自动化学报最新文献

Two new approximate formulations to joint chance-constrained optimization problems are proposed in this paper. The relationships of CVaR (conditional-value-at-risk), chance constrains and robust optimization are reviewed. Firstly, two new upper bounds on E((·) ⁺) are proposed, where E stands for the expectation and x⁺ = max(0, x), based on which two approximate formulations for individual chance-constrained problems are derived. The approximations are proved to be the robust optimization with the corresponding uncertain sets. Then the approximations are extrapolated to joint chance-constrained problem. Finally numerical studies are performed to compare the solutions of individual and joint chance constraints approximations and the results demonstrate the validity of our method.

A pseudospectral method is presented for direct trajectory optimization of optimal control problems using collocation at Chebyshev-Gauss points, and therefore, it is called Chebyshev-Gauss pseudospectral method. The costate and constraint multiplier estimates for the proposed method are rigorously derived by comparing the discretized optimality conditions of an optimal control problem with the Karush-Kuhn-Tucker conditions of the resulting nonlinear programming problem from collocation. The distinctive advantages of the proposed method over other pseudopsectral methods are the good numerical stability and computational efficiency. In order to achieve this goal, the barycentric Lagrange interpolation is substituted for the classic Lagrange interpolation in the state approximation. Furthermore, a simple yet efficient method is presented to alleviate the numerical errors of state differential matrix using the trigonometric identity especially when the number of Chebyshev-Gauss points is large. The method presented in this paper has been taken to two optimal control problems from the open literature, and the results have indicated its ability to obtain accurate solutions to complex constrained optimal control problems.

Synthetic aperture radar (SAR) image is usually polluted by multiplicative speckle noise, which can affect further processing of SAR image. This paper presents a new approach for multiplicative noise removal in SAR images based on sparse coding by shearlets filtering. First, a SAR despeckling model is built by the theory of compressed sensing (CS). Secondly, obtain shearlets coefficient through shearlet transform, each scale coefficient is represented as a unit. For each unit, sparse coefficient is iteratively estimated by using Bayesian estimation based on shearlets domain. The represented units are finally collaboratively aggregated to construct the despeckling image. Our results in SAR image despeckling show the good performance of this algorithm, and prove that the algorithm proposed is robustness to noise, which is not only good for reducing speckle, but also has an advantage in holding information of the edge.

In this work, the affine point set matching is formulated under a variational Bayesian framework and the model points are projected forward into the scene space by a linear transformation. A directed acyclic graph is presented to represent the relationship between the parameters, latent variables, model and scene point sets and an iterative approximate algorithm is proposed for the estimation of the posterior distributions over parameters. Furthermore, the anisotropic covariance is assumed on the transition variable and one Gaussian component is provided for the inference of outlier points. Experimental results demonstrate that the proposed algorithm achieves good performance in terms of both robustness and accuracy.

In this paper, the problem of state feedback stabilization for stochastic feedforward nonlinear systems with input time-delay is considered for the first time. By introducing a variable transformation, skillfully combining the homogeneous domination method, and constructing an appropriate Lyapunov-Krasovskii functional, a state feedback controller is developed to guarantee the closed-loop system globally asymptotically stable in probability.

In this paper, we investigate the nonlinear control problem for multi-agent formations with communication delays in noisy environments and in directed interconnection topologies. A stable theory of stochastic delay differential equations is established and then some sufficient conditions are obtained based on this theory, which allow the required formations to be gained at exponentially converging speeds with probability one for time-invariant formations, time-varying formations, and time-varying formations for trajectory tracking under a special “multiple leaders” framework. Some numerical simulations are also given to illustrate the effectiveness of the theoretical results.