Reactive jammers pose a severe security threat to robotic-swarm networks by selectively disrupting inter-agent communications and undermining formation integrity and mission success. Conventional countermeasures such as fixed power control or static channel hopping are largely ineffective against such adaptive adversaries. This paper presents a multi-agent reinforcement learning (MARL) framework based on the QMIX algorithm to improve the resilience of swarm communications under reactive jamming. We consider a network of multiple transmitter–receiver pairs sharing channels while a reactive jammer with Markovian threshold dynamics senses aggregate power and reacts accordingly. Each agent jointly selects transmit frequency (channel) and power, and QMIX learns a centralized but factorizable action-value function that enables coordinated yet decentralized execution. We benchmark QMIX against a genie-aided optimal policy in a no–channel-reuse setting, and against local Upper Confidence Bound (UCB) and a stateless reactive policy in a more general fading regime with channel reuse enabled. Simulation results show that QMIX rapidly converges to cooperative policies that nearly match the genie-aided bound, while achieving higher throughput and lower jamming incidence than the baselines, thereby demonstrating MARL’s effectiveness for securing autonomous swarms in contested environments.

We present a retrieval-augmented question answering framework for 5G/6G networks, where the Open Radio Access Network (O-RAN) has become central to disaggregated, virtualized, and AI-driven wireless systems. While O-RAN enables multi-vendor interoperability and cloud-native deployments, its fast-changing specifications and interfaces pose major challenges for researchers and practitioners. Manual navigation of these complex documents is labor-intensive and error-prone, slowing system design, integration, and deployment. To address this challenge, we adopt Contextual Retrieval-Augmented Generation (Contextual RAG), a strategy in which candidate answer choices guide document retrieval and chunk-specific context to improve large language model (LLM) performance. This improvement over traditional RAG achieves more targeted and context-aware retrieval, which improves the relevance of documents passed to the LLM, particularly when the query alone lacks sufficient context for accurate grounding. Our framework is designed for dynamic domains where data evolves rapidly and models must be continuously updated or redeployed, all without requiring LLM fine-tuning. We evaluate this framework using the ORAN-Benchmark-13K dataset, and compare three LLMs, namely, Llama3.2, Qwen2.5-7B, and Qwen3.0-4B, across both Direct Question Answering (Direct Q&A) and Chain-of-Thought (CoT) prompting strategies. We show that Contextual RAG consistently improves accuracy over standard RAG and base prompting, while maintaining competitive runtime and CO2emissions. These results highlight the potential of Contextual RAG to serve as a scalable and effective solution for domain-specific Q&A in O-RAN and broader 5G/6G environments, enabling more accurate interpretation of evolving standards while preserving efficiency and sustainability.

LoRa enables long-range, energy-efficient communication for Internet of Things (IoT) applications, making it an essential technology for low-power wide-area networks (LPWANs). However, its security remains a concern, particularly when reliable device identification and authentication are critical. This article addresses these challenges using deep learning (DL) techniques to perform two key tasks: 1) identifying LoRa devices and 2) distinguishing between legitimate signals and rogue signals [generated by kernel density estimation (KDE)]. By training deep neural networks (DNNs) on real LoRa signal data, the study examines the susceptibility of separate models for each task as well as a shared multitask model to untargeted and targeted adversarial attacks, generated with the fast gradient sign method (FGSM). To counter these attacks, a defense strategy using adversarial training is proposed to enhance model robustness. The results highlight vulnerabilities in LoRa security and emphasize the importance of fortifying IoT systems against such sophisticated threats.

In this paper, we address task-oriented (or goal-oriented) communications where an encoder at the transmitter learns compressed latent representations of data, which are then transmitted over a wireless channel. At the receiver, a decoder performs a machine learning task, specifically for classifying the received signals. The deep neural networks corresponding to the encoder-decoder pair are jointly trained, taking both channel and data characteristics into account. Our objective is to achieve high accuracy in completing the underlying task while minimizing the number of channel uses determined by the encoder's output size. To this end, we propose a multi-round, multi-task learning (MRMTL) approach for the dynamic update of channel uses in multi-round transmissions. The transmitter incrementally sends an increasing number of encoded samples over the channel based on the feedback from the receiver, and the receiver utilizes the signals from a previous round to enhance the task performance, rather than only considering the latest transmission. This approach employs multi-task learning to jointly optimize accuracy across varying number of channel uses, treating each configuration as a distinct task. By evaluating the confidence of the receiver in task decisions, MRMTL decides on whether to allocate additional channel uses in multiple rounds. We characterize both the accuracy and the delay (total number of channel uses) of MRMTL, demonstrating that it achieves the accuracy close to that of conventional methods requiring large numbers of channel uses, but with reduced delay by incorporating signals from a prior round. We consider the CIFAR-10 dataset, convolutional neural network architectures, and AWGN and Rayleigh channel models for performance evaluation. Our results show that MRMTL significantly improves the efficiency of task-oriented communications, balancing accuracy and latency effectively.

We consider the communication of time-sensitive information in NextG spectrum sharing where a deep learning-based classifier is used to identify transmission attempts. While the transmitter seeks for opportunities to use the spectrum without causing interference to an incumbent user, an adversary uses another deep learning classifier to detect and jam the signals, subject to an average power budget. We consider timeliness objectives of NextG communications and study the Age of Information (AoI) under different scenarios of spectrum sharing and jamming, analyzing the effect of transmit control, transmit probability, and channel utilization subject to wireless channel and jamming effects. The resulting signal-to-noise-plus-interference (SINR) determines the success of spectrum sharing, but also affects the accuracy of the adversary’s detection, making it more likely for the jammer to successfully identify and jam the communication. Our results illustrate the benefits of spectrum sharing for anti-jamming by exemplifying how a limited-power adversary is motivated to decrease its jamming power as the channel occupancy rises in NextG spectrum sharing with timeliness objectives.

An over-the-air membership inference attack (MIA) is presented to leak private information from a wireless signal classifier. Machine learning (ML) provides powerful means to classify wireless signals, e.g., for PHY-layer authentication. As an adversarial machine learning attack, the MIA infers whether a signal of interest has been used in the training data of a target classifier. This private information incorporates waveform, channel, and device characteristics, and if leaked, can be exploited by an adversary to identify vulnerabilities of the underlying ML model (e.g., to infiltrate the PHY-layer authentication). One challenge for the over-the-air MIA is that the received signals and consequently the RF fingerprints at the adversary and the intended receiver differ due to the discrepancy in channel conditions. Therefore, the adversary first builds a surrogate classifier by observing the spectrum and then launches the black-box MIA on this classifier. The MIA results (based on both simulations and over-the-air software-defined radio (SDR) experiments) show that the adversary can reliably infer signals (and potentially the radio and channel information) used to build the target classifier. Therefore, a proactive defense is developed against the MIA by building a shadow MIA model and fooling the adversary. This defense can successfully reduce the MIA accuracy and prevent information leakage from the wireless signal classifier. Moreover, this defense does not reduce the accuracy of signal classification.

WiFi 5/6 relies on a key feature, Multi-User Multiple-In-Multiple-Out (MU-MIMO), to offer high-volume network throughput and spectrum efficiency. MU-MIMO uses a user selection algorithm, based on each user's channel state information (CSI), to schedule transmission opportunities for a group of users to maximize the service quality and efficiency. In this paper, we discover that such algorithm creates a subtle attack surface for attackers to subvert user selection in MU-MIMO, causing severe disruptions in today's wireless networks. We develop a system, named MU-MIMO user selection strategy inference and subversion (MUSTER), to systematically study the attack strategies and further to seek efficient mitigation. MUSTER is designed to include two major modules: (i) strategy inference, which leverages a new neural group-learning strategy named MC-grouping via combining Recurrent Neural Network (RNN) and Monte Carlo Tree Search (MCTS) to reverseengineer a user selection algorithm, and (ii) user selection subversion, which proactively fabricates CSI to manipulate user selection results for disruption. Experimental evaluation shows that MUSTER achieves a high accuracy rate around 98.6% in user selection prediction and effectively launches the attacks to disrupt the network performance. Finally, we create a Reciprocal Consistency Checking technique to defend against the proposed attacks to secure MU-MIMO user selection.

In this paper, we first present a deep reinforcement learning (deep RL) framework for network slicing in a dynamic environment. We propose three different deep RL algorithms, namely actor-critic, deep Q learning (DQN), and soft DQN, to select slices from the best recorded subset which is updated over time to adapt to the dynamic environment. We evaluate the performances of the proposed deep RL agents for network slicing and provide comparisons. Subsequently, we design intelligent jammers also as deep RL agents that significantly degrade the user's sum reward. Finally, we propose effective defensive measures to mitigate jamming attacks by determining the proper time instants to retrain the network slicing policy. Via simulations, we quantify the improvements in the performance with the defensive retraining.

By moving from massive antennas to antenna surfaces for software-defined wireless systems, the reconfigurable intelligent surfaces (RISs) rely on arrays of unit cells to control the scattering and reflection profiles of signals, mitigating the propagation loss and multipath attenuation, and thereby improving the coverage and spectral efficiency. In this paper, covert communication is considered in the presence of the RIS. While there is an ongoing transmission boosted by the RIS, both the intended receiver and an eavesdropper individually try to detect this transmission using their own deep neural network (DNN) classifiers. The RIS interaction vector is designed by balancing two (potentially conflicting) objectives of focusing the transmitted signal to the receiver and keeping the transmitted signal away from the eavesdropper. To boost covert communications, adversarial perturbations are added to signals at the transmitter to fool the eavesdropper’s classifier while keeping the effect on the receiver low. Results from different network topologies show that adversarial perturbation and RIS interaction vector can be jointly designed to effectively increase the signal detection accuracy at the receiver while reducing the detection accuracy at the eavesdropper to enable covert communications.

Unmanned aerial vehicles (UAVs) are expected to be an integral part of wireless networks, and determining collision-free trajectory in multi-UAV noncooperative scenarios while collecting data from distributed Internet of Things (IoT) nodes is a challenging task. In this article, we consider a path-planning optimization problem to maximize the collected data from multiple IoT nodes under realistic constraints. The considered multi-UAV noncooperative scenarios involve a random number of other UAVs in addition to the typical UAV, and UAVs do not communicate or share information among each other. We translate the problem into a Markov decision process (MDP) with parameterized states, permissible actions, and detailed reward functions. Dueling double deep$Q$-network (D3QN) is proposed to learn the decision-making policy for the typical UAV, without any prior knowledge of the environment (e.g., channel propagation model and locations of the obstacles) and other UAVs (e.g., their missions, movements, and policies). The proposed algorithm can adapt to various missions in various scenarios, e.g., different numbers and positions of IoT nodes, different amount of data to be collected, and different numbers and positions of other UAVs. Numerical results demonstrate that real-time navigation can be efficiently performed with high success rate, high data collection rate, and low collision rate.

Defense strategies have been well studied to combat Byzantine attacks that aim to disrupt cooperative spectrum sensing by sending falsified versions of spectrum sensing data to a fusion center. However, existing studies usually assume network or attackers as passive entities, e.g., assuming the prior knowledge of attacks is known or fixed. In practice, attackers can actively adopt arbitrary behaviors and avoid pre-assumed patterns or assumptions used by defense strategies. In this paper, we revisit this security vulnerability as an adversarial machine learning problem and propose a novel learning-empowered attack framework named Learning-Evaluation-Beating (LEB) to mislead the fusion center. Based on the black-box nature of the fusion center in cooperative spectrum sensing, our new perspective is to make the adversarial use of machine learning to construct a surrogate model of the fusion center's decision model. We propose a generic algorithm to create malicious sensing data using this surrogate model. Our real-world experiments show that the LEB attack is effective to beat a wide range of existing defense strategies with an up to 82 percent of success ratio. Given the gap between the proposed LEB attack and existing defenses, we introduce a non-invasive method named as influence-limiting defense, which can coexist with existing defenses to defend against LEB attack or other similar attacks. We show that this defense is highly effective and reduces the overall disruption ratio of LEB attack by up to 80 percent.

This paper presents channel-aware adversarial attacks against deep learning-based wireless signal classifiers. There is a transmitter that transmits signals with different modulation types. A deep neural network is used at each receiver to classify its over-the-air received signals to modulation types. In the meantime, an adversary transmits an adversarial perturbation (subject to a power budget) to fool receivers into making errors in classifying signals that are received as superpositions of transmitted signals and adversarial perturbations. First, these evasion attacks are shown to fail when channels are not considered in designing adversarial perturbations. Then, realistic attacks are presented by considering channel effects from the adversary to each receiver. After showing that a channel-aware attack is selective (i.e., it affects only the receiver whose channel is considered in the perturbation design), a broadcast adversarial attack is presented by crafting a common adversarial perturbation to simultaneously fool classifiers at different receivers. The major vulnerability of modulation classifiers to over-the-air adversarial attacks is shown by accounting for different levels of information available about the channel, the transmitter input, and the classifier model. Finally, a certified defense based on randomized smoothing that augments training data with noise is introduced to make the modulation classifier robust to adversarial perturbations.

Unmanned aerial vehicles (UAVs) are expected to be an integral part of wireless networks, and determining collision-free trajectories in multi-UAV non-cooperative scenarios is a challenging task. In this paper, we consider a path planning optimization problem to maximize the collected data from multiple Internet of Things (IoT) nodes under realistic constraints. The considered multi-UAV non-cooperative scenarios involve random number of other UAVs in addition to the typical UAV, and UAVs do not communicate with each other. We translate the problem into an Markov decision process (MDP). Dueling double deep Q-network (D3QN) is proposed to learn the decision making policy for the typical UAV, without any prior knowledge of the environment (e.g., channel propagation model and locations of the obstacles) and other UAVs (e.g., their missions, movements, and policies). Numerical results demonstrate that real-time navigation can be efficiently performed with high success rate, high data collection rate, and low collision rate.

We consider a wireless communication network with an adaptive scheme to select the number of packets to be admitted and encoded for each transmission, and characterize the information timeliness. For a network of erasure channels and discrete time, we provide closed form expressions for the Average and Peak Age of Information (AoI) as functions of admission control and adaptive coding parameters, the feedback delay, and the maximum feasible end-to-end rate that depends on channel conditions and network topology. These new results guide the system design for robust improvements of the AoI when transmitting time sensitive information in the presence of topology and channel changes. We illustrate the benefits of using adaptive packet coding to improve information timeliness by characterizing the network performance with respect to the AoI along with its relationship to throughput (rate of successfully decoded packets at the destination) and per-packet delay. We show that significant AoI performance gains can be obtained in comparison to the uncoded case, and that these gains are robust to network variations as channel conditions and network topology change.

We consider a wireless communication system that consists of a background emitter, a transmitter, and an adversary. The transmitter is equipped with a deep neural network (DNN) classifier for detecting the ongoing transmissions from the background emitter and transmits a signal if the spectrum is idle. Concurrently, the adversary trains its own DNN classifier as the surrogate model by observing the spectrum to detect the ongoing transmissions of the background emitter and generate adversarial attacks to fool the transmitter into misclassifying the channel as idle. This surrogate model may differ from the transmitter’s classifier significantly because the adversary and the transmitter experience different channels from the background emitter and therefore their classifiers are trained with different distributions of inputs. This system model may represent a setting where the background emitter is a primary user, the transmitter is a secondary user, and the adversary is trying to fool the secondary user to transmit even though the channel is occupied by the primary user. We consider different topologies to investigate how different surrogate models that are trained by the adversary (depending on the differences in channel effects experienced by the adversary) affect the performance of the adversarial attack. The simulation results show that the surrogate models that are trained with different distributions of channel-induced inputs severely limit the attack performance and indicate that the transferability of adversarial attacks is neither readily available nor straightforward to achieve since surrogate models for wireless applications may significantly differ from the target model depending on channel effects.

We present the DeepWiFi protocol, which hardens the baseline WiFi (IEEE 802.11ac) with deep learning and sustains high throughput by mitigating out-of-network interference. DeepWiFi is interoperable with baseline WiFi and builds upon the existing WiFi's PHY transceiver chain without changing the MAC frame format. Users run DeepWiFi for: i) RF front end processing; ii) spectrum sensing and signal classification; iii) signal authentication; iv) channel selection and access; v) power control; vi) modulation and coding scheme (MCS) adaptation; and vii) routing. DeepWiFi mitigates the effects of probabilistic, sensing-based, and adaptive jammers. RF front end processing applies a deep learning-based autoencoder to extract spectrum-representative features. Then a deep neural network is trained to classify waveforms reliably as idle, WiFi, or jammer. Utilizing channel labels, users effectively access idle or jammed channels, while avoiding interference with legitimate WiFi transmissions (authenticated by machine learning-based RF fingerprinting) resulting in higher throughput. Users optimize their transmit power for low probability of intercept/detection and their MCS to maximize link rates used by backpressure algorithm for routing. Supported by embedded platform implementation, DeepWiFi provides major throughput gains compared to baseline WiFi and another jamming-resistant protocol, especially when channels are likely to be jammed and the signal-to-interference-plus-noise-ratio is low.

An adversarial deep learning approach is presented to launch over-the-air spectrum poisoning attacks. A transmitter applies deep learning on its spectrum sensing results to predict idle time slots for data transmission. In the meantime, an adversary learns the transmitter's behavior (exploratory attack) by building another deep neural network to predict when transmissions will succeed. The adversary falsifies (poisons) the transmitter's spectrum sensing data over the air by transmitting during the short spectrum sensing period of the transmitter. Depending on whether the transmitter uses the sensing results as test data to make transmit decisions or as training data to retrain its deep neural network, either it is fooled into making incorrect decisions (evasion attack) or the transmitter's algorithm is retrained incorrectly for future decisions (causative attack). Both attacks are energy efficient and hard to detect (stealth) compared to jamming the long data transmission period, and substantially reduce the throughput. A dynamic defense is designed for the transmitter that deliberately makes a small number of incorrect transmissions (selected by the confidence score on channel classification) to manipulate the adversary's training data. This defense effectively fools the adversary (if any) and helps the transmitter sustain its throughput with or without an adversary present.

Machine learning finds rich applications in Internet of Things (IoT) networks such as information retrieval, traffic management, spectrum sensing, and signal authentication. While there is a surge of interest to understand the security issues of machine learning, their implications have not been understood yet for wireless applications such as those in IoT systems that are susceptible to various attacks due the open and broadcast nature of wireless communications. To support IoT systems with heterogeneous devices of different priorities, we present new techniques built upon adversarial machine learning and apply them to three types of over-the-air (OTA) wireless attacks, namely denial of service (DoS) attack in terms of jamming, spectrum poisoning attack, and priority violation attack. By observing the spectrum, the adversary starts with an exploratory attack to infer the channel access algorithm of an IoT transmitter by building a deep neural network classifier that predicts the transmission outcomes. Based on these prediction results, the wireless attack continues to either jam data transmissions or manipulate sensing results over the air (by transmitting during the sensing phase) to fool the transmitter into making wrong transmit decisions in the test phase (corresponding to an evasion attack). When the IoT transmitter collects sensing results as training data to retrain its channel access algorithm, the adversary launches a causative attack to manipulate the input data to the transmitter over the air. We show that these attacks with different levels of energy consumption and stealthiness lead to significant loss in throughput and success ratio in wireless communications for IoT systems. Then we introduce a defense mechanism that systematically increases the uncertainty of the adversary at the inference stage and improves the performance. Results provide new insights on how to attack and defend IoT networks using deep learning.

Social connections among network nodes have been well investigated as an additional opportunity in network design (e.g., in routing strategies and trusted networking). This paper presents a paradigm shift that explores the design and performance analysis of combining social links jointly with communication links for message delivery in wireless networks. In a combined multi-layer social and communication network, communication links are based on conventional wireless technologies (e.g., WiFi, Bluetooth) and social links are overlaid over a communication infrastructure (e.g., cellular network) that provides an alternative way for data transmission. The goal is to characterize the performance analytically when routing is designed by combining social and communication links. A distance discretization technique is applied to model the reliability and delay of message delivery. The analytical foundation is developed to analyze the end-to-end delay and success probability under various effects of persistent transmission, potential error in distance estimation, and mobility. Systematic routing strategies that employ network inference are then designed to improve the performance in different aspects, such as delivery delay, delivery success probability, and energy-saving. A network emulation testbed is implemented with actual radios and real-world social network datasets to measure the performance of a heterogeneous network with social and communication links. The results in this paper show that the integration of social links in wireless network routing as a multi-layer design leads to substantial performance improvement for delay and reliability of message delivery.

A novel approach of training data augmentation and domain adaptation is presented to support machine learning applications for cognitive radio. Machine learning provides effective tools to automate cognitive radio functionalities by reliably extracting and learning intrinsic spectrum dynamics. However, there are two important challenges to overcome, in order to fully utilize the machine learning benefits with cognitive radios. First, machine learning requires significant amount of truthed data to capture complex channel and emitter characteristics, and train the underlying algorithm (e.g., a classifier). Second, the training data that has been identified for one spectrum environment cannot be used for another one (e.g., after channel and emitter conditions change). To address these challenges, a generative adversarial network (GAN) with deep learning structures is used to 1) generate additional synthetic training data to improve classifier accuracy, and 2) adapt training data to spectrum dynamics. This approach is applied to spectrum sensing by assuming only limited training data without knowledge of spectrum statistics. Machine learning classifiers are trained to detect signals using either limited, augmented or adapted training data. Results show that training data augmentation increases the classifier accuracy significantly and this increase is sustained with domain adaptation while spectrum conditions change.

An adaptive rate optimization solution is presented to utilize the benefits of multiuser multiple-input multiple-output (MU-MIMO) communications in a wireless network with distributed and decentralized control that adapts to dynamic channel, interference, and traffic conditions. First, the ergodic sum rates of MIMO multiple access channel (MAC) and interference channel (IC) configurations are determined by jointly integrating the error and overhead effects due to channel estimation (training) and feedback into the rate optimization. Then, a distributed channel access protocol that leverages local information without any centralized scheduler is developed to select and activate MU-MIMO configurations with the maximum achievable sum rates depending on channel, interference, and traffic conditions. In a mobile ad hoc network (MANET), MU-MIMO is shown to provide major gains in network throughput (after accounting for the control message overhead) compared with single antenna and point-to-point (P2P) MIMO communications.

This paper presents a novel system, synthetic high-fidelity social media data generator (SHIELD), for generating the synthetic social media data. SHIELD jointly generates time-varying, directed and weighted interaction graph structures and topic-driven text features similar to the input social media data. A synthetic interaction graph is generated by a social network model to minimize the distance to real graph and is enhanced by adding various patterns, such as anomalies and information cascades, interaction types, and temporal dynamics. A synthetic text generator based on the$n$-gram Markov model is trained under each topic identified by topic modeling. Synthetic text and graph structures are combined through the assignment of synthetic social media entities. Extensive performance evaluation via a graph and text analysis is provided to demonstrate the statistical fidelity of large-scale synthetic data generated by SHIELD. A data evaluation exercise with human participants is executed to identify how difficult it is for a human to distinguish between tweets that were generated by SHIELD and tweets that were posted by real users. Experimental results followed by a statistical significance analysis showed that human participants cannot reliably distinguish between real and synthetic tweets.

There are different forms of uncertainty in satellite communications, including cognitive interferers, channel conditions, packet traffic, and spectrum occupancy of users across channels. In addition, delay (such as propagation delay observed over satellite links) increases spectrum uncertainty and makes spectrum sensing and spectrum access two challenging tasks. To address such challenges, this paper presents a regret minimization solution for primary user (PU) and secondary user (SU) spectrum access to satellite resources in the presence of cognitive interferers. This robust game theoretic solution supports hierarchical spectrum sharing and dynamic spectrum access over multiple channels. Users select channels for data transmission and perform power control to optimize individual utility functions that are random due to different forms of uncertainty. The proposed game engine based on regret minimization framework provides a low-complexity and fast solution compared with traditional game solutions based on expected utility maximization. Detailed numerical results evaluate throughput and delay of PUs and SUs in the presence of cognitive interferers and compare the robust game theory-enabled approach with two benchmark schemes (with and without knowledge on channel availability). To support controllable and repeatable test and evaluation with real radios, an emulation testbed is built with software-defined radios connected with a network channel emulator that generates channel, mobility, and interference effects for satellite communications. GNU Radio modules are developed for cognitive network functionalities and run on USRP N210 radios that represent SU, PU, interferer, and satellite nodes. Emulation tests validate the effectiveness of the proposed solution under real radio effects.

Distributed interference cancellation is considered for a single source and multiple collaborative receivers in the presence of multiple unknown interferers. Receivers observe interference in a distributed setting and collaborate to mitigate the common interference from multiple unknown interferers. Under Rayleigh frequency-flat fading conditions, the ergodic capacity is shown to improve with collaboration surpassing the communication overhead due to collaboration. In particular at SINR of -15 dB, the performance improvement relative to the case without collaboration is a factor of 15. Numerical results are also provided for constrained alphabet signaling, along with simulation results based on DVB-S2#x002F;T2 standard. Finally, results from hardware-in-the-loop tests using USRPs are presented to show the feasibility of practical implementation. Both simulation and radio tests demonstrate the benefit of collaboration for distributed cancellation of unknown interference.

The backpressure algorithm is known to provide throughput optimality in routing and scheduling decisions for multi-hop networks with dynamic traffic. The essential assumption in the backpressure algorithm is that all nodes are benign and obey the algorithm rules governing the information exchange and underlying optimization needs. Nonetheless, such an assumption does not always hold in realistic scenarios, especially in the presence of security attacks with intent to disrupt network operations. In this paper, we propose a novel mechanism, called virtual trust queuing, to protect backpressure algorithm based routing and scheduling protocols against various insider threats. Our objective is not to design yet another trust-based routing to heuristically bargain security and performance, but to develop a generic solution with strong guarantees of attack resilience and throughput performance in the backpressure algorithm. To this end, we quantify a node's algorithm-compliance behavior over time and construct a virtual trust queue that maintains deviations of a give node from expected algorithm outcomes. We show that by jointly stabilizing the virtual trust queue and the real packet queue, the backpressure algorithm not only achieves resilience, but also sustains the throughput performance under an extensive set of security attacks. Our proposed solution clears a major barrier for practical deployment of backpressure algorithm for secure wireless applications.

This paper addresses the problem of joint inference and optimization in wireless networks. An optimization framework based on information-geometric network inference is developed and implemented (using a real radio emulation testbed) with scalable solutions to infer the end-to-end rate distributions of stochastic network flows from link rate measurements. The proposed low-complexity solutions apply when the underlying network inference (network tomography) problem can be decomposed to smaller-size subproblems that are solved independently by partially inferring only the flow rates of interest. The solutions are extended to infer flow rates jointly with link loss rates when retransmissions are considered over unreliable wireless links. By using the inferred distributions of flow rates, an inference mechanism is presented to optimize the network performance. First, the distributions of flow rates are inferred from the average rate measurements on selected links. Then, the distributions of all links rates are computed from the inferred distributions of flow rates. Finally, these inference results are used with network optimization. The weighted sum of link outage probabilities is minimized by adapting power control or routing decisions in mobile wireless access. This approach is iterated between network inference and optimization, providing link outage and end-to-end throughput gains compared to static approaches with fixed network (inference and optimization) parameters. The joint network inference and optimization framework is implemented with real configurable radios and the performance is verified with hardware-in-the-loop emulation test results that are obtained with actual radio transmissions over emulated channels.

This paper presents a game theoretic solution for hierarchical spectrum sharing between primary users (PUs) and secondary users (SUs) in the presence of cognitive interferers. There exist several forms of uncertainty, including channel conditions, packet traffic, and spectrum occupancy of users across channels. This uncertainty is further aggravated by delays (such as propagation delays observed over satellite links) that make spectrum-efficient communication a challenging task. A robust game theoretic framework is developed for dynamic spectrum access (DSA) management over multiple channels. Cognitive functionalities employed in the game solution include selecting channels for data transmission and performing power control at each user to sustain target rates. By considering random utility functions, the game engine based on regret minimization provides low complexity and fast solutions compared to traditional game solutions based on expected utility maximization. Detailed numerical results with comparison to benchmark schemes (the ideal case and the random case where users have perfect or no knowledge on channel availability, respectively) are provided to show the effectiveness of robust game theory-enabled approach.

The backpressure algorithm is known to provide throughput optimality in routing and scheduling decisions for multi-hop networks with dynamic traffic. The essential assumption in the backpressure algorithm is that all nodes are benign and obey the algorithm rules governing the information exchange and underlying optimization needs. Nonetheless, such an assumption does not always hold in realistic scenarios, especially in the presence of security attacks with intent to disrupt network operations. In this paper, we propose a novel mechanism, called virtual trust queuing, to protect backpressure algorithm based routing and scheduling protocols from various insider threats. Our objective is not to design yet another trust-based routing to heuristically bargain security and performance, but to develop a generic solution with strong guarantees of attack resilience and throughput performance in the backpressure algorithm. To this end, we quantify a node's algorithm-compliance behavior over time and construct a virtual trust queue that maintains deviations from expected algorithm outcomes. We show that by jointly stabilizing the virtual trust queue and the real packet queue, the backpressure algorithm not only achieves resilience, but also sustains the throughput performance under an extensive set of security attacks.

We study deadline-aware scheduling with adaptive network coding (NC) for real-time traffic over a single-hop wireless network. To meet hard deadlines of real-time traffic, the block size for NC is adapted based on the remaining time to the deadline so as to strike a balance between maximizing the throughput and minimizing the risk that the entire block of coded packets may not be decodable by the deadline. This sequential block size adaptation problem is then cast as a finite-horizon Markov decision process. One interesting finding is that the optimal block size and its corresponding action space monotonically decrease as the deadline approaches, and that the optimal block size is bounded by the “greedy” block size. These unique structures make it possible to significantly narrow down the search space of dynamic programming, building on which we develop a monotonicity-based backward induction algorithm (MBIA) that can find the optimal block size in polynomial time. Furthermore, a joint real-time scheduling and channel learning scheme with adaptive NC is developed to adapt to channel dynamics in a mobile network environment. Then, we generalize the analysis to multiple flows with hard deadlines and long-term delivery ratio constraints. We devise a low-complexity online scheduling algorithm integrated with the MBIA, and then establish its asymptotical utility optimality. The analysis and simulation results are corroborated by high-fidelity wireless emulation tests, where actual radio transmissions over emulated channels are performed to demonstrate the feasibility of the MBIA in finding the optimal block size in real time.

This paper models and analyzes the problem of search (navigation) with local information in combined social and wireless communication networks. Social networks are modeled with short-range and long-range connections representing small-world and scale-free network characteristics. By distinguishing the delay and success probability on different link types, the end-to-end delay distribution and success probability are first derived as functions of the social separation from the destination. New routing algorithms are then developed to improve the delay and chain completion success, and the effects of delay deadline on success probability are evaluated. The analysis is extended to the multi-layer combined social and communication network model, where wireless communication becomes the underlay to route information with the aid of social connections. The analytical results on delay and success probability are validated by comparing them with search results on a real-world social and communication network. Results of this paper show how social connections can help reduce the search delay and increase the success probability in chain completion that runs on interdependent social and wireless communication network structures.

A fundamental challenge in wireless networks is how to handle packet loss due to noise, interference, and dynamic channel effects, especially when there is no per-packet acknowledgement due to additional delay and potential loss of feedback packets. We design and implement a packet coding optimization scheme, applied at the source node, to enhance end-to-end transmission reliability in a lossy multi-hop network. Specifically, each source node transmits a file of packets to its destination node in the network where the end-to-end packet acknowledgement is not readily available because of the possible error or delay effects over multiple hops. By simply retransmitting packets, a source node cannot guarantee innovative packet arrivals at the destination node. Thus, we design an optimization scheme for adaptive packet coding applied at the source node to avoid transmitting redundant packets, thereby significantly improving the throughput, even for the case of a single unicast session. This scheme does not require any knowledge of network topology and can seamlessly operate with any routing or network coding protocol used at the intermediate relay nodes. We analyze the throughput properties of coded transmissions and verify the feasibility of performance gains via simulations. Then, we provide high fidelity emulation testbed results with real radio transmissions over emulated channels to evaluate the throughput gains. Without relying on end-to-end acknowledgment for each packet, we show that the adaptive packet coding optimization scheme can achieve significantly higher throughput than the retransmission scheme in lossy networks with unicast traffic.

We consider the problem of reliable spectrum sensing and opportunistic access on channels with stochastic traffic in batch processing systems such as network coding (NC). We show how a secondary user (SU) can leverage the structure induced by block-based NC on primary users' (PUs) channels to mitigate the effects of channel sensing errors and improve the detection of idle PU spectrum and the throughput. NC is known to improve the transmission efficiency and therefore when applied on a PU channel, it can extend the spectrum availability for the SUs. We refer to the additional gain of spectrum predictability from NC and show that under possible sensing errors the SU can more reliably detect the idle spectrum if the PUs' channels carry network-coded transmissions even when the channel utilization is fixed. We consider two different objectives at the SU. For quickest detection, the SU applies the Cumulative Summation (CUSUM) algorithm to detect idle slots on a PU channel and further improves the detection capability with the Viterbi algorithm, if the PU spectrum dynamics are known. For throughput maximization, the SU tracks the PU spectrum with the Partially Observable Markov Decision Process (POMDP) approach. Our results show that NC renders the spectrum more predictable, which can be used by the SUs to mitigate the effects of sensing errors and improve the throughput. We validate these results with real radio measurements taken in software-defined radio based wireless network tests.

This paper addresses the problem of search with local information in combined social and communication networks. Social networks are modeled with short-range and long-range connections representing small-world and scale-free network characteristics. By distinguishing the delay and success probability on different social links, the end-to-end delay distribution and success probability are derived as functions of the social separation from the destination. Also, greedy routing algorithms are developed to improve the delay and chain completion success. Then, the analysis is extended to the combined social and communication networks, where wireless communication becomes the underlay to route information with the aid of social connections. The analytical results are validated via simulated search results on a combined social and communication network. Our results show how social connections can help reduce the search delay and increase the success probability in chain completion.

In this demonstration, we present new capabilities of Intelligent Automation Inc. (IAI) Radio Frequency Network Emulation Simulation Tool (RFnestTM) (www.i-a-i.com/rfnest). RFnestTM is a Field Programmable Gate Array (FPGA) based network channel emulator that allows all of the channels of a full network of radio nodes to be emulated in real time, with all communication nodes experiencing a realistic channel impulse response. This allows RFnestTM to be used for protocol testing, replaying field tests, and model validation. RFnestTM has a modular design with three main capabilities: 1) FPGA based emulation hardware with RF front ends that allows nodes with real radios to send their RF signal over an emulated channel without any modification to the radio, 2) modeling of time-varying channel impulse responses within the emulation hardware, with channel properties based on mobility defined with a scripted or interactive Graphic User Interface (GUI) environment, and 3) integration with network emulators and monitoring functionality that allows the user to instantiate, manage, and monitor real and virtual network nodes within the scenario. In the new version of RFnestTM we now support: Frequency programmability: The new RFnestTM allows users to change the center frequency through software from 800 Mhz to 2.7GHz. Increased bandwidth: RFnestTM supports a bandwidth of 60 MHz. Higher fidelity channel modeling: Through software and hardware updates RFnestTM support wireless channels with up to 20 taps and a wide range of Doppler frequencies. MIMO: RFnestTM now supports different combination of MIMO channel between transmitters and receivers.

The throughput and stability properties of wireless network coding are evaluated for an arbitrary number of terminals exchanging broadcast traffic with the aid of a relay. First, coding and scheduling schemes are derived that minimize the number of transmissions needed for each node to broadcast one packet. For stochastically varying traffic, the stable throughput is then compared under both digital and analog network coding schemes. The initial analysis focuses on a network with a single relay. Extensions to arbitrary terminal-relay configurations are then outlined for a general multihop network. Backpressure-like algorithms for jointly achieving throughput optimal scheduling and network coding are given for each network coding scheme.

We consider a source node broadcasting to two receivers over a general erasure channel with receiver feedback. We characterize the capacity region of the channel and construct algorithms based on linear network coding (either randomized or depending on channel dynamics) that achieve this capacity. We then consider stochastic arrivals at the source for the two destinations and characterize the stable throughput region achieved by adapting the same algorithms that achieve capacity. Next, we modify these algorithms to improve their delay performance and characterize their stable throughput regions. Although the capacity and stability regions obtained by the algorithms are not always identical (because of the extra overhead needed for the algorithms to handle stochastic traffic), they are within a few bits of each other and have similar forms. This example exhibits an unusual relationship between capacity and stability regions and extends similar prior studies for multiple access channels.

We study adaptive network coding (NC) for scheduling real-time traffic over a single-hop wireless network. To meet the hard deadlines of real-time traffic, it is critical to strike a balance between maximizing the throughput and minimizing the risk that the entire block of coded packets may not be decodable by the deadline. Thus motivated, we explore adaptive NC, where the block size is adapted based on the remaining time to the deadline, by casting this sequential block size adaptation problem as a finite-horizon Markov decision process. One interesting finding is that the optimal block size and its corresponding action space monotonically decrease as the deadline approaches, and the optimal block size is bounded by the "greedy" block size. These unique structures make it possible to narrow down the search space of dynamic programming, building on which we develop a monotonicity-based backward induction algorithm (MBIA) that can solve for the optimal block size in polynomial time. Since channel erasure probabilities would be time-varying in a mobile network, we further develop a joint real-time scheduling and channel learning scheme with adaptive NC that can adapt to channel dynamics. We also generalize the analysis to multiple flows with hard deadlines and long-term delivery ratio constraints, devise a low-complexity online scheduling algorithm integrated with the MBIA, and then establish its asymptotic throughput-optimality. In addition to analysis and simulation results, we perform high fidelity wireless emulation tests with real radio transmissions to demonstrate the feasibility of the MBIA in finding the optimal block size in real time.

We consider network tomography and monitoring in dynamic wireless systems and leverage analytical tools from optimization theory and information geometry to infer the invariant statistical network structures. We extend the classical network tomography problem beyond average link rate measurements and develop a systematic optimization mechanism to infer the end-to-end wireless network behavior. This involves estimating the distributions of global network flow rates from the arbitrary statistics collected for the wireless link (channel) rates subject to the topology and link capacity constraints. We develop first a centralized network inference framework based on minimizing the distance of network flow rates from the prior information in the probability space that is spanned by the measurement constraints. Then, distributed implementation follows from message passing among the individual probes in the network and balances the complexity and convergence trade-offs. This formulation facilitates multi-scale multi-resolution inference of flow rates along with link capacity estimation. The underlying optimization framework for information-geometric network inference adapts to wireless network dynamics and offers robust operation with respect to the measurement errors and conflicts as well as the temporal and spatial variations in wireless networks.

It is known that distributed power control in wireless ad-hoc networks is challenging, due to the inherent global coupling between concurrent transmissions interfering with each other. Observing that the globally optimal point lies on the boundary of the feasible region, we transform the utility maximization problem into a more structured problem in the form of maximizing the minimum weighted utility. Then, we develop a centralized algorithm for the minimum weighted utility maximization problem as a benchmark. Next, by using extended duality theory, we introduce penalty multipliers and decompose the minimum weighted utility maximization problem into subproblems for individual users. Appealing to the simulated annealing method, we propose a distributed stochastic power control algorithm, where each user stochastically adjusts its target utility to improve the overall system utility. Although the underlying optimization problem is nonconvex, our algorithm can guarantee global optimality although the convergence rate may be slow due to the usage of simulated annealing. We improve the convergence rate further by devising an enhanced algorithm based on the geometric cooling schedule.

We consider a cognitive radio network where primary users (PUs) employ network coding for data transmissions. We view network coding as a spectrum shaper, in the sense that it increases spectrum availability to secondary users (SUs) and offers more structure of spectrum holes, which in turn improves the predictability of the primary spectrum. With this spectrum shaping effect of network coding, each SU can carry out adaptive channel sensing by dynamically updating the list of the (predicted) idle PU channels and giving priority to these channels for spectrum sensing. This dynamic spectrum access approach with network coding improves how SUs detect and utilize spectrum holes over PU channels. Our results show that compared to the existing approaches based on retransmission, both PUs and SUs can achieve higher stable throughput, thanks to the spectrum shaping effect of network coding.

We consider two sources in a wireless network exchanging stochastically varying traffic using an intermediate relay. Each relay use incurs some cost, which, for example, could be transmission energy. This cost is shared between the sources when packets from both are transmitted simultaneously by the relay using network coding. If the relay transmits a packet originating from one source only, the cost is incurred by that source only. In this setting, we study transmission policies that tradeoff the average cost with the average packet delay. We first present the cost-delay tradeoff for a centralized scheme using Lyapunov stability arguments. Next, we consider a distributed policy, where each source aims to optimize its own cost-delay tradeoff. We determine the Nash equilibrium of the resulting non-cooperative game and show that it performs worse than the centralized algorithm. To overcome this limitation, we introduce a pricing mechanism at the relay, which is shown to achieve the centralized performance. These algorithms, though oblivious to the arrival statistics, do require global knowledge of queue backlogs. Lastly, we consider distributed algorithms that overcome this requirement. Among those, we observe that simple queue-length threshold algorithms perform remarkably well.

Due to the broadcast nature of the wireless medium, wireless networks are highly susceptible to jamming attacks. Such attacks are often studied in a game theoretic framework under the assumption of uninterrupted traffic subject to continuous jamming opportunities. Instead, we analyze the effect of dynamically changing traffic on jamming games for power controlled medium access. Random packet arrivals raise the possibility that the transmitter queues may be empty when jamming attacks start and thus waste the energy of jammers. We consider a non-cooperative game in which transmitters and jammers select their transmission power to balance the transmission cost subject to delay and energy constraints. We show that jammers incur a significant performance loss when they do not have knowledge of transmitter queue states. Dynamic traffic increases the immunity to jamming attacks and gives insights into defense mechanisms.

We study wireless multicasting from two sources to two destinations with the assistance of a single half-duplex relay. The objective is to evaluate the throughput and error performance of different analog and digital relay schemes with linear network coding at the relay. The analog relay node forwards either a scaled version of the received signal to the destinations, or alternatively, first filters the received signals to generate a linear Minimum Mean Squared Error (MMSE) estimate, which is subsequently forwarded. The digital relay scheme first detects the source transmissions, combines the packets with a network code, and forwards the resulting symbols to the destinations. For all schemes the destinations recover the source and relay signals by first applying linear MMSE filters, followed by decoding of the source bits. The performance of the schemes are compared in terms of normalized throughput (bits per channel use accounting for the delay due to the relay) and uncoded error probability, given a normalized power constraint. Both narrowband and wideband transmission schemes are considered. Our results show that the analog relay schemes outperform the digital network coding scheme with respect to both throughput and error probability because of error propagation through the relay. Numerical results are presented, which illustrate throughput-reliability trade-offs for all schemes considered.

We consider multicast communications from a single source to multiple destinations through a wireless network with unreliable links. Random linear network coding achieves the min-cut flow capacity; however, additional overhead is needed for end-to-end error protection and to communicate the network coding matrix to each destination. We present a joint coding and training scheme in which training bits are appended to each source packet, and the channel code is applied across both the training and data. This scheme allows each destination to decode jointly the network coding matrix along with the data without knowledge of the network topology. It also balances the reliability of communicating the network coding matrices with the reliability of data detection. The throughput for this scheme, accounting for overhead, is characterized as a function of the packet size, channel properties (error and erasure statistics), number of independent messages, and field size. We also compare the performance with that obtained by individual channel coding of training and data. Numerical results are presented for a grid network that illustrate the reduction in throughput due to overhead.

The transmission of packets is considered from one source to multiple receivers over single-hop erasure channels. The objective is to evaluate the stability properties of different transmission schemes with and without network coding. First, the throughput limitation of retransmission schemes is discussed and the stability benefits are shown for randomly coded transmissions, which, however, need not optimize the stable throughput for finite coding field size and finite packet block size. Next, a dynamic scheme is introduced for distributing packets among virtual queues depending on the channel feedback and performing linear network coding based on the instantaneous queue contents. The difference of the maximum stable throughput from the min-cut rate is bounded as function of the order of erasure probabilities depending on the complexity allowed for network coding and queue management. This queue-based network coding scheme can asymptotically optimize the stable throughput to the max-flow min-cut bound, as the erasure probabilities go to zero. This is realized for a finite coding field size without accumulating packet blocks at the source to start network coding. The comparison of random and queue-based dynamic network coding with plain retransmissions opens up new questions regarding the tradeoffs of stable throughput, packet delay, overhead, and complexity.

In wireless networks, throughput optimization is an essential performance objective that cannot be adequately characterized by a single criterion (such as the minimum transmitted or sum-delivered throughput) and should be specified over all source-destination pairs as a rate region. For a simple and yet fundamental model of tandem networks, a cross-layer optimization framework is formulated to derive the maximum throughput region for saturated multicast traffic. The contents of network flows are specified through network coding (or plain routing) in network layer and the throughput rates are jointly optimized in medium access control layer over fixed set of conflict-free transmission schedules (or optimized over transmission probabilities in random access). If the network model incorporates bursty sources and allows packet queues to empty, the objective is to specify the stability region as the set of maximum throughput rates that can be sustained with finite packet delay. Dynamic queue management strategies are used to expand the stability region toward the maximum throughput region. Network coding improves throughput rates over plain routing and achieves the largest gains for broadcast communication and intermediate network sizes. Throughput optimization imposes fundamental tradeoffs with transmission and processing energy costs such that the throughput-optimal operation is not necessarily energy efficient.

This paper addresses network coding in wireless networks in conjunction with medium access control (MAC). It is known that coding over wired networks enables connections with rates that cannot be achieved by routing. However, the properties of wireless networks (e.g., omnidirectional transmissions, destructive interference, single transceiver per node, finite energy) modify the formulation of time-varying network coding in a way that reflects strong interactions with underlying MAC protocols and deviates from the classical approach used in wired network coding. To perform network coding over conflict-free transmission schedules, predetermined network realizations are separately activated by a time-division mechanism and the content of network flows is derived through network coding to optimize performance measures such as achievable throughput and energy costs. A systematic method is presented to construct linear wireless network codes and interactions with MAC schedules are discussed under wireless assumptions. Network coding is also extended to operate with arbitrary (random or scheduled access based) MAC protocols. Alternatively, conflict-free transmission schedules are jointly constructed with network codes by decomposing wireless networks into subtrees and employing graph coloring on simplified subtree graphs. Finally, network coding and plain routing are compared in terms of throughput, energy and delay performance under different MAC solutions.

In this paper, we compare the effects of the saturated and possibly emptying packet queues on wireless network coding (or plain routing as a special case) in a simple tandem network. We consider scheduled or random access with omnidirectional transmissions and assume the classical collision channel model without simultaneous transmission and reception by any node. For the case of multiple source nodes, we evaluate the multicast throughput rates jointly achievable by different source-destination pairs under the separate assumptions of network coding and plain routing only. Particularly, we specify the throughput region for saturated queues and stability region for possibly emptying queues. We also evaluate the fundamental trade-offs among the performance objectives of throughput and transmission and processing energy costs. Finally, we extend the analysis to non-cooperative network operation with selfish nodes competing for limited network resources. We point at the inefficiency of competitive medium access control and network coding (or plain routing) decisions at individual nodes, and introduce a pricing-based cooperation stimulation mechanism to improve the throughput and energy efficiency performance

In this paper, we address the joint design and distributed implementation of medium access control (MAC) and network coding in wireless ad hoc networks. We consider a slotted wireless network with links modeled as classical collision channels. We assume omnidirectional packet transmissions and do not allow simultaneous transmission and reception by any node. These wireless network properties demand a new formulation of time-varying network coding with additional constraints that reflect the strong interactions with the underlying MAC operation. First, we outline how to construct network coding over a predetermined set of wireless network realizations with conflict-free transmission schedules. The wireless network realizations are separately activated using a time division mechanism and the network flows are chosen to optimize the performance measures (such as stable throughput or average energy costs) through wireless network coding. Next, we follow an alternative approach of joint wireless network coding and link scheduling with a distributed implementation. The analysis is based on decomposing wireless networks into subtrees and applying graph coloring on the simplified subtree network representations to derive the wireless network codes that are further converted to the conflict-free sets of transmission schedules. Finally, we extend network coding to operate with arbitrary single-receiver MAC protocols within each receiver's area

We rediscover the value of scheduled access through a detailed foray into the questions of throughput and energy consumption for MAC protocols in ad hoc wireless networks, where the optimal channel access scheduling is NP-complete. We propose a two-layered time-division heuristic of receiver activation and Group TDMA as polynomial-time solutions to throughput and energy-efficient link scheduling and resource allocation in networks with dynamically changing transmitter-receiver pairs.

In this paper, we address the collision resolution (CR) problem from an energy-efficiency point of view and develop a residual-energy-based collision resolution algorithm (CRA) for energy-limited terminals. In this algorithm, which is based on tree-splitting, packets involved in a collision are partitioned into subsets according to the amount of residual battery energy left at the corresponding terminals, and retransmissions are scheduled according to a tree structure. We extend the proposed energy-based CR approach to cases without hard energy constraints but, rather, with energy-efficiency objectives. The algorithm then utilizes the distance from the receiver as the criterion. We evaluate the proposed algorithm via simulation for communication systems ranging from simple single-cell classical collision channel models to general multihop wireless ad-hoc networks.