Automatic texture feature selection for image pixel classification

Pixel-based texture classifiers and segmenters are typically based on the combination of texture feature extraction methods that belong to a single family (e.g., Gabor filters). However, combining texture methods from different families has proven to produce better classification results both quantitatively and qualitatively. Given a set of multiple texture feature extraction methods from different families, this paper presents a new texture feature selection scheme that automatically determines a reduced subset of methods whose integration produces classification results comparable to those obtained when all the available methods are integrated, but with a significantly lower computational cost. Experiments with both Brodatz and real outdoor images show that the proposed selection scheme is more advantageous than well-known general purpose feature selection algorithms applied to the same problem.     

A new compression based approach for reconfiguration overhead reduction in virtex based RTR systems

The advantage of RTR systems usually comes with some costs. The required time to map some areas of a program to an FPGA is considerable and affects the performance of RTR systems. Several methods have been developed to speed up the configuration process in these systems. Configuration compression can reduce the total number of write operations to load a configuration and it has been proven to be an efficient technique for dealing with the configuration overhead. In this paper, we have developed a new approach for reconfiguration overhead reduction in Virtex Based RTR Systems by using a compression technique based on Lempel-Ziv (LZ) algorithm. Since the order of the sequence of configuration frames affects the compression rate, we have proposed an algorithm based on Genetic Algorithm for finding the optimal configuration sequence of frames. The proposed algorithm will be applied to the input configuration file in a batch (offline) manner, and its time complexity is tolerable considering the overhead reduction obtained by having the optimal sequence of frames in run-time configuration decompression. Also, corresponding to our approach, a hardware model has been designed for configuration decompression.     

Stability regions in the parameter space: D-decomposition revisited

The challenging problem in linear control theory is to describe the total set of parameters (controller coefficients or plant characteristics) which provide stability of a system. For the case of one complex or two real parameters and SISO system (with a characteristic polynomial depending linearly on these parameters) the problem can be solved graphically by use of the so-called D-decomposition. Our goal is to extend the technique and to link it with general M-Δ framework. In this way we investigate the geometry of D-decomposition for polynomials and estimate the number of root invariant regions. Several examples verify that these estimates are tight. We also extend D-decomposition for the matrix case, i.e. for MIMO systems. For instance, we partition real axis or complex plane of the parameter k into regions with invariant number of stable eigenvalues of the matrix A+kB. Similar technique can be applied to double-input double-output systems with two parameters.     

Stability analysis and stabilization control of multi-variable switched stochastic systems

In this paper, the mean square (MS) stability and exponential mean square (EMS) stability of multi-variable switched stochastic systems are investigated. Based on the concept of the average dwell-time and the ratio of the total time running on all unstable subsystems to the total time running on all stable subsystems, some sufficient conditions are given to ensure the MS stability and EMS stability of the switched stochastic systems involved. Further, for the switched stochastic control systems with all subsystems controllable or stabilizable, EMS stabilization controls and sufficient conditions on EMS stabilization are presented, and the convergent rates of the closed-loop systems are obtained.     

Continuous stabilization controllers for singular bilinear systems: The state feedback case

Global asymptotic stabilization for a class of singular bilinear systems is first studied in this paper. New approaches are developed by means of the LaSalle invariant principle for nonlinear systems. A new set of sufficient condition is first derived via the continuous static state feedback, the feedback not only guarantees the existence of solution but also the global asymptotical stabilization for the closed loop system.     

On a stochastic sensor selection algorithm with applications in sensor scheduling and sensor coverage

In this note we consider the following problem. Suppose a set of sensors is jointly trying to estimate a process. One sensor takes a measurement at every time step and the measurements are then exchanged among all the sensors. What is the sensor schedule that results in the minimum error covariance? We describe a stochastic sensor selection strategy that is easy to implement and is computationally tractable. The problem described above comes up in many domains out of which we discuss two. In the sensor selection problem, there are multiple sensors that cannot operate simultaneously (e.g., sonars in the same frequency band). Thus measurements need to be scheduled. In the sensor coverage problem, a geographical area needs to be covered by mobile sensors each with limited range. Thus from every position, the sensors obtain a different view-point of the area and the sensors need to optimize their trajectories. The algorithm is applied to these problems and illustrated through simple examples.     

Adaptive output feedback control methodology applicable to non-minimum phase nonlinear systems

An adaptive output feedback control methodology is developed for a class of uncertain multi-input multi-output nonlinear systems using linearly parameterized neural networks. The methodology can be applied to non-minimum phase systems if the non-minimum phase zeros are modeled to a sufficient accuracy. The control architecture is comprised of a linear controller and a neural network. The neural network operates over a tapped delay line of memory units, comprised of the system's input/output signals. The adaptive laws for the neural-network weights employ a linear observer of the nominal system's error dynamics. Ultimate boundedness of the error signals is shown through Lyapunov's direct method. Simulations of an inverted pendulum on a cart illustrate the theoretical results.     

Controller design for Markov jumping systems subject to actuator saturation

In this paper, the stochastic stabilization problem for a class of Markov jumping linear systems (MJLS) subject to actuator saturation is considered. The concept of domain of attraction in mean square sense is used to analyze the closed-loop stability. When the jumping mode is available, a mode-dependent state feedback controller is developed. Otherwise, we give a less conservative approach to design the mode-independent state feedback controller. Both design procedures can be converted into a set of linear matrix inequalities (LMIs). Finally, a numerical example is provided to show the effectiveness of the techniques.     

The control of a two-level Markov decision process by time aggregation

The solution of Markov Decision Processes (MDPs) often relies on special properties of the processes. For two-level MDPs, the difference in the rates of state changes of the upper and lower levels has led to limiting or approximate solutions of such problems. In this paper, we solve a two-level MDP without making any assumption on the rates of state changes of the two levels. We first show that such a two-level MDP is a non-standard one where the optimal actions of different states can be related to each other. Then we give assumptions (conditions) under which such a specially constrained MDP can be solved by policy iteration. We further show that the computational effort can be reduced by decomposing the MDP. A two-level MDP with M upper-level states can be decomposed into one MDP for the upper level and M to M(M-1) MDPs for the lower level, depending on the structure of the two-level MDP. The upper-level MDP is solved by time aggregation, a technique introduced in a recent paper [Cao, X.-R., Ren, Z. Y., Bhatnagar, S., Fu, M., & Marcus, S. (2002). A time aggregation approach to Markov decision processes. Automatica, 38(6), 929-943.], and the lower-level MDPs are solved by embedded Markov chains.     

Advertising an event

The paper considers a problem of how to minimize advertising costs to sell seats for a particular event, for instance, a sports game, a rock concert or a ballet performance. We take into consideration a word-of-mouth effect, which means that people buying a ticket tell their friends about it, so that advertising is unnecessary to inform those people.The problem is one of optimal control and the number of seats sold and the advertising effort of the organizers are the state and control variables, respectively. We show that, besides being dependent on the cost and revenue parameters, the optimal advertising policy is also affected by the length of the planning period and the relation between the number of seats and the total number of potential attendees.