This paper describes an identifier for a class of nonlinear systems based on continuous recurrent neural networks (CRNN). The identifier is proposed considering the approximation properties of artificial neural networ...
详细信息
ISBN:
(纸本)9798331517519;9798331517526
This paper describes an identifier for a class of nonlinear systems based on continuous recurrent neural networks (CRNN). The identifier is proposed considering the approximation properties of artificial neural networks. The learning or adaptive laws are obtained using Barrier Lyapunov functions with an exponentially decreasing barrier. The application of such a function results in a bounded identification error with an exponential convergence and predefined decay. Additionally, it ensures the convergence of the weights for the activation functions to the fitting values. The proposed identifier was used to identify a virtual Cartesian robot with two degrees of freedom. The results showed the performance of the identification error, which does not violate the imposed exponential barrier. Moreover, the effect of the predefined convergence parameter was observed in the identification error evolution without the need for the change of any other parameter in the CRNN.
Industrial controlsystems (ICSs), such as Supervisory control and Data Acquisition (SCADA) systems, are increasingly popular for manufacturing applications, leading to significant improvements in efficiency and produ...
详细信息
ISBN:
(纸本)9798350313871
Industrial controlsystems (ICSs), such as Supervisory control and Data Acquisition (SCADA) systems, are increasingly popular for manufacturing applications, leading to significant improvements in efficiency and productivity. However, the vulnerability of these systems to ransomware attacks has become a major concern. This vulnerability is mainly due to the centralized nature of ICSs, which prioritize efficiency over security. To address this issue, this paper proposes a decentralized Blockchain-Based ICS (BBICS) architecture. Such architecture uses a peer-to-peer network of nodes to replicate critical data and distribute transactions using a consensus mechanism, which synchronizes nodes and resolves single points of failure. Additionally, BBICS encrypts critical data in a tamper-resistant manner to prevent attackers from decrypting or manipulating data. Moreover, zero-trust authorization and authentication further enhance security by preventing the broadcasting of ransomware attacks in internal networks of devices. The evaluation of the proposed system with respect to performance and reliability under normal and ransomware attack situations suggest BBICS' feasibility and practicality.
This paper investigates the output-based predictive control problem for networked controlsystems (NCSs) with delays and data loss using the dynamic event-triggered mechanism (DETM). And the output-based DETM is propo...
详细信息
Ad hoc network is a sort of self-organizing network that relies on no fixed infrastructure. Each node can act as a host to send and receive its own data, or as a possible router to forward data to other nodes. In the ...
详细信息
With the increasing computing power, using data-driven approaches to co-design a robot's morphology and controller has become a promising way. However, most existing data-driven methods require training the contro...
详细信息
ISBN:
(纸本)9781665491907
With the increasing computing power, using data-driven approaches to co-design a robot's morphology and controller has become a promising way. However, most existing data-driven methods require training the controller for each morphology to calculate fitness, which is time-consuming. In contrast, the dual-network framework utilizes data collected by individual networks under a specific morphology to train a population network that provides a surrogate function for morphology optimization. This approach replaces the traditional evaluation of a diverse set of candidates, thereby speeding up the training. Despite considerable results, the online training of both networks impedes their performance. To address this issue, we propose a concurrent network framework that combines online and offline reinforcement learning (RL) methods. By leveraging the behavior cloning term in a flexible manner, we achieve an effective combination of both networks. We conducted multiple sets of comparative experiments in the simulator and found that the proposed method effectively addresses issues present in the dual-network framework, leading to overall algorithmic performance improvement. Furthermore, we validated the algorithm on a real robot, demonstrating its feasibility in a practical application.
The Space-Air-Ground Integrated network (SAGIN) includes LEO satellite communication and an air-based network with high-altitude UAVs, communication aircraft, and strategic nodes. The air network boasts rapid access, ...
详细信息
The rapid growth of the digital industry has created a higher demand for robust network Intrusion Detection systems (NIDS) to protect valuable information and the integrity of network infrastructures as the digital in...
详细信息
ISBN:
(纸本)9798331518097
The rapid growth of the digital industry has created a higher demand for robust network Intrusion Detection systems (NIDS) to protect valuable information and the integrity of network infrastructures as the digital industry grows rapidly. One of the most important challenges in the current intrusion detection landscape is the growing sophistication of cyber threats, including zero-day attacks, polymorphic malware, and advanced persistent threats, which are difficult to detect using traditional methods. Furthermore, systems often suffer from high false positive rates and struggle to scale effectively in real-time applications. Traditionally, intrusion detection methods were quite effective, but performance is still lacking due to the inability to adapt to evolving threats. Recent breakthroughs include deep learning approaches, ensemble methods, and hybrid detection models. However, these are still plagued by high computational overhead and a lack of transparency in their decision-making processes. The work exploits Optuna for the optimization of hyperparameters, specifically in the performance improvement of various ML models. Among the best-ranked frameworks for the optimization of hyperparameters, Optuna provides a principled method for tuning hyperparameters, resulting in significantly enhanced accuracy and efficiency of the intrusion detection model. The implication of this research work is that it searches for the best configuration of parameters for each algorithm with balanced false positives and detection rates. The study includes an overall scenario of recent development in NIDS. More precisely, this paper shows how Hyperparameter tuning attains very superior model performance compared to other models. The comparative results presented have shown that models which are optimized using Optuna surpass the non-optimized ones by a huge margin with respect to accuracy, recall, precision, and F1-score. The paper also discusses ensemble techniques by integrating the
Neural networkcontrollers (NNCs) have shown great promise in autonomous and cyber-physical systems. Despite the various verification approaches for neural networks, the safety analysis of NNCs remains an open problem...
详细信息
ISBN:
(纸本)9781577358800
Neural networkcontrollers (NNCs) have shown great promise in autonomous and cyber-physical systems. Despite the various verification approaches for neural networks, the safety analysis of NNCs remains an open problem. Existing verification approaches for neural networkcontrolsystems (NNCSs) either can only work on a limited type of activation functions, or result in non-trivial over-approximation errors with time evolving. This paper proposes a verification framework for NNCS based on Lipschitzian optimisation, called DeepNNC. We first prove the Lipschitz continuity of closed-loop NNCSs by unrolling and eliminating the loops. We then reveal the working principles of applying Lipschitzian optimisation on NNCS verification and illustrate it by verifying an adaptive cruise control model. Compared to state-of-the-art verification approaches, DeepNNC shows superior performance in terms of efficiency and accuracy over a wide range of NNCs. We also provide a case study to demonstrate the capability of DeepNNC to handle a real-world, practical, and complex system. Our tool DeepNNC is available at https://***/TrustAI/DeepNNC.
In this paper, we consider the problem of reference tracking in uncertain nonlinear systems. A neural State-Space Model (NSSM) is used to approximate the nonlinear system, where a deep encoder network learns the nonli...
详细信息
ISBN:
(纸本)9798350358513;9798350358520
In this paper, we consider the problem of reference tracking in uncertain nonlinear systems. A neural State-Space Model (NSSM) is used to approximate the nonlinear system, where a deep encoder network learns the nonlinearity from data, and a state-space component captures the temporal relationship. This transforms the nonlinear system into a linear system in a latent space, enabling the application of model predictive control (MPC) to determine effective control actions. Our objective is to design the optimal controller using limited data from the target system (the system of interest). To this end, we employ an implicit model-agnostic meta-learning (iMAML) framework that leverages information from source systems (systems that share similarities with the target system) to expedite training in the target system and enhance its controlperformance. The framework consists of two phases: the (offine) meta-training phase learns an aggregated NSSM using data from source systems, and the (online) meta-inference phase quickly adapts this aggregated model to the target system using only a few data points and few online training iterations, based on local loss function gradients. The iMAML algorithm exploits the implicit function theorem to exactly compute the gradient during training, without relying on the entire optimization path. By focusing solely on the optimal solution, rather than the path, we can meta-train with less storage complexity and fewer approximations than other contemporary meta-learning algorithms. We demonstrate through numerical examples that our proposed method can yield accurate predictive models by adaptation, resulting in a downstream MPC that outperforms several baselines.
To reduce CO2 emissions and tackle increasing fuel costs, the aviation industry is swiftly moving towards the electrification of aircraft. From the viewpoint of systems and control, a key challenge brought by this tra...
详细信息
ISBN:
(纸本)9798350382662;9798350382655
To reduce CO2 emissions and tackle increasing fuel costs, the aviation industry is swiftly moving towards the electrification of aircraft. From the viewpoint of systems and control, a key challenge brought by this transition corresponds to the management and safe operation of the propulsion system's onboard electrical power distribution network. In this work, for a series-hybrid-electric propulsion system, we propose a distributed adaptive controller for regulating the voltage of a DC bus that energizes the electricity-based propulsion system. The proposed controller-whose design is based on principles of back-stepping, adaptive, and passivity-based control techniques-also enables the proportional sharing of the electric load among multiple converter-interfaced sources, which reduces the likelihood of over-stressing individual sources. Compared to existing control strategies, our method ensures stable, convergent, and accurate voltage regulation and load sharing even if the effects of power lines of unknown resistances are considered. The performance of the proposed control scheme is illustrated via numerical simulations of an exemplary propulsion architecture.
暂无评论