Many large-scale distributed Multi-Agent Systems (MAS) exchange information over low- power communication networks. In such scenarios, agents communicate intermittently with each other, often with limited power and over unlicensed spectrum bands that are susceptible to interference, eavesdropping, and Denial-of-Service (DoS) attacks. In this work, we consider a popular low-power, long-range communication protocol known as LoRa. Despite LoRa's high tolerance for noise and interference, it was found vulnerable to interference from particular chirp-type signals. State-of-the-art signal jamming techniques that exploit this property require the knowledge of two sensitive parameters - Bandwidth (BW) and Spreading Factor (SF). However, such information is available only to authenticated parties on the network and not to an eavesdropping adversary. We expose LoRa's vulnerability to DoS attacks by designing an intelligent jammer that surpasses the need for prior knowledge of these parameters. Exploiting a structural pattern in LoRa signals, we propose a Neural Network (NN) implementation for jointly inferring the two parameters by eavesdropping. Through simulation and experimentation, we analyze the detection vulnerability of LoRa for each combination of these parameters at various Signal to Noise Ratio (SNR) values. This work also presents a Radio Frequency (RF) dataset of LoRa signals, which is used to validate our inference model through experimentation. / Master of Science / When many independent devices (or agents) work together in a large system, they often need to communicate with each other. They do so using low-powered networks and often in an intermittent manner. These networks operate on unlicensed radio frequencies, which are open to interference, unwanted snooping, and 'denial-of-service' attacks that could shut down communication. In our study, we focus on a popular low-power, long-distance communication protocol called LoRa. Despite being designed to handle interference and noise well, related literature revealed that LoRa is vulnerable to a specific type of interference caused by 'chirp' signals. Current techniques to jam these signals and disrupt communication require the knowledge of two important factors - bandwidth and spreading factor. Normally, only authorized parties in the network would know these details, not any outsiders looking to interfere. However, we exploit LoRa's vulnerability without knowing these two parameters. By identifying a pattern in LoRa signals, we designed an artificial intelligence model that can determine these two parameters just by listening in. We then ran simulations and conducted experiments to understand how susceptible LoRa is to being detected under various levels of signal strength and noise. We also prepared a dataset of LoRa signals and used this data to confirm the effectiveness of our model.

No description available.

We implement reinforcement learning in the context of wireless communications in two very different settings. In the first setting, we study the use of reinforcement learning in an underwater acoustic communications network to adapt its transmission frequencies to avoid interference and potential malicious jammers. To that effect, we implement a reinforcement learning algorithm called contextual bandits. The harsh environment of an underwater channel provides a challenging problem. The channel may induce multipath and time delays which lead to time-varying, frequency-selective attenuation. These factors are also influenced by the distance between the transmitter and receiver, the subbands the interference is located within, and the power of the transmitter. We show that the agent is effectively able to avoid frequency bands that have degraded channel quality or that contain interference, both of which are dynamic or time-varying . In the second setting, we study the use of reinforcement learning to adapt the modulation and power scheme of a jammer seeking to disrupt a wireless communications system. To achieve this, we make use of a linear contextual bandit to learn to jam the victim system. Prior work has shown that with the use of linear bandits, improved convergence is achieved to jam a single-carrier system using time-domain jamming schemes. However, communications systems today typically employ orthogonal frequency division multiplexing (OFDM) to transmit data, particularly in 4G/5G networks. This work explores the use of linear Thompson Sampling (TS) to jam OFDM-modulated signals. The jammer may select from both time-domain and frequency-domain jamming schemes. We demonstrate that the linear TS algorithm is able to perform better than a traditional reinforcement learning algorithm, upper confidence bound-1 (UCB-1), in terms of maximizing the victim's symbol error rate. We also draw novel insights by observing the action states, to which the reinforcement learning algorithm converges. We then investigate the design and modification of the context vector in the hope of in- creasing overall performance of the bandit, such as decreased learning period and increased symbol error rate caused to the victim. This includes running experiments on particular features and examining how the bandit weights the importance of the features in the context vector. Lastly, we study how to jam an OFDM-modulated signal which employs forward error correction coding. We extend this to leverage reinforcement learning to jam a 5G-based system implementing some aspects of the 5G protocol. This model is then modified to introduce unreliable reward feedback in the form of ACK/NACK observations to the jammer to understand the effect of how imperfect observations of errors can affect the jammer's ability to learn. We gain insights into the convergence time of the jammer and its ability to jam the victim, as well as improvements to the algorithm, and insights into the vulnerabilities of wireless communications for reinforcement learning based jamming. / Master of Science / In this thesis we implement a class of reinforcement learning known as contextual bandits in two different applications of communications systems and jamming. In the first setting, we study the use of reinforcement learning in an underwater acoustic communications network to adapt its transmission frequencies to avoid interference and potential malicious jammers. We show that the agent is effectively able to avoid frequency bands that have degraded channel quality or that contain interference, both of which are dynamic or time-varying. In the second setting, we study the use of reinforcement learning to adapt the jamming type, such as using additive white Gaussian noise, and power scheme of a jammer seeking to disrupt a wireless communications system. To achieve this, we make use of a linear contextual bandit which implies that the contexts that the jammer is able to observe and the sampled probability of each arm has a linear relationship with the reward function. We demonstrate that the linear algorithm is able to outperform a traditional reinforcement learning algorithm in terms of maximizing the victim's symbol error rate. We extend this work by examining the impact of the context feature vector design, LTE/5G-based protocol specifics (such as error correction coding), and imperfect reward feedback information. We gain insights into the convergence time of the jammer and its ability to jam the victim, as well as improvements to the algorithm, and insights into the vulnerabilities of wireless communications for reinforcement learning based jamming.

Many species of bats primarily use echolocation, a type of active sensing wherein bats emit ultrasonic pulses and listen to echoes, for guidance and navigation. Swarms of such bats are a unique type of multi-agent systems that feature bats's echolocation and flight behaviors. In the work of this dissertation, we used bat swarming as an inspiration for multi-agent systems to study various topics which include predation success, active sensing, and collision avoidance. To investigate the predation success, we modeled a group of bats hunting a number of collectively behaving prey. The modeling results demonstrated the benefit of localized grouping of prey in avoiding predation by bats. In the topics regarding active sensing and collision avoidance, we studied individual behavior in swarms as bats could potentially benefit from information sharing while suffering from frequency jamming, i.e., bats having difficulty in distinguishing between self and peers's information. We conducted field experiments in a cave and found that individual bat increased biosonar output as swarm size increased. The experimental finding indicated that individual bat acquired more sensory information in larger swarms even though there could be frequency jamming risk. In a simulation wherein we modeled bats flying through a tunnel, we showed the increasing collision risk in larger swarms for bats either sharing information or flying independently. Thus, we hypothesized that individual bat increased pulse emissions for more sensory information for collision avoidance while possibly taking advantage of information sharing and coping with frequency jamming during swarming. / Ph. D.

An integral part in the field of soft robotics is the ability to tune material stiffness. This adaptability is inspired from the natural ability of organisms to alter their stiffness to perform various tasks. The most common approach to mimic this ability is through granular jamming, where a granular material switches between fluid and solid-like states based on density alterations caused by vacuum pressure. In this thesis, a cuboid composite material is introduced, containing internal cylindrical chambers arranged in distinct matrix configurations (2x2, 3x3, and 4x4). A custom-designed pneumatic system enables precise control over this transition, allowing for selective modulation of stiffness across different regions of the material by applying differing pressures to specific regions of the composite material. This approach not only allows for rapid changes in stiffness, but enables stiffness to be adjusted uniformly throughout the material or localized to specific areas. This approach also allows for predictive modeling of granular composites to better understand its mechanical response under differential pressures. / Master of Science / Soft robotics is a field that mimics the flexibility of living organisms such as octopi, geckos, etc., to create machines that can adapt to various tasks and environments. One of the unique features of these robots is their ability to change how stiff or soft they are, much like an octopus can alter the rigidity of its tentacles when gripping an object. A method called granular jamming is at the heart of this technology. It involves using materials made up of tiny particles, like coffee grounds or sand, that can switch between flowing freely like a liquid and locking together like a solid. This switch is controlled by changing the space between the particles, usually by sucking out air to pack them tightly. The research in this thesis introduces a special type of material designed as a rubber-like cube containing multiple small cylindrical compartments arranged in different patterns, such as 2x2 or 4x4 grids. Each compartment is filled with these unique particle-based materials, in this particular instance, the material is coffee grounds. We use a specially designed air pressure system to selectively adjust the air pressure in these compartments, making the material stiffer or softer as needed. This allows us to control the stiffness with great precision, either uniformly across the whole block or in specific areas. The experiments conducted in this thesis show a clear pattern: the more air pressure is decreased (making it more negative), the stiffer the material becomes. This finding confirms that granular jamming is a promising strategy for rapidly and precisely controlling material stiffness for future soft robotic applications.

Global Navigation Satellite Systems (GNSS) provide critical positioning, navigation, and timing (PNT) services across various sectors. GNSS signals are weak when they reach Earth from Medium Earth Orbit (MEO), making them vulnerable to jamming. The jamming threat has been growing over the past decade, putting critical services at risk. In response, the National Space-Based PNT Advisory Board and the White House advocate for policies and technologies to protect, toughen, and augment GPS for a more resilient PNT. Time-sequential estimation improves navigation accuracy and allows for the augmentation of GNSS with other difficult-to-interfere sensors. Safety-critical navigation applications (e.g., GNSS/INS-based aircraft localization) that use time-sequential estimation require high-integrity measurement error time correlation models to compute estimation error bounds. In response, two new methods to identify high-integrity measurement error time correlation models from experimental data are developed and evaluated in this thesis. As opposed to bounding autocorrelation functions in the time domain and power spectra in the frequency domain, methods proposed in this thesis use bounding of lagged product distributions in the time domain and scaled periodogram distributions in the frequency domain. The proposed methods can identify tight-bounding models from empirical data, resulting in tighter estimation error bounds. The sample distributions are bound using theoretical First-order Gauss-Markov process (FOGMP) model distributions derived in this thesis. FOGMP models provide means to account for error time correlation while being easily incorporated into linear estimators. The two methods were evaluated using simulated and experimental GPS measurement error data collected in a mild multipath environment. To protect and alert GNSS end users of jamming, this thesis proposes and evaluates an autonomous algorithm to detect jamming using publicly available data from large receiver networks. The algorithm uses carrier-to-noise ratio (C/N0)-based jamming detectors that are optimal, self-calibrating, receiver-independent, and while adhering to a predefined false alert rate. This algorithm was tested using data from networks with hundreds of receivers, revealing patterns indicative of intentional interference, which provided an opportunity to validate the detector. This validation activity, described in this thesis, consists of designing a portable hardware setup, deriving an optimal power-based jamming monitor for independent detection, and time-frequency analysis of wideband RF (WBRF) data collected during jamming events. The analysis of the WBRF data from a genuine jamming event detected while driving on I-25 in Denver, Colorado, USA, revealed power variations resembling a personal privacy device (PPD), validating the C/N0 detector's result. Finally, this thesis investigates the cause of recurring false alerts in our power-based jamming detectors. These false alerts are caused by a few short pulses of power increases, which other researchers also observe. The time-frequency analysis of signals from the pulses revealed binary data encoded using frequency shift keying (FSK) in the GPS L1 band. Various experiments confirmed the signals are not aliases of out-of-band signals. A survey of similar encoded messages identified the source as car key fobs and other devices transmitting at 315 MHz, nowhere near the GPS L1 band, with an unattenuated 5$^{th}$ harmonic in the GPS L1 band. The RF emission regulations were analyzed to identify mitigation. / Doctor of Philosophy / Global Navigation Satellite Systems (GNSS) have become integral to modern-day life. Many essential services rely on GNSS-provided Positioning, Navigation, and Timing (PNT) services; power grids rely on accurate GNSS-provides timing for synchronization; stock markets use them for time-stamping trades; aircraft and ships use GNSS to correct accumulated position errors regularly; to name a few. In addition, the availability of cheap and accessible PNT services combined with mobile internet spawned new service sectors through mobile applications. A 2019 study published by the National Institute of Standards and Technology (NIST) estimates that GPS has generated $1.4 trillion in U.S. economic benefits since the system became available in the 1980s. With the wide adoption of GNSS services comes new motives for interference. These motives can range from delivery workers and truck drivers trying to hide their location from their employers to something more nefarious, such as criminals trying to evade law enforcement surveillance. GNSS jamming is a type of interference in which the attacker drowns out the faint GNSS signals, broadcast from medium Earth Orbit (MEO) at 20,000 km, with a powerful RF transmitter. Some commonly used devices are transmitters are cheaply available for as low as $10 on Amazon, known as personal privacy devices (PPDs). Another source of jamming comes from militaries in conflict zones overseas, jamming GNSS signals over large areas of a country or a city. However, two major incidents in the US have disrupted air traffic over busy airspace, such as in Denver and Dallas. This threat of GNSS interference has grown over the past decade and is only getting worse. The White House and other organizations advocate for policies for a more resilient PNT; to protect, toughen, and augment GNSS. % This thesis contributes to protecting GNSS frequencies through autonomous algorithms that process publicly available signal quality data from large receiver networks for jamming detection. This autonomous algorithm uses detectors that are self-calibrating and optimal, i.e., minimizing the probability of missed detection while targeting a predefined false alert probability. Several jamming event patterns consistent with intentional interference were detected using this algorithm. The signal-quality-based detectors were validated using an independent power-based optimal jamming detector derived in this thesis. Spurious recurring false alerts triggered the power detector. An investigation described in the thesis discovered that car key fobs and other devices emit RF energy in restricted GPS frequencies. Based on the analysis of FCC regulation for RF transmitters, mitigation is proposed for power-based jamming detectors to prevent false alarms. Time-sequential estimation improves navigation accuracy and allows for the augmentation of GNSS with other difficult-to-interfered sensors such as IMU or LIDAR. Safety-critical navigation applications can benefit from time-sequential estimation, but they require high-integrity measurement error time correlation models to compute bounds on positioning errors. Two new methods to derive high-integrity measurement error time correlation models from experimental data are developed and evaluated in this thesis. These methods can derive tighter bounding models compared to the existing methods, reducing the uncertainty in position estimates. The two methods were implemented and evaluated using simulated and experimental GPS measurement error data collected in a mild multipath environment.

Esta dissertação apresenta uma reflexão sobre as interseções entre práticas artísticas e ativismo contemporâneo, especialmente nas décadas de 1990 e 2000. A partir de diferentes contextos, o estudo investiga os conceitos e objetivos de uma arte coletiva e engajada socialmente, considerando seus modos de experimentação estética e expressão política. Utilizando-se de entrevistas, manifestos, textos críticos, reportagens e documentos como fotografias, vídeos e filmes, a dissertação apresenta no primeiro capítulo um histórico detalhado sobre as diversas concatenações entre arte, ativismo político e produção coletiva no século XX. No segundo capítulo, este trabalho analisa a formulação de uma \"estética anti-corporativa\", baseada em táticas intervencionistas criadas por artistas e coletivos radicados nos EEUU, Espanha, França, Canadá, Austrália e Brasil. Seus projetos envolvem instalações artísticas com experimentos biológicos, mídia tática, cartografias, protestos contra a globalização capitalista, performances e Culture Jamming. O terceiro capítulo apresenta um estudo sobre o coletivismo artístico no Brasil e algumas de suas estratégias de ação, como intervenções urbanas, circuitos alternativos de produção e de distribuição, projetos com comunidades específicas e colaborações com movimentos sociais. Além disso, o texto faz uma breve reflexão sobre a atitude e o impacto destes grupos sobre o sistema de arte, caracterizado pelo apoio institucional de museus, galerias, mostras internacionais, críticos, curadores e patrocínio corporativo / This dissertation presents a reflection about the intersections between artistic practices and contemporary activism, especially in the decades of 1990 and 2000. From different contexts, teh study investigates the concepts and objectives of a collective art, socially engaged, considering their modes of aesthetic experimentation and political expression. Utilizing interviews, manifests, critical texts, newsprints and documents as photographies, videos and movies, the dissertation presnts in the first chapter a historical account about the concatenations between art, political activism and collective production in the twentieth century. In the second chapter, this work analyses a formulation of an \"anti-corporate aesthetics\", based in interventionist tactics created by artists and collectives in USA, Spain, France, Canada, Australia and Brazil. Their projects involve artistic installations with biological experiments, tactical media cartographies, protests against capitalist globalization, performances and culture jamming. The third chapter presents a study about the artistic collectivism in Brazil and some of their strategies of action, as urban interventions, alternative circuits of production and distribution, projects with specific communities and collaborations with social movements. Besides, the next makes a brief reflection about the attitude and impact of these groups in the art system, characterized by institutional support of museums, galleries, international exhibitions, art critics, curators and corporate sponsorship

Esta dissertação apresenta uma reflexão sobre as interseções entre práticas artísticas e ativismo contemporâneo, especialmente nas décadas de 1990 e 2000. A partir de diferentes contextos, o estudo investiga os conceitos e objetivos de uma arte coletiva e engajada socialmente, considerando seus modos de experimentação estética e expressão política. Utilizando-se de entrevistas, manifestos, textos críticos, reportagens e documentos como fotografias, vídeos e filmes, a dissertação apresenta no primeiro capítulo um histórico detalhado sobre as diversas concatenações entre arte, ativismo político e produção coletiva no século XX. No segundo capítulo, este trabalho analisa a formulação de uma \"estética anti-corporativa\", baseada em táticas intervencionistas criadas por artistas e coletivos radicados nos EEUU, Espanha, França, Canadá, Austrália e Brasil. Seus projetos envolvem instalações artísticas com experimentos biológicos, mídia tática, cartografias, protestos contra a globalização capitalista, performances e Culture Jamming. O terceiro capítulo apresenta um estudo sobre o coletivismo artístico no Brasil e algumas de suas estratégias de ação, como intervenções urbanas, circuitos alternativos de produção e de distribuição, projetos com comunidades específicas e colaborações com movimentos sociais. Além disso, o texto faz uma breve reflexão sobre a atitude e o impacto destes grupos sobre o sistema de arte, caracterizado pelo apoio institucional de museus, galerias, mostras internacionais, críticos, curadores e patrocínio corporativo / This dissertation presents a reflection about the intersections between artistic practices and contemporary activism, especially in the decades of 1990 and 2000. From different contexts, teh study investigates the concepts and objectives of a collective art, socially engaged, considering their modes of aesthetic experimentation and political expression. Utilizing interviews, manifests, critical texts, newsprints and documents as photographies, videos and movies, the dissertation presnts in the first chapter a historical account about the concatenations between art, political activism and collective production in the twentieth century. In the second chapter, this work analyses a formulation of an \"anti-corporate aesthetics\", based in interventionist tactics created by artists and collectives in USA, Spain, France, Canada, Australia and Brazil. Their projects involve artistic installations with biological experiments, tactical media cartographies, protests against capitalist globalization, performances and culture jamming. The third chapter presents a study about the artistic collectivism in Brazil and some of their strategies of action, as urban interventions, alternative circuits of production and distribution, projects with specific communities and collaborations with social movements. Besides, the next makes a brief reflection about the attitude and impact of these groups in the art system, characterized by institutional support of museums, galleries, international exhibitions, art critics, curators and corporate sponsorship

In this report metrics for jamming resistance and radar performance of waveform sets are described and developed, and different sets of waveforms are optimized, evaluated and compared. It is shown that without additional processing or PRI jitter, waveform sets can reach jamming resistance a few dB worse than what is provided by PRI jitter alone, and together with PRI jitter a few dB better. Waveforms with better jamming resistance tend to have worse range sidelobes and Doppler tolerance, but show less structure in their spectrograms, suggesting better LPI properties. The Doppler tolerance metric is new, as well as the comparative analysis of waveform sets on multiple metrics including jamming resistance. / Radar är fundamentalt i modern krigsföring. Med en radar kan man avfyra vapen från säkra avstånd och med precision mäta in mål. En radarstörare har som mål att förhindra en radar från att mäta in sitt mål. Då radarn fungerar genom att sända ut specifikt modulerade radiovågspulser och lyssna efter ekot från omgivningen kan störaren förhindra detta genom att antingen sända mycket starkt brus, eller genom att sända radiovågspulser med samma specifika modulation. Den senare metoden kallas för DRFM-störning, där förkortningen står för Digitalt RadioFrekvens-Minne, vilket antyder att störaren kan minnas radarns modulation och själv använda den. Om radarn använder en ny modulation (eng: waveform) för varje puls kan störaren inte använda modulationen den minns från förra pulsen utan måste vänta på att nästa puls träffar den innan den kan repetera pulsen, vilket begränsar dess störförmåga. Denna rapport tänker sig att radarn har en begränsad uppsättning av modulationer att byta mellan, och undersöker olika sådana uppsättningar och bedömer och jämför dem på olika mått av radarprestanda och störtålighet. Radioprestandamåtten inkluderar hur mycket förstärkning och hur fin upplösning man får av modulationen, hur väl modulationen kan hantera mycket snabba mål, och hur stora "sidolober" som uppstår runt starka mål. Sidolobsfenomenet är jämförbart med det optiska fenomenet där små men ljusstarka saker på natten kan se ut att ha en ljus halo eller ljusa utstrålningar runt sig. Störtålighetsmåtten kvantifierar hur distinkta de olika modulationerna i radarns uppsättning är, och på så vis hur väl radarn kan urskilja en modulation från de andra, tillsammans med hur liten sannolikheten är att störaren lyckas välja just den modulation vi kommer använda till nästa puls. Resultaten visar att metoden av modulationsbyten kan ge nästan lika stor störtålighet som en välkänd metod, PRI-jitter, ger själv och något högre i kombination med den metoden. Bättre störtålighet visas gå hand i hand med sämre mått på radarprestanda, men mindre strukturerade spektrogram vilket antyder att de kan vara svårare att upptäckas av radarspanare. Försämringen i måtten på radarprestanda innebär inte nödvändigtvis en lika stor försämring i faktisk radarprestanda, då sidoloberna tar an en brusartad karaktär vilket leder till praktiska fördelar gentemot de vanliga fixa sidoloberna.

The ability of tissues and cells to move and rearrange is central to a broad range of diverse biological processes such as tissue remodeling and rearrangement in embryogenesis, cell migration in wound healing, or cancer progression. These processes are linked to a solidlike to fluid-like transition, also known as unjamming transition, a not rigorously defined framework that describes switching between a stable, resting state and an active, moving state. Various mechanisms, that is, proliferation and motility, are critical drivers for the (un) jamming transition on the cellular scale. However, beyond the scope of these fundamental mechanisms of cells, a unifying understanding remains to be established. During embryogenesis, the proliferation rate of cells is high, and the number density is continuously increasing, which indicates number-density-driven jamming. In contrast, cells have to unjam in tissues that are already densely packed during tumor progression, pointing toward a shape-driven unjamming transition. Here, we review recent investigations of jamming transitions during embryogenesis and cancer progression and pursue the question of how they might be interlinked. We discuss the role of density and shape during the jamming transition and the different biological factors driving it.