Real-time anomaly detection in Open Radio Access Networks (O-RAN) demands high accuracy, low false alarms, and computational efficiency for resource-constrained edge deployment. Traditional methods struggle with computational overhead, inconsistent cross-domain performance, and suboptimal feature representations that miss subtle attacks on O-RAN's open interfaces. We present CoMeT-Net (Consensus Memory Template Network), a framework achieving state-of-the-art detection through three innovations: (1) structured memory banks enabling template-based consensus voting with $O(N \cdot C)$ complexity; (2) adaptive gating that downweights ambiguous features as a learned noise filter; (3) contrastive alignment unifying feature learning and classification. Deployed in O-RAN infrastructure via edge servers and Near-RT RIC xApp, CoMeT-Net enables dynamic threat mitigation through PRB throttling and RRC connection release. On network traffic datasets, CoMeT-Net achieves 99.35% F1 score with 10$\times$ lower false alarm rates than baselines while maintaining 0.3-3ms inference across hardware tiers from servers to Raspberry Pi 4. O-RAN testbed validation demonstrates effective isolation, degrading attacker latency to >1400ms while preserving 15-20ms for legitimate users.

Open Radio Access Network (O-RAN) is expected to support a diverse range of cutting-edge, real-time, and heterogeneous dynamic traffic demands. In response, we formulate a network-slicing resource allocation optimization problem to enhance the Quality of Service (QoS) in O-RAN. The proposed problem is formulated as a Nonlinear Programming (NLP) problem, designed to efficiently meet User Equipment (UE) traffic demands in a dynamic environment while maintaining QoS and optimizing network radio resources. Given the combinatorial nature of this problem, finding an optimal solution is challenging and often impractical in a dynamic traffic environment. To address this, we propose decomposing the prime problem into two sub-problems: a Long-Term Resource Allocation (LTRA) problem focusing on resource allocation decisions, and a Short-Term Resource Scheduling (STRS) problem that manages the scheduling of demanded resources. We propose the Long Short-Term Memory Actor-Critic-based (LS-RLSlice) algorithm, a novel approach that modifies LSTM and Deep Deterministic Policy Gradient (DDPG) algorithms for efficiently solving LTRA and STRS. We perform simulations to show that our proposed algorithm outperforms state-of-the-art schemes and significantly reduces the utilization of network resources. Finally, our proposed solution is validated using the real-world POWDER testbed supporting the O-RAN stack and provides configuration setups for Non-Real-Time and Near-Real-Time RAN Intelligent Controllers (Non-RT RIC and Near-RT RIC).

Next-Generation Radio Access Networks (NG-RAN) aim to support diverse vertical applications with strict security, latency, and Service-Level Agreement (SLA) requirements. These demands introduce challenges in securing the infrastructure, allocating resources dynamically, and enabling real-time reconfiguration. This demo presents SnSRIC, a secure and intelligent network slicing framework that mitigates a range of Distributed Denial-of-Service (DDoS) attacks in Open RAN environments. SnSRIC incorporates an AI-driven xApp that dynamically allocates Physical Resource Blocks (PRBs) to active users while enforcing slice-level security. The system detects anomalous behavior, distinguishes between benign and malicious devices, and uses the E2 interface to throttle rogue signaling while maintaining service continuity for legitimate users.

Next Generation Radio Access Networks (NG-RANs) are a promising paradigm for meeting 6G and future application requirements. However, the practical implementation of NG-RAN systems faces significant challenges due to novel technologies, network densification, and more complex applications. Specifically, the limited capacity of front-haul links and privacy concerns have posed severe constraints that must be addressed. To overcome these obstacles, we present a novel approach, called FedBNG, which is a disaggregated and distributed Federated Learning (FL)-based algorithm for NG-RAN. This algorithm enables collaboration between User Equipment (UEs) and the NG-RAN infrastructure through a learning process and shared prediction models, ultimately improving privacy and alleviating the burden on the front-haul interface. Using a shared predictive model, our proposed approach facilitates cooperative learning between Radio Units (RUs) and Distributed Units (DUs). To accomplish this, we initially used the first-phase training models of RUs and DUs as input for local training. Subsequently, the suboptimal DU models are uploaded to the Central Unit (CU) for the next phase of global training. We present numerical results to evaluate the efficacy of our proposed approach in terms of accuracy, service latency, and traffic volume. Our algorithm’s convergence properties demonstrate that it outperforms the current state-of-the-art solution based on FedAvg.

Task offloading with Mobile-Edge Computing (MEC) is envisioned as a promising technique to prolong battery lifetime and enhance the computational capacity of mobile devices. In this paper, we consider a multi-user MEC system with a Base Station (BS) equipped with a computation server that assists users in executing computation-intensive tasks via offloading. Exploiting approximate computing in MEC, we can trade the output accuracy over a subset of offloading data instead of the entire dataset. We formulate the Energy-Latency-aware Task Offloading and Approximate Computing (ETORS) problem, aiming to optimize the trade-off between energy consumption and latency. Due to the mixed-integer nature of this problem, we employ the Dual-Decomposition Method (DDM) to decompose the original problem into three subproblems—namely the Task-Offloading Decision (TOD), the CPU Frequency Scaling (CFS), and the Quality of Computation Control (QoCC). Our approach consists of two iterative layers: in the outer layer, we adopt the duality technique to find the optimal value of the Lagrangian multiplier associated with the primal problem; and in the inner layer, we formulate the subproblems that can be solved efficiently using convex optimization techniques. Simulation results coupled with real-time experiments on a small-scale MEC testbed show the effectiveness of our proposed resource allocation scheme and its advantages over existing approaches.

Cloud Radio Access Network (C-RAN) is a key architecture for 5G cellular wireless network that aims at improving spectral and energy efficiency of the network by uniting traditional RAN with cloud computing. In this paper, a novel resource allocation scheme that optimizes the network energy efficiency of a C-RAN is designed. First, an energy consumption model that characterizes the computation energy of the BaseBand Unit (BBU) is introduced based on empirical results collected from a programmable C-RAN testbed. Then, an optimization problem is formulated to maximize the energy efficiency of the network, subject to practical constraints including Quality of Service (QoS) requirement, radio remote head transmit power, and fronthaul capacity limits. The formulated Network Energy Efficiency Maximization (NEEM) problem jointly considers the tradeoff among the network accumulated data rate, BBU power consumption, fronthaul cost, and beamforming design. To deal with the non-convexity and mixed-integer nature of the problem, we utilize successive convex approximation methods to transform the original problem into the equivalent Weighted Sum-Rate (WSR) maximization problem. We then propose a provably-convergent iterative method to solve the resulting WSR problem. Extensive simulation results coupled with real-time experiments on a small-scale C-RAN testbed show the effectiveness of our proposed resource allocation scheme and its advantages over existing approaches.

Mobile Augmented Reality (AR) has become a reality thanks to improvements in mobile hardware. Still, mobile AR lags behind its desktop counterpart in both latency and performance. Simple offloading to external computers has been attempted, but is not practical due to high communication latency and adverse user experience. In this paper, we propose a novel Mobile Edge Computing framework for Augmented Reality applications (MEC-AR). MEC-AR is designed to take advantage of 5G cellular networks and make optimized computation-offloading decisions in a multi-tiered hierarchy. A three-layered architecture involving the end user, the mobile edge, and finally the cloud is envisioned. In the context of MEC resource management, we cast a Mixed Integer Linear Program (MILP) that aims at finding an efficient application placement on the MEC-AR layers to minimize the network latency. We evaluate the performance of our proposed MEC-AR framework by conducting extensive experimental analysis using images taken around Rutgers University. Simulation results coupled with real-time experiments on a small-scale MEC testbed show that our hierarchical computation mechanism improves the performance of mobile AR applications in terms of both energy consumption and network latency.

Task offloading with Mobile-Edge Computing (MEC) is envisioned as a promising technique for prolonging battery lifetime and enhancing the computation capacity of mobile devices. In this paper, we consider a multi-user MEC system with a Base Station (BS) equipped with a computation server assisting mobile users in executing computation-intensive real-time tasks via offloading technique. We formulate the Energy-Latency-aware Task Offloading and Approximate Computing (ETORS) problem, which aims at optimizing the trade-off between energy consumption and application completion time. Due to the centralized and mixed-integer natures of this problem, it is very challenging to derive the optimal solution in practical time. This motivates us to employ the Dual-Decomposition Method (DDM) to decompose the original problem into three subproblems-namely the Task-Offloading Decision (TOD), the CPU Frequency Scaling (CFS), and the Quality of Computation Control (QoCC). Our approach consists of two iterative layers: in the outer layer, we adopt the duality technique to find the optimal value of Lagrangian multiplier associated prime problem; and in the inner layer, we formulate the subproblems that can be solved efficiently using convex optimization techniques. We show that the computation offloading selection depends not only on the computing workload of a task, but also on the maximum completion time of its immediate predecessors and on the clock frequency as well as on the transmission power of the mobile device. Simulation results coupled with real-time experiments on a small-scale MEC testbed show the effectiveness of our proposed resource allocation scheme and its advantages over existing approaches.

Our goal is to design, develop, and validate via computer simulations and testbed experiments novel resource allocation and computation offloading algorithms aimed at improving the spectral and energy efficiency in next generation cellular networks. To achieve this goal, we exploit the high degree of cooperation provided by Software-Defined Networking (SDN) and Network Function Virtualization (NFV) for cloud-assisted wireless networks, including Cloud Radio Access Network (C-RAN) and Mobile Edge Computing (MEC) infrastructures.

Mobile Edge Computing (MEC) has recently emerged as a promising paradigm to enhance mobile networks' performance by providing cloud-computing capabilities to the edge of the Radio Access Network (RAN) with the deployment of MEC servers right at the Base Stations (BSs). Meanwhile, in-network caching and video transcoding have become important complementary technologies to lower network cost and to enhance Quality of Experience (QoE) for video-streaming users. In this paper, we aim at optimizing the QoE for dynamic adaptive video streaming by taking into account the Distortion Rate (DR) characteristics of videos and the coordination among MEC servers. Specifically, a novel Video-streaming QoE Maximization (VQM) problem is cast as a Mixed-Integer Nonlinear Program (MINLP) that jointly determines the integer video resolution levels and video transmission data rates. Due to the challenging combinatorial and non-convex nature of this problem, the Dual-Decomposition Method (DDM) is employed to decouple the original problem into two tractable subproblems, which can be solved efficiently using standard optimization solvers. Real-time experiments on a wireless video streaming testbed have been performed on a FDD-downlink LTE emulation system to characterize the performance and computing resource consumption of the MEC server under various realistic conditions. Emulation results of the proposed strategy show significant improvement in terms of users' QoE over traditional approaches.

Cloud Radio Access Network (C-RAN) is emerging as a transformative paradigmatic architecture for the next generation of wireless cellular networks. In this demo, a programmable C-RAN testbed is implemented where the Base Band Unit (BBU) is virtualized using the OpenAirInterface (OAI) software platform, and the eNodeB and User Equipment (UEs) are implemented using Software-Defined Radio (SDR) USRP boards. Based on our testbed architecture, we further develop a novel hierarchical computation mechanism to improve the performance of mobile Augmented Reality (AR) applications.

Cloud radio access network (C-RAN) is emerging as a transformative paradigmatic architecture for the next generation of cellular networks. In this paper, a novel resource allocation solution that optimizes the energy consumption of a C-RAN is proposed. First, an energy consumption model that characterizes the computation energy of the base band unit (BBU) pool is introduced based on the empirical results collected from a programmable C-RAN testbed. Then, the resource allocation problem is split into two subproblems-namely the bandwidth power allocation (BPA) and the BBU energy-aware resource allocation (EARA). The BPA, which is first cast via mixed-integer nonlinear programming and then reformulated as a convex problem, aims at assigning a feasible bandwidth and power to serve all users while meeting their quality of service (QoS) requirements. The second subproblem, i.e., the BBU EARA, is defined as a bin-packing problem that aims at minimizing the number of active virtual machines in the BBU pool to save energy. Simulation results coupled with the real-time experiments on a small-scale C-RAN testbed show that the proposed resource allocation solution optimizes the energy consumption of the network while meeting practical constraints and QoS requirements, and outperforms competing algorithms, such as best fit decreasing, RRH-clustering, and SINR-based.