In this paper, we present an H∞filter for a giant cavity system which couples to a single-mode waveguide at two spatially separated points. The dynamics of the cavity is described by a time-delay Langevin equation in the Heisenberg picture whose time-delay term results from the non-local interactions to a waveguide. The input-output relation with a time-delay feature is also derived. In addition, we design an H∞filter for estimating the evolution of the mode in the cavity using a quadrature representation state-space representation and Linear Matrix Inequality (LMI) approach. Finally, the efficacy of our method is numerically demonstrated in an example.

Federated learning is vulnerable to poisoning attacks by malicious adversaries. Existing methods often involve high costs to achieve effective attacks. To address this challenge, we propose a sybil-based virtual data poisoning attack, where a malicious client generates sybil nodes to amplify the poisoning model’s impact. To reduce neural network computational complexity, we develop a virtual data generation method based on gradient matching. We also design three schemes for target model acquisition, applicable to online local, online global, and offline scenarios. In simulation, our method outperforms other attack algorithms since our method can obtain a global target model under non-independent uniformly distributed data.

Lane change for automated vehicles (AVs) is an important but challenging task in complex dynamic traffic environments. Due to difficulties in guaranteeing safety as well as a high efficiency, AVs are inclined to choose relatively conservative strategies for lane change. To avoid the conservatism, this paper presents a cooperation-aware lane change method utilizing interactions between vehicles. We first propose an interactive trajectory prediction method to explore possible cooperations between an AV and the others. Further, an evaluation on safety, efficiency and comfort is designed to make a decision on lane change. Thereafter, we propose a motion planning algorithm based on model predictive control (MPC), which incorporates AV’s decision and surrounding vehicles’ interactive behaviors into constraints so as to avoid collisions during lane change. Quantitative testing results show that compared with the methods without an interactive prediction, our method enhances driving efficiencies of the AV and other vehicles by 14.8% and 2.6%, respectively, which indicates that a proper utilization of vehicle interactions can effectively reduce the conservatism of the AV and promote the cooperation between the AV and others.

Forecasting the trajectories of neighbor vehicles is a crucial step for decision making and motion planning of autonomous vehicles. This paper proposes a graph-based spatial-temporal convolutional network (GSTCN) to predict future trajectory distributions of all neighbor vehicles using past trajectories. This network tackles spatial interactions using a graph convolutional network (GCN), and captures temporal features with a convolutional neural network (CNN). The spatial-temporal features are encoded and decoded by a gated recurrent unit (GRU) network to generate future trajectory distributions. Besides, we propose a weighted adjacency matrix to describe the intensities of mutual influence between vehicles, and the ablation study demonstrates the effectiveness of our scheme. Our network is evaluated on two real-world freeway trajectory datasets: I-80 and US-101 in the Next Generation Simulation (NGSIM). Comparisons in three aspects, including prediction errors, model sizes, and inference speeds, show that our network can achieve state-of-the-art performance.

In this paper, we focus on detection of quantum colored noise and present a novel detection method which employs a quantum harmonic oscillator as a noise probe. With respect to an unknown spectrum of the quantum colored noise, quantum spectral decomposition theorem is applied to obtain a linear system model for the internal modes of the quantum colored noise, where the spectrum is parameterized. Then we establish a linear augmented model for the whole system which consists of a probe of a quantum oscillator and the quantum linear system model of the quantum colored noise. Due to the equivalence between the augmented model and the actual system, we estimate unknown parameters of the linear system model for the noise by solving an optimization problem. The effectiveness of our method is verified by an example of identification of quantum Lorentzian noise.

This paper presents a new method to estimate real-time queue lengths at a signalized intersection by utilizing limited data extracted from surveillance videos. This method focuses on reconstructing vehicles' trajectories on an entire road segment. The real-time queue length can be derived from these reconstructed trajectories. In order to improve the accuracy of the trajectory reconstruction, a built-up car-following model is proposed to reconstruct the trajectories of vehicles joining and leaving the queue, respectively, which are further corrected by a fusion algorithm. The proposed method was validated in the Next Generation Simulation dataset, and the queue length for each signal cycle can be estimated with a high accuracy. The results show that the proposed method has a higher precision compared to three baseline models.

In privacy-preserving machine learning, individual parties are reluctant to share their sensitive training data due to privacy concerns. Even the trained model parameters or prediction can pose serious privacy leakage. To address these problems, we demonstrate a generally applicable Distributed Privacy-Preserving Prediction (DPPP) framework, in which instead of sharing more sensitive data or model parameters, an untrusted aggregator combines only multiple models' predictions under provable privacy guarantee. Our framework integrates two main techniques to guarantee individual privacy. First, we introduce the improved Binomial Mechanism and Discrete Gaussian Mechanism to achieve distributed differential privacy. Second, we utilize homomorphic encryption to ensure that the aggregator learns nothing but the noisy aggregated prediction. Experimental results demonstrate that our framework has comparable performance to the non-private frameworks and delivers better results than the local differentially private framework and standalone framework.

This paper examines the event-triggered consensus problem of the multi-agent system on signed networks. An event-based distributed control protocol is proposed for multiagent networks where the dynamics of each agent is characterized by a controllable linear time-invariant system ( A, B). By exploring the Gauge transformation between Laplacian matrix and signed Laplacian matrix, we show that the bipartite consensus and trivial consensus can be achieved, respectively, in terms of the structural balance of the underlying signed network. Furthermore, the Zeno phenomenon of the closed-loop system is shown to be non-existed when employing the proposed event-based control protocol. Simulation results are finally provided to demonstrate our results.

In this paper, we present a linear least squares method to identify a damping rate function for a non-Markovian single qubit system. The dynamics of the system is described by a time-convolutionless master equation where the unknown damping rate function contains all the information of the environment. By expressing the function as a polynomial of time with unknown coefficients, we convert the identification problem into a parameter estimation problem. We transform the master equation into a Bloch differential equation, and obtain a reduced system whose output is the time trace of an observable. Then we use a linear least squares method to estimate unknown coefficients in the polynomial by utilizing the measured outputs. Finally, the effectiveness of our method is shown in an example of a two-level atom non-Markovian system.

In this paper, we present a gradient algorithm to identify a damping rate function for a non-Markovian single qubit system. The dynamics of the single qubit system in a non-Markovian environment are assumed to obey a time convolutionless master equation, where all the non-Markovian effects of the environment are combined in the unknown damping rate function. To identify the damping rate function, we measure time trace observables of the qubit such that we can formulate the identification procedure as an optimization problem. Thus, we design a gradient algorithm to optimally reveal the damping rate function.