Anti-saturation attack (ASA) strategy is vital for the survival of a warship group, and attracts the focus of many researchers. In this paper, the dynamics of ASA is formulated as a Markov Decision Process (MDP) with ...
详细信息
Anti-saturation attack (ASA) strategy is vital for the survival of a warship group, and attracts the focus of many researchers. In this paper, the dynamics of ASA is formulated as a Markov Decision Process (MDP) with an enhanced states space since those characters are involved, such as the formation and detection and interception areas of warship group. A reinforcement learning method (Double Deep Q-leaning, DDQN) is developed to solve the problem and deal with the curse of dimensionality whereby the cost-to-go value is calculated by a marine engagement simulation system. A heuristic defense algorithm guided by field expert knowledge is designed for comparison. The experimental results show that the DDQN method performs better in anti-saturation attack scenarios.
The four-step transportation model plays an important role in urban planning. The quality of the first phase, i.e. trip generation, determines the performance of the global course. The majority of trip generation fore...
详细信息
The four-step transportation model plays an important role in urban planning. The quality of the first phase, i.e. trip generation, determines the performance of the global course. The majority of trip generation forecasting models highly rely on mathematical derivation and have many predictor variables during the prediction, which leads to low accuracy of results and requires laboriously hand-crafted design of input vectors. This paper is the first to introduce the gradient boosting decision tree (GBDT) algorithm for trip generation prediction, and harmonizes such a powerful machine learning method with traditional urban planning requirements to achieve better prediction performance. Unlike the commonly used linear regression method, GBDT can automatically perform feature selection and model the non-linear relationships between input and output variables. Experimental results on real-world residential travel census in Beijing prove that the GBDT model significantly outperforms the baseline and can forecast the trip generation more accurately.
In sensor networks, due to inevitable sensor faults, malfunctions, or deliberate attacks, sensors may transmit erroneous, inaccurate, or misleading data, thereby degrading overall system performance. To address this i...
详细信息
Despite remarkable achievements in neural network applications using spin–orbit torque (SOT) devices, current research predominantly focuses on unimodal classification tasks, while cross-modal learning tasks often su...
详细信息
Despite remarkable achievements in neural network applications using spin–orbit torque (SOT) devices, current research predominantly focuses on unimodal classification tasks, while cross-modal learning tasks often suffer from suboptimal generation quality. To address this limitation, we employ generative adversarial networks (GANs) based on CoPt-SOT to achieve cross-modal learning and generation from speech to handwritten digit images. Initially, we explore the field-free switching characteristics of single-layer CoPt devices, developing spin-based implementations of Scaled Rectified Linear Unit (SReLu), Sigmoid, and Tanh neuronal functions. Simultaneously, we leverage the nonvolatile multistate characteristics of CoPt devices to realize excitatory-inhibitory synaptic plasticity. Subsequently, for the first time, we construct half-spin and full-spin GAN networks using these spin-based neurons and synapses, enabling cross-modal learning and generation between speech and image domains. Finally, we evaluate the performance of these networks through handwritten digit recognition tasks, achieving impressive recognition accuracies of 93.78 % and 88.61 % for half-spin and full-spin implementations, respectively. Notably, this work overcomes the existing bottleneck in SOT device applications, which have been largely confined to unimodal classification tasks, and demonstrates significant potential for expanding the scope of SOT-based technologies.
The ℋ ∞ consensus control problem is handled for a class of discrete time-varying multi-agent systems with round-robin protocol and missing measurements. The round-robin protocol is considered, which allows only one...
详细信息
The ℋ ∞ consensus control problem is handled for a class of discrete time-varying multi-agent systems with round-robin protocol and missing measurements. The round-robin protocol is considered, which allows only one agent can send its measurement data at each transmission step in order to prevent the data collisions. And the missing measurements phenomenon is described by a sequence of Bernoulli distributed random variables with known probabilities. We focus on designing the controller parameters to guarantee that the closed-loop multi-agent systems with missing measurements satisfy the ℋ ∞ consensus performance. On this basis, some sufficient and necessary conditions are obtained by solving coupled backward recursive Riccati difference equations, where the existence and feasibility of the control scheme can be obtained with the help of the completing squares method and the stochastic analysis technique. At last, a numerical example is given to illustrate the effectiveness of the developed controller design scheme.
The H consensus control problem is handled for a class of discrete time-varying multi-agent systems with round-robin protocol and missing measurements. The round-robin protocol is considered, which allows only one age...
详细信息
The H consensus control problem is handled for a class of discrete time-varying multi-agent systems with round-robin protocol and missing measurements. The round-robin protocol is considered, which allows only one agent can send its measurement data at each transmission step in order to prevent the data collisions. And the missing measurements phenomenon is described by a sequence of Bernoulli distributed random variables with known probabilities. We focus on designing the controller parameters to guarantee that the closed-loop multi-agent systems with missing measurements satisfy the H consensus performance. On this basis, some sufficient and necessary conditions are obtained by solving coupled backward recursive Riccati difference equations, where the existence and feasibility of the control scheme can be obtained with the help of the completing squares method and the stochastic analysis technique. At last, a numerical example is given to illustrate the effectiveness of the developed controller design scheme.
Traffic congestion is a serious problem around the world and to a great extent influences urban communities in various manners including increased stress levels, delayed deliveries, fuel wastage, and monetary losses. ...
详细信息
Traffic congestion is a serious problem around the world and to a great extent influences urban communities in various manners including increased stress levels, delayed deliveries, fuel wastage, and monetary losses. Therefore, an accurate congestion prediction algorithm to limit these misfortunes is fundamental. This paper presents a comparative study of traffic congestion prediction systems including decision tree, logistic regression, and neural networks. Five days of traffic information (1,231,200 samples) are utilized to drive the prediction model. The TensorFlow and the Clementine machine learning platforms are used for data preprocessing, training, and testing of the model. The confusion matrix clears that decision tree has better prediction performance and leads the other two methods with accuracy (97%), macro-average precision (95%), macro-average recall (96%), and macro-average F1_score (96%) in the python programming environment. Moreover, performance of the three prediction models is verified in Clementine environment and decision tree outperforms all other models with an accuracy of 97.65%.
We study the formation control problem for a group of mobile agents in a plane, in which each agent is modeled as a kinematic point and can only use the local measurements in its local frame. The agents are required t...
详细信息
The integration of signals from physical, social and cyber spaces, known as Cyber-Physical-Social systems (CPSS), is a new research paradigm for urban transportation, where the traffic control and management (C&M)...
详细信息
The integration of signals from physical, social and cyber spaces, known as Cyber-Physical-Social systems (CPSS), is a new research paradigm for urban transportation, where the traffic control and management (C&M) is collaborative optimized among the three sub-systems. Though some technologies and optimization methods have been studied since its proposition, there is a lack of a systemic architecture as well as an overall implementation about how to efficiently exploit the social signals. For this reason, this paper proposes a general framework of CPSS for urban transportation and presents a feasible solution for traffic optimization based on knowledge automation. The specific implementation includes basic modeling of CPSS, knowledge evolution and reasoning, and collaborative optimization of C&P strategies. As a remarkable highlight, the influence of both individual activities and social learning is concerned during knowledge evolution and reasoning part. A case study from the application in the city of Dongguan is also given to validate our proposed framework and methods, showing that they can efficiently improve the average speed of the actual transportation.
Cyber-Physical-Social systems (CPSS) provides a novel perspective for constructing “Smart City”, which is also known as the Human-Machine-Things-System (HMTS), focusing on the fusion of ternary space: social network...
详细信息
Cyber-Physical-Social systems (CPSS) provides a novel perspective for constructing “Smart City”, which is also known as the Human-Machine-Things-System (HMTS), focusing on the fusion of ternary space: social network of human society, network of machines and the Internet of things. In this paper, we propose a specific implementation framework of CPSS for Smart City based on intelligent loops, including basic modeling and interactive fusion, state perception and cognition, and adaptive learning. On this basis, an overall architecture of the CPSS platform is designed, which is applied in the urban transportation management in Hangzhou. The application results demonstrate that the intelligent loop could optimize the control and management strategies for actual urban transportation.
暂无评论