This paper studies the problem of global decentralized control by output feedback for large-scale uncertain systems whose subsystems are interconnected not only by their outputs but also by their unmeasurable states. We show that under a linear growth condition, there is a decentralized output feedback controller rendering the closed-loop system globally exponentially stable. This is accomplished by extending an output feedback domination design that requires only limited information about the nonlinear system. We will apply our design to lower, upper, and non-triangular nonlinear systems. The significance of our results is that we do not need to have prior information about the nonlinearities of the system. Furthermore, we need to only employ a linear observer in combination with a linear controller to stabilize the system. A time-varying output feedback controller is also constructed for use with large-scale systems that have unknown parameters.

In this paper we consider the completeness problem of reasoning about planar spatial relationships in pictorial retrieval systems. We define a large class of two-dimensional scenes, the extended pseudo-symbolic pictures. The existing rule system R is proved to be complete for (extended) pseudo-symbolic pictures. We also introduce a new iconic indexing, the (extended) pseudo- 2D string representation, for them. The (extended) pseudo-2D string has the good properties of the 2D string. It is unambiguous, like the augmented 2D string, and has a compact form suitable for image retrieval. We then present efficient algorithms to determine whether a given planar picture is (extended) pseudo-symbolic or not, and if it is, these algorithms also return its (extended) pseudo-2D string representation. Picture retrieval by (extended) pseudo-2D strings is also discussed.

Optimization based flow control has been proposed in [2] to improve the network performance with congested bottle links. This rate-based technique has advantages over traditional window based heuristic algorithms in that the optimal performance in terms of maximal aggregate utility function can be achieved when each source adaptively adjusts its data rate. Several decentralized optimization algorithms have been applied to the flow control. However, one of most important features of these algorithms: the relation between the convergence speed and network parameters is not well studied, experimentally or theoretically. The contribution of this paper is two-fold. The first contribution is that we propose Aitken-extrapolation to accelerate the convergence process. Secondly, we compare the convergence speed of various algorithms by theoretic analysis and simulations. Based on the results, the network parameters can be appropriately chosen to improve network performance.

The problem of constructing and approximating control theoretic smoothing splines is considered in this paper. It is shown that the optimal approximating function can be given as the solution of a forced Hamiltonian system, that can be explicitly solved using the Riccati transform, and an explicit linear filter can be constructed. We show that the bandwidth of the filter can be naturally controlled and thus for control theoretic smoothing splines the far past and the far future are unimportant. Hence smoothing splines are “local” in nature rather than "global". We conclude that while spline approximations are not causal the far future is not important.