Deploying transformer-based anomaly detection in Industrial Internet of Things (IIoT) networks poses unique challenges including stringent latency requirements, heterogeneous protocols, and safety-critical operations. The proposed work combines multimodal federated learning with adaptive resource optimization, specifically designed for diverse edge devices by addressing three key challenges: (1) a Cross-Modal Progressive Attention (CMPA) mechanism that enables the efficient fusion of vibration, temperature, and pressure sensor data through trainable routing matrices, thus, reducing the complexity fromO(T2max) toO(TmaxlogTmax); (2) a Resource-Adaptive Federated Optimization (RAFO) algorithm with provable convergence guarantees under non-IID and asynchronous conditions; and (3) an adaptive differential privacy scheme that dynamically adjusts privacy budgets based on model convergence and data sensitivity. The theoretical analysis confirms that the convergence bounds under asynchronous updates and privacy composition is guaranteed. Evaluation is performed on four IIoT datasets (TON-IoT, SWAT, WADI, and HAI) that demonstrates that FedTransformer-Edge attains 91.3% F1-score (±2.1%, 95% CI) while reducing communication overhead by 62% and energy consumption by 48% compared to other state-of-the-art base-lines. The framework maintains real-time performance (3.7ms average latency) on resource-constrained devices with 512MB RAM, validated through deployment on 50 heterogeneous edge devices.

Information assurance (IA) is a growing concern, since almost every aspect of our lives depends on distributed information systems and the frequency and sophistication of cyber attacks targeting these systems is on the rise. However, IA cannot be considered in isolation, as it also affects the Quality of Service (QoS); with limited resources, the security mechanisms employed for IA (e.g. firewalls, antivirus, encryption) usually adversely affect QoS levels delivered by a system. The system therefore needs to make tradeoffs between IA and QoS. These tradeoffs are complicated by the facts that users’ relative preferences over QoS–IA change based on the situation, the preferences of different users conflict and tradeoff decisions made at one node in the distributed system typically affect other nodes as well. We address the problem of distributed computation of tradeoffs among various aspects of QoS and IA in a way that maximizes the satisfaction of all stakeholders. Specifically, we want the nodes in the system to make coordinated decisions as to what local actions to take to optimize QoS/IA levels delivered by the system. Our first contribution is formulating this problem as a distributed constraint optimization problem (DCOP). This entails quantifying various notions involved in tradeoffs to be able to compare options in the course of optimization, as well as encoding the effects of various decisions on the quantities we want to optimize. The DCOPs we obtain have cost functions where multiple local configurations result in the same cost. In addition, the corresponding factor graphs contain many cycles. To deal with these issues, our second contribution is a value propagation (VP) algorithm that helps nodes reach a consistent set of decisions even in cyclic factor graphs with non-unique local optima. We present experimental results comparing the performance of the max-sum algorithm with and without VP against other algorithms when applied to domain-inspired and random instances. On the domain-inspired instances, max-sum with VP achieves near optimal solutions at a fraction of time taken by Distributed Pseudotree Optimization Procedure, with the reduction in solution cost due to VP most pronounced in scenarios with resource contention. On random instances, VP obtains solutions with cost 1.7× of optimal, even when performed without a utility propagation algorithm first.

The complexity and tight integration of electromechanical systems often makes them "brittle" and hard to modify in response to changing requirements. We aim to remedy this by capturing expert knowledge as functional blueprints, an idea inspired by regulatory processes that occur in natural morphogenesis. We then apply this knowledge in an intelligent design variation tool. When a user modifies a design, our tool uses functional blueprints to modify other components in response, thereby maintaining integration and reducing the need for costly search or constraint solving. In this paper, we refine the functional blueprint concept and discuss practical issues in applying it to electromechanical systems. We then validate our approach with a case study applying our prototype tool to create variants of a miniDroid robot and by empirical evaluation of convergence dynamics of networks of functional blueprints.

This paper discusses the receiver design of a destination node supporting multiple antennas for an alternate-relaying decode-and-forward (DF) cooperative communication system. We exploit the structure of the received signal to develop an optimal data detection algorithm at the destination. It is shown that the optimal algorithm can be implemented by parallel detectors, each based on a family of Bahl, Cocke, Jelinek and Raviv (BCJR) algorithms. The proposed optimal algorithm requires to receive and store all packets before performing data detection. To avoid this, a sub-optimal algorithm is also proposed. Unlike the optimal algorithm, the sub-optimal one exploits two consecutive received packets to detect one packet. It turns out that the sub-optimal algorithm has less reduced delay, complexity, memory size and bandwidth loss with a slight increase of the bit-error-rate. The detection performance of the proposed algorithms is evaluated via Monte Carlo simulations, and the results demonstrate their effectiveness.

A key issue in the full‐rate alternate‐relaying cooperative communication systems is the interference which is caused by the simultaneous transmission of the source and one of the relays. In this study, the authors propose maximum‐likelihood (ML) detectors to mitigate the interference in such systems. Unlike previous work in which interference cancellation is required at the destination, the authors exploit the interference signal as a beneficial resource to develop an optimal detector. It is shown that the optimal detector can be implemented by parallel Viterbi algorithms. The major drawback of the proposed optimal detector is the delay because the destination has to receive and store the entire frame before performing data detection. Owing to the inevitable delay restriction, a sub‐optimal detector is developed. In contrast with the optimal detector, the sub‐optimal detector exploits two consecutive received packets to decode one packet. It turns out that the sub‐optimal detector significantly reduces the required delay, memory size and bandwidth loss, with a slight increase of the bit‐error‐rate and the computational complexity. Extensive simulation results have been presented to demonstrate the effectiveness of the proposed detectors.

Many natural organisms exhibit canalization: small genetic changes are accommodated by adaptation in other systems that interact with them. Engineered systems, however, are typically quite brittle, making design automation extremely difficult. We propose to address this problem with a generative representation of design based on manifold operators. The operator set we propose combines the intuitive simplicity of top-down rewrite rules with the flexibility and distortion tolerance of bottom-up GRN-based models. An embryogeny specified using this representation thus places constraints on a developing design, rather than specifying a fixed body plan, allowing canalization processes to modulate the design as it continues to develop. We demonstrate our ideas in the domain of electromechanical design and validate them with simulations at different levels of abstraction.

In this paper, we propose the use of bit-interleaved coded modulation iterative decoding (BICM-ID) in the full-rate decode and forward cooperative communication systems. At the destination, we exploit the interference signal to develop the optimal detector. It is shown that the proposed detector is implemented by parallel concatenating maximum a posteriori (MAP) algorithms and dampers to the decoders. The detector exchanges soft information between decoders and MAP algorithms in an iterative way, thus, improving communication reliability. Due to the inevitable delay restriction of the optimal detector, a sub-optimal detector is developed. Extensive simulation results are presented to demonstrate the effectiveness of proposed detectors. Results indicate that the proposed detectors outperform conventional relaying detectors in terms of their bit error rate.

We present a novel ensemble architecture for learning problem-solving techniques from a very small number of expert solutions and demonstrate its effectiveness in a complex real-world domain. The key feature of our “Generalized Integrated Learning Architecture” (GILA) is a set of heterogeneous independent learning and reasoning (ILR) components, coordinated by a central meta-reasoning executive (MRE). The ILRs are weakly coupled in the sense that all coordination during learning and performance happens through the MRE. Each ILR learns independently from a small number of expert demonstrations of a complex task. During performance, each ILR proposes partial solutions to subproblems posed by the MRE, which are then selected from and pieced together by the MRE to produce a complete solution. The heterogeneity of the learner-reasoners allows both learning and problem solving to be more effective because their abilities and biases are complementary and synergistic. We describe the application of this novel learning and problem solving architecture to the domain of airspace management, where multiple requests for the use of airspaces need to be deconflicted, reconciled, and managed automatically. Formal evaluations show that our system performs as well as or better than humans after learning from the same training data. Furthermore, GILA outperforms any individual ILR run in isolation, thus demonstrating the power of the ensemble architecture for learning and problem solving.

In this paper, we propose maximum likelihood (ML) detectors to mitigate the influence of the interference signal for the two-relays full- rate cooperative systems. At the relays, the proposed ML detector is employed by averaging out the interference signal. Furthermore, at the destination, we exploit the interference signal to develop the ML detector. It is shown that the optimal detector is implemented by parallel Viterbi algorithms. The major drawback of the proposed optimal detector is the delay, i.e a destination has to receive and store the whole received packets before performing data detection. Due to the inevitable delay restriction, sub-optimal detector is developed. In contrast with the optimal detector, the sub-optimal detector exploits two consecutive received packets to decode one packet. It turns out that the sub-optimal detector outperforms the optimal detector in terms of the required delay, memory size, bandwidth loss, and computational complexity. Extensive simulation results have been presented to demonstrate the effectiveness of the proposed detectors. Results indicate that the proposed detectors outperform conventional relaying detectors in terms of their bit error rate and packet error rate.

A key issue in the full-rate two-relay cooperative communication systems is the interference which is caused by the simultaneous transmission of the source and one of the relays at any time. In this paper, we exploit the interference signal at the destination to develop a maximum likelihood (ML) detector for decode and forward full-rate cooperative systems. It is shown that the Viterbi algorithm can be employed to find the ML solution. To reduce the complexity of the proposed ML detector, a sub-optimal detector is also introduced. Further, we propose a ML interference cancellation scheme at the relays. The performance of the proposed schemes is evaluated through Monte Carlo simulations. Results indicate that the proposed schemes outperform direct transmission and conventional relaying schemes in terms of their bit error rate.