective speech intelligibility and speech quality in different testing conditions. Secondly, a complementary binaural feature set is proposed to increase the ability of source separation in adverse environment with non-stationary background noise and high reverberation using 2-channel recordings. Experimental results show that the combination of spatial features with this complementary feature set improves significantly the speech intelligibility and speech quality in noisy and reverberant conditions. Thirdly, a novel dilated convolution neural network is proposed to improve the generalization of the monaural supervised speech enhancement model to different untrained speakers, unseen noises and simulated rooms. This model increases the speech intelligibility and speech quality of the recovered speech significantly, while being computationally more efficient and requiring less memory in comparison to other models. In addition, the proposed model is modified with recurrent layers and dilated causal convolution layers for real-time processing. This model is causal which makes it suitable for implementation in hearing aid devices and ASR system, while having fewer trainable parameters and using only information about previous time frames in output prediction. The main goal of the proposed algorithms are to increase the intelligibility and the quality of the recovered speech from noisy and reverberant environments, which has the potential to improve both speech processing applications and signal processing strategies for hearing aid and cochlear implant technology.

The non-ending growth of data traffic resulting from the continuing emergence of Internet applications with high data-rate demands sets huge capacity requirements on optical interconnects and transport networks. This requires the adoption of optical communication technologies that can make the best possible use of the available bandwidths of electronic and electro-optic components to enable data transmission with high spectral efficiency (SE). Therefore, advanced modulation formats are required to be used in conjunction with energy-efficient and cost-effective transceiver schemes, especially for medium- and short-reach applications. Important challenges facing these goals are the stringent requirements on the characteristics of optical components comprising these systems, especially laser sources. Laser phase noise is one of the most important performance-limiting factors in systems with high spectral efficiency. In this research work, we study the effects of the spectral characteristics of laser phase noise on the characterization of lasers and their impact on the performance of digital coherent and self-coherent optical communication schemes. The results of this study show that the commonly-used metric to estimate the impact of laser phase noise on the performance, laser linewidth, is not reliable for all types of lasers. Instead, we propose a Lorentzian-equivalent linewidth as a general characterization parameter for laser phase noise to assess phase noise-related system performance. Practical aspects of determining the proposed parameter are also studied and its accuracy is validated by both numerical and experimental demonstrations. Furthermore, we study the phase noises in quantum-dot mode-locked lasers (QD-MLLs) and assess the feasibility of employing these devices in coherent applications at relatively low symbol rates with high SE. A novel multi-heterodyne scheme for characterizing the phase noise of laser frequency comb sources is also proposed and validated by experimental results with the QD-MLL. This proposed scheme is capable of measuring the differential phase noise between multiple spectral lines instantaneously by a single measurement. Moreover, we also propose an energy-efficient and cost-effective transmission scheme based on direct detection of field-modulated optical signals with advanced modulation formats, allowing for higher SE compared to the current pulse-amplitude modulation schemes. The proposed system combines the Kramers-Kronig self-coherent receiver technique, with the use of QD-MLLs, to transmit multi-channel optical signals using a single diode laser source without the use of the additional RF or optical components required by traditional techniques. Semi-numerical simulations based on experimentally captured waveforms from practical lasers show that the proposed system can be used even for metro scale applications. Finally, we study the properties of phase and intensity noise changes in unmodulated optical signals passing through saturated semiconductor optical amplifiers for intensity noise reduction. We report, for the first time, on the effect of phase noise enhancement that cannot be assessed or observed by traditional linewidth measurements. We demonstrate the impact of this phase noise enhancement on coherent transmission performance by both semi-numerical simulations and experimental validation.

The memory for events is a central component in human cognition, but we have yet to see artificial agents that can demonstrate the same range of event memory capabilities as humans. Some machine learning systems are capable of behaving as if they remember and reason about events, but often times, their behavior is produced by an ad hoc assemblage of opaque statistical algorithms which yield little new insights on the nature of event memory. We propose a novel, psychologically plausible theory of event memory with an accompanying implementation which affords integrated agents the ability to remember events, present details about their past experiences, and reason about future events. We propose to demonstrate such event memory reasoning capabilities in three different experiments. First, we evaluate the fundamental capabilities of our theory to explain different event memory phenomena, such as remembering. Second, we aim to show that our event memory theory provides a unified framework for building intelligent agents that generate explanations of their own behavior and  make inferences about the goals and intentions of other actors. Third, we evaluate whether our event memory theory facilitates cooperative behavior of computational agents in human-robot teams. The proposed work will be completed in December 2020. If our efforts are successful, we believe it will change the way humans interact with autonomous agents. People will better understand why robots, self-driving vehicles, and other agents behave the way they do, and as a result, will know when to trust them. This in turn will speed adoption of autonomous systems not only in military settings but, in everyday life.